Conceived and designed the experiments: MA CB LV. Performed the experiments: MA GS EM. Analyzed the data: MA. Contributed reagents/materials/analysis tools: JH LIR CB MMR LV. Wrote the paper: MA LV. Collected samples, provided field information & field-based infrastructure: JH LIR CB MMR.
The authors have declared that no competing interests exist.
An assessment of population size and structure is an important first step in devising conservation and management plans for endangered species. Many threatened animals are elusive, rare and live in habitats that prohibit directly counting individuals. For example, a well-founded estimate of the number of great apes currently living in the wild is lacking. Developing methods to obtain accurate population estimates for these species is a priority for their conservation management. Genotyping non-invasively collected faecal samples is an effective way of evaluating a species' population size without disruption, and can also reveal details concerning population structure.
We opportunistically collected wild chimpanzee faecal samples for genetic capture-recapture analyses over a four-year period in a 132 km2 area of Loango National Park, Gabon. Of the 444 samples, 46% yielded sufficient quantities of DNA for genotyping analysis and the consequent identification of 121 individuals. Using genetic capture-recapture, we estimate that 283 chimpanzees (range: 208–316) inhabited the research area between February 2005 and July 2008. Since chimpanzee males are patrilocal and territorial, we genotyped samples from males using variable Y-chromosome microsatellite markers and could infer that seven chimpanzee groups are present in the area. Genetic information, in combination with field data, also suggested the occurrence of repeated cases of intergroup violence and a probable group extinction.
The poor amplification success rate resulted in a limited number of recaptures and hence only moderate precision (38%, measured as the entire width of the 95% confidence interval), but this was still similar to the best results obtained using intensive nest count surveys of apes (40% to 63%). Genetic capture-recapture methods applied to apes can provide a considerable amount of novel information on chimpanzee population size and structure with minimal disturbance to the animals and represent a powerful complement to traditional field-based methods.
Obtaining reliable estimates of a species' population size is an important component in determining its conservation status and provides a baseline for evaluating demographic change and/or conservation success over time. Rare and elusive species living in low visibility environments, like many of the world's threatened tropical animals, are usually impossible to count directly and difficult or expensive to detect by indirect methods
Great apes have been particularly difficult to survey due to their shy nature, low densities and occurrence in remote and inaccessible areas. Furthermore, ethical and practical concerns regarding trapping and collaring animals which are cognitively advanced, socially-bonded and susceptible to human disease has prevented the use of certain population estimation techniques, such as direct counts or capture-mark-recapture
Faeces, hair, feathers and other non-invasively collected materials are reliable sources of DNA and have allowed evolutionary and ecological processes to be inferred for elusive species
Currently, the vast majority of our knowledge on the behavior and ecology of chimpanzees comes from long-term studies on the eastern and western subspecies (
Non-invasive genetic sampling of apes offers a complement to traditional field-based approaches for understanding some aspects of wild chimpanzee society. Studies on kin relationships and patterns of relatedness within and between social groups
In this study, we aimed to estimate the number of chimpanzees and their distribution into groups in a 132 km2 area of Loango National Park, Gabon using the genetic capture-recapture method. To do so, we amplified 8 rapidly evolving, highly variable, autosomal microsatellite markers from central chimpanzee faecal samples collected opportunistically over a four-year period, which allowed us to reliably distinguish even closely related individuals. Because chimpanzees are male-philopatric with males remaining in their natal community for life, we also amplified 13 Y-chromosome microsatellite markers for all males. We hypothesized that the resulting paternally inherited haplotypes should be the same or similar within groups, while differing between groups, as has been previously observed in a study of multiple communities of eastern chimpanzees
Samples were collected across the Loango Ape Project research site, a 132 km2 area in the central sector of Loango National Park, Gabon
Between February 2005 and July 2008, two to four field teams conducting ape habituation and biomonitoring activities in the study area opportunistically and unsystematically collected up to three-day-old chimpanzee faecal samples; due to the presence of dung beetles, rain and maggots, ape faeces do not persist for more than three days at Loango. Faeces were preserved using the two-step ethanol-silica procedure
A total of 452 putative chimpanzee samples were collected from beneath night nests and from where chimpanzees had defecated as they moved through the forest during the day
