Tracing the origins of rescued chimpanzees reveals widespread chimpanzee hunting in Cameroon

Abstract

Background: While wild chimpanzees are experiencing drastic population declines, their numbers at African rescue and rehabilitation projects are growing rapidly. Chimpanzees follow complex routes to these refuges; and their geographic origins are often unclear. Identifying areas where hunting occurs can help law enforcement authorities focus scarce resources for wildlife protection planning. Efficiently focusing these resources is particularly important in Cameroon because this country is a key transportation waypoint for international wildlife crime syndicates. Furthermore, Cameroon is home to two chimpanzee subspecies, which makes ascertaining the origins of these chimpanzees important for reintroduction planning and for scientific investigations involving these chimpanzees.

Results: We estimated geographic origins of 46 chimpanzees from the Limbe Wildlife Centre (LWC) in Cameroon. Using Bayesian approximation methods, we determined their origins using mtDNA sequences and microsatellite (STRP) genotypes compared to a spatial map of georeferenced chimpanzee samples from 10 locations spanning Cameroon and Nigeria. The LWC chimpanzees come from multiple regions of Cameroon or forested areas straddling the Cameroon-Nigeria border. The LWC chimpanzees were partitioned further as originating from one of three biogeographically important zones occurring in Cameroon, but we were unable to refine these origin estimates to more specific areas within these three zones.

Conclusions: Our findings suggest that chimpanzee hunting is widespread across Cameroon. Live animal smuggling appears to occur locally within Cameroon, despite the existence of local wildlife cartels that operate internationally. This pattern varies from the illegal wildlife trade patterns observed in other commercially valuable species, such as elephants, where specific populations are targeted for exploitation. A broader sample of rescued chimpanzees compared against a more comprehensive grid of georeferenced samples may reveal 'hotspots' of chimpanzee hunting and live animal transport routes in Cameroon. These results illustrate also that clarifying the origins of refuge chimpanzees is an important tool for designing reintroduction programs. Finally, chimpanzees at refuges are frequently used in scientific investigations, such as studies investigating the history of zoonotic diseases. Our results provide important new information for interpreting these studies within a precise geographical framework.

Figure 1

Chimpanzee subspecies and georeferenced DNA sampling distributions. A) Distribution of chimpanzee subspecies. Chimpanzees belong to a single species (Pan troglodytes) that is divided into four subspecies [1]. Phylogenetic analyses of mtDNA suggest that these subspecies are divided into two geographically and genetically defined groups that split about 0.5 mya: a western African group (P. t. verus and P. t. ellioti [22] [known until recently as P. t. vellerosus [23]) and a central/eastern African group (P. t. troglodytes and P. t. schweinfurthii) [24]. A phylogeographic break between these two groups occurs at the Sanaga River in central Cameroon, separating populations of P. t. ellioti north of the river from P. t. troglodytes south of the river. However, the Sanaga does not stop dispersal between subspecies completely because some gene flow between them occurs near the confluence of the Sanaga and its main tributary, the Mbam River [24,28]. B) Map of Cameroon and Nigeria showing collection sites of georeferenced chimpanzee DNA samples. Sampling sites shown on the map are: Ise Forest Reserve (ISFR), Cross River National Park (CRNP), Akoh Zanto (AKZN), Gashaka Gumti National Park (GGNP), Mount Cameroon (MTCM), Mosse (MSSE), Manb'ra (MANB), Douala-Edea Forest Reserve (DEFR), Campo-Ma'an National Park (CMNP) and Dja Biosphere Reserve (DJBR).

Figure 2

Median joining network of mtDNA HVRI sequences. This network is composed of 464 georeferenced chimpanzee samples spanning Cameroon and Nigeria that were reported in previous studies [22,23,26-28] and LWC chimpanzee samples (n = 46). Haplotypes were color coded denoting their region of origin. Samples shown in purple were collected in Nigeria and western Cameroon west and north of the Sanaga River. Samples shown in orange were collected in southern Cameroon south of the Sanaga River.

Figure 3

Representation of confidence in CAM assignments for georeferenced chimpanzee samples. The green circles represent 100 point estimates drawn randomly from the set of all possible locations weighted according to their probability. Stars represent the median point coordinate estimate for each georeferenced sample. The concentration of these 100 circles in any given area is a guide to the probability that the sample arose from that area, where tighter concentrations of circles indicate higher confidence in the median point estimate. The top panel shows a georeferenced sample (ISFR001) with estimated CAM origins near its actual location of origin, along with a high concentration of circles near its location origin. The bottom panel shows a georeferenced sample with estimated CAM origins also near its actual sampling location (MANB014), but with more dispersed coordinate point estimates.

Figure 4

Estimated CAM origins of 46 LWC chimpanzees. Samples were color coded to denote their mtDNA haplotype membership shown in Figure 2. Samples shown in purple clustered with mtDNA Haplotype 1, whereas those shown orange clustered with mtDNA Haplotype 2.

Figure 5

Representation of confidence in CAM assignments for LWC chimpanzee samples. The green circles represent 100 point estimates drawn randomly from the set of all possible locations weighted according to their probability. Stars represent the median point coordinate estimate for each sample, color coded according to each sample's mtDNA haplotype from Figure 2. The concentration of these 100 circles in any given area is a guide to the probability that the sample arose from that area, where tighter concentrations of circles indicate higher confidence in the median point estimate. The top panel shows LWC026 with a tight concentration of circles near its median point estimate. The bottom panel shows LWC040 with more dispersed coordinate point estimates.