Comparative efficiency of three bait types for live trapping of small rodents in Central Africa

1 Département de Biologie, Université des Sciences et Techniques de Masuku (USTM), Franceville B.P. 901, Gabon. 2 EA 7462 GTUBE, Université de Rennes 1, 35042 Rennes cedex, France. 3 IGEPP, Université de Rennes 1, Avenue du Général Leclerc 35042 Rennes, France. 4 Department of Biological Sciences, University of New Orleans, New Orleans, Louisiana 70148, USA. 5 University of California-Riverside UCCE Riverside County, 21150 Box Springs Rd, Suite 202, Moreno Valley, CA, 92557-8718, USA . 6 Tulane University, 6823 St. Charles Avenue, New Orleans LA 70118, USA.

Although peanut butter has been used in several previous rodent surveys across the region, its efficacy relative to other traditional bait types has not been systematically compared.Bâ et al. (2012) used peanut butter as bait with a relatively high trap success (Ts) of 12.8%.Similarly, Webala et al. (2006) obtained a T S of 7.25% using a mixture of peanut butter, "maize meal", fried coconut and mashed ripe bananas, while Kok et al. (2013) obtained a Ts of 11.7% using a mix of oats and peanut butter.However, in another study by O'Brien et al. (2006), peanut butter mixed with oats yielded a relatively low trap success (Ts = 0.4%).The disparity between these results and the absence of a standardized experimental design for comparing bait types prompted us to explicitly test the efficiency of a mixture of peanut butter and corn flakes (PbCf) relative to manioc and palm nut in order to establish whether a more efficient sampling strategy could be used in future surveys.

Small rodent sampling strategy
All rodent live trapping was carried out using Sherman live traps (7.5 x 9 x 23 cm).At each site, four trap lines of 250 m in length were laid out at a distance of at least 25 m apart.Each trap line was made up of 50 Sherman live traps (7.5 x 9 x 23 cm) placed at regular intervals of 5 m along the length of each transect (Gautun et al., 1986;Nicolas and Colyn, 2003;Fichet-Calvet et al., 2009;Nicolas et al., 2010a;Denys et al., 2014).Traps were baited in equal proportions with either manioc, palm nut or PbCf.Trapping was conducted for a maximum of five consecutive nights at each site.
Traps were checked two times per day in the early morning and late afternoon and fresh bait was applied when necessary.At each site, up to 30 captured individuals were euthanized by cervical dislocation, measured, weighed and sexed for species identification and tissue samples were stored in ethanol 96% until DNA extraction.Any additional specimens were identified and released at their point of capture.All animal handling and processing was carried out according to UNO IACUC protocol 12-008.

Species identification
In order to avoid misidentification of specimens within the morphologically similar species complexes within the genera Hylomyscus and Praomys, species were identified based on external morphology (Kingdon, 2003;Musser and Carleton, 2005;Happold, 2013); and sequencing mitochondrial genes 16S rRNA and cytochrome b gene sequences were employed to clarify identifications where external morphology was ambiguous (Nicolas and Colyn, 2003;Nicolas et al., 2010b;Mboumba et al., 2011).

Statistical analysis
Trap success (Ts) = [(Nm/NTN) ×100] was estimated at each site; where Nm is the total number of individuals and NTN is the number of trap-nights (Nicolas and Colyn, 2003;Kadjo et al., 2013).Ts was compared among bait types using a chi-squared test, and P-values for all pairwise comparisons were adjusted using the False Discovery Rate correction (FDR; Benjamini and Hochberg, 1995).The number of individuals trapped by bait type was also analyzed separately for each species using an exact multinomial test.Pairwise comparisons between bait types were performed using exact binomial tests with FDR-corrected P-values.Statistical analyses were performed using the R software (R Core Team, 2014).

RESULTS
A total of seven species of rodents were captured all at night.These comprise: Hylomyscus alleni, H. walterverheyeni, Praomys misonnei, Nannomys minutoides, Lemniscomys striatus, Heimyscus fumosus and an unidentified Praomys sp.The latter two species were not trapped in sufficient numbers to allow for statistical analyses.
The global trap success observed in the present study was comparable to levels of success obtained in West African savannas by Bâ et al. (2012;Ts = 12%) and South African montane savannas (Kok et al., 2013; Ts = approx.12.8%), and superior to capture rates in the Ivory Coast (Traoré et al., 1980; Ts = 1.7% -2.43%)

DISCUSSION
Overall, our results show that PbCf bait is generally more effective than either palm nut or manioc.The global trap success with PbCf was far higher than that of previous studies in Central Africa using Sherman traps baited with either manioc or palm nut (Nicolas and Colyn, 2003).In particular, the T S for rodents caught in Moukalaba-Doudou National Park using PbCf was far better than the 0.4% obtained at the same site by O'Brien et al. (2006).Both studies used the same trap type but in the case of O"Brien et al. (2006) were baited with peanut butter and oats instead of peanut butter and corn flakes, which could be one of the factors explaining these differences in Ts.
Palm nuts had been used routinely in field sampling studies in CA because they are a local native forest product, easily available and inexpensive.Similarly, manioc is also readily available locally and is regularly cultivated in fields in savannas and forest clearings.However, palm nuts are only effective when fresh and decay rapidly after being harvested, posing a serious problem for long-term field sampling missions away from villages.Also, their fruiting period is seasonal, limiting their availability at certain times of the year.In contrast, PbCf is available at all times of the year, is nonperishable and has been shown here to be extremely effective relative to both local bait types.We therefore recommend using PbCf bait routinely as bait in field sampling studies of small rodents in CA since the cost of purchasing these products will rapidly be outweighed by the greater capture success.

Conclusion
This study shows that PbCf is a more effective at capturing small rodents than other more traditional baits.Specifically, the number of individuals and species captured was greater using PbCf than either manioc or palm nut.We therefore recommend its use in studies of small mammal biodiversity in the central African tropics as it both saves time and improves capture success, thus overcoming two, major constraints in studies of small rodent populations.

Figure 2 .
Figure 2. Total number of individuals captured per bait type, for each rodent species.Different letters indicate significantly different counts.

Table 1 .
Number of captures of each species by bait type and in each habitat type.