Faecal samples were extracted from one month to one year after collection, using the QIAmp Stool kit (QIAGEN) with slight modifications
At least one sample from each male individual identified in the data set was further genotyped at the 13 Y-chromosome loci previously described in
Up to four different PCR products were combined and electrophoresed on an ABI PRISM 3100 Genetic Analyser and alleles were sized relative to an internal size standard (ROX labeled HD400) using GeneMapper Software version 3.7 (Applied Biosystems). Heterozygous genotypes were validated by observing each allele in two or more independent reactions and depending on the quantity of DNA in the extract, homozygous genotypes were confirmed in up to five independent PCR amplifications
We used CERVUS 3.0 to identify independent samples with matching autosomal genotypes. We estimated the minimum number of autosomal loci necessary to obtain a PIDsibs value of ≤0.001
The number, composition and minimum territory size of chimpanzee groups were determined using the following criteria (illustrated in
Each letter represents an individual's genotype and its sampling location over the course of the entire study period. Females are denoted by ♀, males are denoted by ♂. Underlined samples were found at the same location on the same day. (i) Relative geographic locations of samples from individuals A through F (ii) Assuming that individuals found together belong to the same group, here male B links together samples A and C, thus A,B & C are all members of a single community. (iii) If males B, E & D all carry the same Y-haplotype we assume they belong to the same community and draw a minimum convex polygon (MCP) around these individuals. As chimpanzees are territorial, we assume that females found within this MCP belong to the males' community. Thus, individuals, A, B, C, D & E all belong to the same community. (iv) Because female C was found within the Y-haplotype defined MCP at one collection event, we can extend the MCP to include any other sampling events of female C. By doing so, female F now also falls within the MCP of the group so that individuals A, B, C, D, E & F all belong to the same group. See text for exceptions to these rules.
Second, as chimpanzees are a patrilocal species we hypothesized that if males of each community carry a unique set of Y-chromosome haplotypes as suggested by previous research
Third, females and males found within a given Y-chromosome-delineated MCP community were attributed to that community, as we assumed no territory overlap. Individuals found in association with samples from different group affiliations over the study period could not be attributed to any community and were not used in MCP construction. When a female was found both within and outside of an MCP, the MCP was redrawn to include the exterior female data points. The area covered by the final MCP was considered the minimum territory size for that chimpanzee community.
If females did not fall into any MCP and were either found alone, with other females who were also not attributed to any group, or from a collection site where only the one sample contained usable DNA, then they could not be attributed to any group. Because a dead individual has ‘left’ the population this violates the assumption of closure in our population and we do not include individual C12 (known to have been killed in an intercommunity attack in August 2005
Grouping all samples into a single-sampling session scheme and using individual genotypes that were identified from one (initial capture) or more (recaptured) samples, we calculated a genetic capture-recapture estimates using the maximum likelihood two innate rates model (ML-TIRM) estimator implemented in the software Capwire (
Calculating a population estimate using samples collected over the entire four-year study period may violate the assumption of closure in our models. Thus, to compare inferences made over the entire study period with those from a more restricted time period (and consequently smaller spatial area), we calculated a population estimate using samples collected from February 2005 to June 2008 as well as a population estimate from samples collected from each year separately.
In total 202 chimpanzee samples yielded usable genotypes, resulting in a 46% (202/444) extraction success rate over the four year period. Extraction success was not obviously related to time of year (data not shown) and was consistently low every year, ranging from 41% (in 2007) to 63% (in 2005).
Genotypes from the 202 samples were on average 98.9% complete with 88.6% of extracts (179/202) genotyped at all eight loci and 9.9% genotyped at seven loci. After identifying matching genotypes from multiple samples and assigning consensus names to the matches, genotypes from the resulting 121 chimpanzees were on average 99.5% complete (
In all cases where two or more samples produced identical genotypes at seven or all eight loci, we obtained a PIDsibs value of ≤0.001, strongly suggesting that in these cases the samples did indeed come from the same individual and were not derived from full-siblings who happened to be identical at these loci. Two samples C74 and C120 produced confirmed genotypes at 6 of the 8 loci with only 1 allele confirmed at the other two loci. However, both of these samples mismatched all other samples at a minimum of 4 loci so we are quite certain that they represent unique individuals.
