Sampling scheme and protocols
Elements of the trapping procedures followed here have already been
partly described in Dalecky et al. (2015), Diagne et al. (2021), and
Granjon et al. (2021). The live traps used were of two types: locally
made wire-mesh live traps (8.5 × 8.5 × 26.5 cm) and Sherman [H.B.
Sherman Traps, Inc., Tallahassee, Florida, USA] folding box traps (8 ×
9 × 23 cm). Traps were set inside housing or working units (e.g.
dwelling houses, storehouses, shops, workshops) which potentially
included inner yards and associated parts (e.g. exterior staircases,
verandas). In each of these buildings, traps were set between one and
three consecutive nights in different “rooms”. Most of the time, a
trap of each type was placed in each room, on the floor, on furniture,
or even high up (wall tops, frame…). When the traps were
initially set up, each room sampled was georeferenced (geographic
coordinates GPS-recorded with an accuracy of +/- 5m). The rooms were
classified as belonging to eight “room types”, namely bedrooms,
granaries, food shops, kitchens, non-food stores, outdoors, stock rooms
and workshops. In the rooms, the presence or absence of food, and the
nature (materials) of the floor, walls and ceiling (see modalities in
Fig. 1 legend) was noted. This information represents markers of the
type of habitat (traditional vs. modern) in which the small
mammals studied live in contact with their human hosts. Traps were
checked and then baited once a day with peanut butter spread on a slice
of fresh onion.
Captured rodents were morphologically identified (following keys
provided in Granjon and Duplantier 2009), euthanized by cervical
dislocation as recommended by Mills et al. (1995), then weighed to the
nearest 0.5g, sexed and dissected. When necessary, molecular data were
generated to allow unambiguous species identification of rodents
(following procedures described by Lecompte et al., 2005; Dobigny et
al., 2011).