Morphometric identification of small mammal footprints from ink tracking tunnels in the Brazilian Cerrado

1. What they did

Clear footprints of small mammals are rarely found in nature, and many Neotropical rodents and marsupials have similar tracks, so they are poorly covered by track guides. The authors set out to adapt ink tracking tunnels for the forests of central Brazil, build a reference collection of known footprints, develop a way to identify footprints by measurement, and compare the tunnels with traps for surveying small mammals.

• Tunnels. They cut PVC water pipe into tunnels of three sizes (10, 7.5 and 5 cm across), lined them with paper, and placed an ink pad (blue stamp ink on a strip of plastic from a soft drink bottle) in the middle, with bait out of reach behind it. Animals walking through picked up ink and left prints on the paper on either side.
• Reference collection (2000 to 2002). Animals live-captured in Cerrado habitats were identified and encouraged to walk through a tunnel at least three times. For each animal they chose the three best front and three best hind prints, scanned them, and marked the centre of every toe pad and sole pad.
• Measuring prints. They measured a set of distances between the pads, built first around the sole pads (the rigid part of the foot) and then the toes (the mobile part), and used discriminant function analysis (DFA) to tell species apart.
• Identifying unknown prints. All prints of one species in one tunnel counted as one record. A record was only accepted when every replicate was classified to the same species with high probability and fell inside that species’ cloud of reference prints, to avoid misidentifying rare species missing from the collection.
• Inventories. In a gallery forest near Brasília, they set 36 tunnels in August 2000 (half on the ground, half on vines or branches 1 to 2 m up). In August 2001 they ran two transects, one with 60 Sherman live traps and one with 60 tunnels, and swapped them after four nights.

2. What they found

The tunnels

• They recorded animals from about 25 g (Gracilinanus agilis) to about 1000 g (Didelphis albiventris), arboreal to semi-fossorial, on the ground and in the understory, in dry and rainy periods.
• Baited tunnels were so attractive that one tunnel could hold records of up to four species after four nights, including trap-shy Monodelphis.
• Some animals left too many prints. Gracilinanus agilis averaged 110 prints per tunnel, other groups 16 to 28.
• Each tunnel cost US$1.20 and took about one minute to check.

The reference collection and identification

• 1408 reference prints from 251 individuals of 30 species (21 rodents and 9 marsupials).
• Prints of Monodelphis, Rattus, Proechimys and Nectomys were distinctive enough to identify by eye against the collection.
• DFA correctly classified 93.8% of front and 97.3% of hind prints of marsupials, and 74.0% of front and 88.3% of hind prints of rodents. Correct classification per genus ranged from 67 to 100% for rodents and 88 to 100% for marsupials.
• Hind prints separated species better than front prints.
• With fewer species in each analysis, results improved: hind prints separated species within the same genus 97.3 to 100% of the time (for example 97.9% for three Oryzomys species).
• Species separated along two main lines: body size, and a gradient from terrestrial through scansorial (climbing at times) to arboreal. Terrestrial rodents had longer, narrower hind soles than arboreal ones, and arboreal and scansorial marsupials had a clearly opposable hind toe.
• Separation was hardest among the many rodents weighing 30 to 60 g.

Tunnels versus traps

• August 2000: 65 track records of at least seven species, a record success rate of 25.8%.
• August 2001: 149 track records against 68 captures, of at least eight species. Track success was 31.0% against 14.1% trap success.
• Captures and track records of each species were not significantly correlated, suggesting species responded differently to traps and tunnels. Monodelphis and Oryzomys megacephalus were detected only in tunnels; no species was detected only by trapping.
• Rain splashing just outside the tunnels blurred 14% of track records. Fitting downward-angled PVC elbows to the tunnel ends solved this in later grassland surveys.

3. What they concluded

• Tunnels can sample small mammals in different habitats, vegetation layers and locomotion types, and footprint measurements analysed together can identify species, even closely related species living side by side.
• Tunnels were a more efficient inventory method than trapping and gave a good picture of the gallery forest small mammal community, including trap-shy species.
• The main limitation is the need for a good local reference collection, including rare species, built from well-identified live animals. Because prints are permanent, the collection can be built or added to during and after a survey.
• Individuals cannot be told apart by their prints unless they are toe-clipped, so density, home range and survival need traps as well.
• Tunnels suit studies needing presence or absence data or an activity index (inventories, distribution, daily activity, habitat selection), and could be used to study variation by age, sex and region. They are simple, cheap and give a permanent record, and could complement trapping.

4. How tracking was or was not used

The whole study is about identifying small mammals from their footprints. Prints were collected on paper in ink tunnels, not found in the wild. Identification relied mainly on measuring distances between toe and sole pads and statistical classification against a reference collection of prints from known animals, with some distinctive genera identified by direct comparison. No trailing was involved.

5. Tracker or observer certification

Not reported. The authors identified the prints themselves; most identifications rested on statistical classification rather than an observer’s judgement, but nothing is said about anyone’s tracking experience.

6. Other technologies, skills or innovations

• Ink tracking tunnels made from PVC pipe in three sizes, matched to each target species’ step length; medium and small tunnels cut in two halves with a moulded flange so ink and bait could be recharged.
• Blue stamp ink that gave clear prints for at least 16 days in the laboratory and seven days in the field.
• Prints scanned at 400 dots per inch, left prints flipped to match right ones, and pad centres digitised with Carta Linx software.
• The Truss method, a minimal set of distances linking the pads, adapted to feet.

7. Implications for conservation (from the paper)

• Tunnels avoid stress to animals and avoid direct contact with species that can carry diseases such as hantaviruses and American trypanosomiasis.
• A researcher does not need to be in the field every day, unlike trapping.
• Tracks that fit no species in a good reference collection may point to a rare species, or one not known from the region.