The rates of allelic dropout and the appearance of irreproducible, sporadic alleles were calculated and found to be on average 16% and 2% per PCR, respectively. Using the multiple tubes approach with DNA quantification we estimated the number of independent PCRs necessary to ensure with >99% certainty that homozygote genotypes are authentic and not the result of allelic dropout
Six of the 13 Y-chromosome microsatellite loci under investigation were polymorphic, although only two alleles were seen at each of these six loci (
Groups were first identified by visually evaluating whether Y-chromosome haplotypes of the 58 males (N = 92 observations) in the study area clustered geographically. Males carrying haplotypes E, F and H appeared to each cluster geographically to the exclusion of all other haplotypes (
Males are designated by their Y-chromosome haplotype (uppercase A-I). Females are designated by the lowercase letter(s) of the group in whose minimum convex polygon (MCP) they were found (a, bi, c, dg, e, f or h). Females that did not occur in any MCP or that were found in association with more than one group throughout the study are represented by black circles. In cases where females were found both within and outside of a given MCP, the MCP was enlarged to include the “exterior” geographic location of the female. MCPs represent minimum territory boundaries of each chimpanzee community. Area of MCPs stated in parentheses in legend. For group H (southern most points) only 2 individuals were identified and so no MCP could be drawn. Inset, map of Africa with Gabon highlighted in white, arrow indicates location of Loango field site.
The 20 males carrying haplotype B also clustered to the exclusion of all other male haplotypes except in one instance where the male with haplotype I (C134) was found with three males with haplotype B (C103, C105, C106) (
ID is consensus name given to matching chimpanzee genotypes. In row “sex”, M = male, F = female. Y-hap refers to the Y-chromosome haplotype of the male individual. Y-hap MCP is the group membership identity of each individual based on their exclusive inclusion in any of the Y-haplotype defined MCPs. B?/A? means individual may belong or have belonged to either group B or A over the study period. A?/C? individual may belong or have belonged to either group C or A. “?” indicates females could not be attributed to any group. Grey boxes bound first and last instance when group members detected over the 4-year study period.
Similarly, haplotypes D and G co-occur within a very small geographical space (
As shown in
The 47 females falling within these MCPs were considered to belong to their respective Y-haplotype defined groups. Four females (C14, C22, C29 and C62, see below) had ambiguous group affiliations and 13 females (listed at the end of
Behavioral and genetic evidence also suggest that groups A and C are distinct entities. In addition to the purported August 2005 killing of C-haplotype male C12 by group A males reported in
In sum, seven groups were identified (A, BI, C, DG, E, F and H), however only groups A, BI and C were detected on more than 10 occasions (89, 47 and 11 times, respectively), making inferences about minimum group size and territory size limited to these groups. Thus, minimum group size ranged from 7 to 47 individuals, and minimum territory size ranged from 2.6 km2 to 45.0 km2 (
Group | Minimum # individuals | Minimum # males | Minimum # females | Minimum territory size | # occasions group detected |
47 | 21 | 26 | 45.0 km2 | 89 | |
35 | 20 | 15 | 33.9 km2 | 47 | |
7 |
3 |
4 | 2.6 km2 | 11 |
|
4 | 3 | 1 | n.d, | 5 | |
4 | 3 | 1 | n.d. | 4 | |
4 | 4 | 0 | n.d, | 4 | |
2 | 2 | 0 | - | 2 | |
17 | - | 17 | - | 24 | |
2 | 2 | - | - | 4 | |
122 |
58 |
64 | - | 190 |
Group H was only detected twice, so no minimum territory size could be calculated. n.d. – not determined as groups were sampled fewer than 10 times.
*- an additional dead male (C12, from group C) was identified in a previous study and included in the totals presented.
Of the 202 chimpanzee samples from the study site yielding usable genotypes, there were 13 instances of samples collected at the same location and on the same day as the other samples representing the same individuals; these cases were collapsed into single captures. Of the 83 samples collected in 2007, 61 unique genotypes were identified. The number of faeces successfully genotyped per individual ranged from 1 to 6 (mean 1.56, SD 1.05) for the entire study period and from 1 to 5 (mean 1.36, SD 0.80) for the 2007 sampling period, with the majority of individuals sampled only once in either sampling schemes (entire study period: 84/121; 2007 sampling period: 47/61) (
Applying the ML-TIRM population estimator, resulted in a point estimate of 283 (CITIRM-AllData: 208–316) chimpanzees using the 132 km2 area over the entire study period and 176 (CITIRM-2007: 113–220) chimpanzees using a 73 km2 subset of the study area in 2007 alone. By adjusting for area sampled, we obtain similar density estimates from the entire data set and the 2007 data: 2.14 (CITIRM-AllData 1.58–2.39) chimpanzees/km2 and 2.41 (CITIRM-2007 1.55–3.01) chimpanzees/km2, respectively. The precision of the estimates, measured as the entire width of the 95% confidence interval divided by the estimate itself, was 38% and 61% of the point estimate for the entire data set and 2007 samples, respectively.
Of the 13 Y-chromosome loci genotyped, only six were variable, and then, only dimorphic. This low amount of Y-chromosome variation differs from the pattern observed in eastern chimpanzees in Kibale forest, Uganda
As compared to other chimpanzee subspecies, very little is known about social organization and grouping patterns in central chimpanzees
Several considerations suggest that the MCPs derived here may underestimate the territory sizes of the chimpanzee groups at Loango. MCPs can overestimate the territory size of species by including areas that are not used by the individuals
Most of group C's small territory is overlapped by group A's territory. The circumstantial evidence suggests that group A is expanding its territory and replacing group C by making incursions into group C's territory and killing group C males and infants. A similar pattern of group extinction was observed in two eastern chimpanzee populations. At Mahale, one group (M) was suspected of exterminating the males of another (K), co-opting many of the group females and expanding into its territory over a 12 year period
As an alternative explanation for the apparent overlap between groups A and C, it has been suggested that neighborhoods exist in some chimpanzee communities
We show in this study that despite reliance on opportunistically collected faecal samples with poor extraction success (46% on average), we can obtain useful population estimates, albeit with moderate 95% confidence intervals. As gorilla samples collected in the same manner from the same site over a similar time period had a higher success rate (82%,
Using the ML-TIRM method we obtain a population estimate of 283 chimpanzees (ranging from 208 to 316 individuals) using the Loango study area from 2005–2008. Chimpanzees are long-lived primates with slow life histories. Adult deaths and female dispersals are rare events and chimpanzee females give birth only once every 5 to 6 years
The density estimate of 2.14 chimpanzees/km2 (range: 1.58–2.39) is in the upper range of those previously reported for other central chimpanzee sites [0.03–2.78 chimpanzees/km2, 14]. It is important to note however that published chimpanzee density estimates are from nest surveys which have been shown to underestimate the density of chimpanzees by 70% or more when compared to estimates obtained from direct observations in eastern and central
Most individuals were only captured a single time, resulting in a population estimate with moderate precision. Consequently, the number of samples genotyped was smaller than the number of individuals estimated to live in the population for both sampling schemes. Previous studies have shown that genotyping at least twice as many samples as the number of individuals that exist in the study population dramatically decreases the width of the 95% confidence interval surrounding the obtained population estimate
Despite these limitations, the precision of the genetic estimate is comparable to that of traditional nest count estimates used to evaluate ape population size, for which 95% confidence interval widths (from lower to upper confidence bound) of 40% to 63% of the estimate are reported
We show here that genetic monitoring provides a useful and informative complement to field-based research. Although the 444 chimpanzees samples used here were collected opportunistically over a four-year period, had the focus of a team (or teams) been to search out and obtain faeces, the same number of samples could have been collected over a much shorter period of time. On the other hand, collecting the samples over time allowed us to monitor the movements of individuals. The cost of collection materials and laboratory materials for genetic monitoring are not prohibitive but neither are they trivial. A similar study to the one presented here (695 gorilla samples collected, 384 samples extracted, and 16 microsatellite markers amplified) within the context of a pre-existing field infrastructure, estimated additional laboratory costs for the analysis of the samples to be approximately 12,000 Euros (not including the cost of labor,
For chimpanzees specifically, their fission-fusion social system provides additional challenges to evaluating group composition as samples from individuals in the same social group cannot be linked together as easily as for gorillas
Genetic surveys can play an important role in assessing wild ape population dynamics when used in addition to traditional surveys, which provide a wealth of information on ape ecology and anthropogenic disturbances. Traditional transect based nest-count surveys can often give rapid assessments to conservation managers that is not possible with genetic-based methods. However, even though genetic surveys will increase the expense of a survey and require increased time for analysis, we demonstrate that the information gained from the additional time and expense is worthwhile, even with opportunistic sampling and a poor success rate. It is clear that opportunistic genetic sampling provides a wealth of information and is a valuable biomonitoring tool for elusive species and we highly recommend its inclusion in forest monitoring activities in the future.
Mismatch distributions for the Loango chimpanzee genotypes. The majority of individuals were compared at 8 autosomal loci, however a subset were also compared at 3 additional autosomal loci. Y-chromosome haplotypes were also compared for all the males (with the haplotype coded as a single “homozygous” locus). Values above columns represent number of dyads in each locus category.
(1.10 MB TIF)
Median-joining networks depicting the phylogenetic relationships of Y-chromosomal haplotypes for the Loango chimpanzees. Each circle represents one Y haplotype. Circle size is proportional to haplotype frequency, with the smallest circle representing a haplotype carried by one individual. * denotes haplotypes found in group BI, ̂ denotes haplotypes found in group DG.
(2.31 MB TIF)
Movements of individuals C14, C22 and C79 suspected of moving between groups C and A and location of suspected intergroup aggression (infanticides and adult male killing). In June 2006, after following chimpanzee vocalizations, we observed a group of eight chimpanzees that were displaying and vocalizing. Once the chimpanzees had dispersed from the site, bloodspots, chunks of flesh and an infant foot, were found. In June 2007, we observed several chimpanzees vocalizing with hair bristled and appearing distressed. We found fresh blood and bone at the contact site and upon following the group, one male was observed eating what appeared to be an infant chimpanzee. In both cases, diarrhea, a sign of stress, was present at the contact sites. Female C14, originally found in early 2005 with females otherwise associated with group C, was subsequently found in the center of group A's MCP in November 2006. Furthermore, female C22 was found in June 2005 in association with haplotype C male C32 and within the group C MCP in March 2005. She was then found just north (within 300 m) of the group C MCP on three later occasions (April 2006, June 2006, June 2008). In fact, her sample from June 2006 was the only successfully genotyped sample from the nine samples collected in the area of the infant killing described above. In June 2008, C22 was sampled with two males: haplotype C male C79 and haplotype A male C136. Male C79 was initially found in the center of the haplotype C MCP in April 2007 but then in the northeastern limit of group BI's MCP in December 2007 and finally with C136 as described above. Inset: Map of study site (
(1.17 MB TIF)
Minimum territory size of group A calculated using male chimpanzee samples with Y-haplotype A that were captured two or more times only. Males sampled more than once noted with their consensus ID. Inset: Map of study site (
(1.03 MB TIF)
Genotypes of 125 chimpanzees from Loango National Park, Gabon. * individual C12 is an adult male chimpanzee that was killed in the study area in an intercommunity attack in August of 2005 and genotyped in a previous study (Boesch et al 2007).
(0.04 MB DOC)
The nine Y-chromosome haplotypes (Y Hap) identified in the Loango Ape Project Study area. The 6 polymorphic loci are shaded in grey with the two alleles of the locus in black or white.
(0.03 MB DOC)
Primer sequences, annealing temperature, repeat type and allelic size ranges of Y-chromosomal microsatellite loci. Ta: annealing temperature for singleplex PCR. F: forward primer (the forward primer used in the multiplex and singleplex PCRs are identical except that the forward primer used in the singleplex PCR is fluorescently labeled with FAM, HEX or NED dyes). R: reverse primer. Rnest: reverse nested primer. §: only nested reverse primers were designed for this study, forward and reverse primers are published elsewhere (Erler et al. 2004; Gusmao et al. 2002a; Gusmao et al. 2002b).
(0.02 MB DOC)
We thank the Agence Nationale des Parcs Nationaux (ANPN) and the Centre National de la Recherche Scientifique et Technique (CENAREST) of Gabon for permission to conduct our research in Loango National Park. We also thank H. Kühl, L. Makaga, E. R. Guizzard, N. Tagg, B. Graw, E. Fairet, M. Gregoire, L. Rankin and the other field assistants of the Loango Ape project for their help collecting the samples. We are very grateful to L. Cowen for advice on capture-recapture methodologies, G. Campbell, H. Kühl, D. Morgan, C. Sanz, K. Langergraber, D. Caillaud, R. Mundry, A. Melnyk, R. Arandjelovic, D. Ward & J. Junker for helpful discussions and A. Abraham & C. Lang for laboratory assistance. The project is a collaboration among the Société pour la Conservation et le Développement (SCD), Wildlife Conservation Society (WCS), and the Max Planck Society.