The bongo (Tragelaphus eurycerus) is the largest forest antelope found in equatorial Africa. With imposing horns on both males and females, its distinct large size and striking reddish-brown coat, with thick white stripes, makes the species an appealing symbol for forest conservation ( Figure 1 ). For the same reasons, the species has also been prized by trophy hunters (Koopmans et al., 2021). There are two existing bongo subspecies: the lowland or western bongo (T. e. eurycerus), which is found across Central and West Africa (Guinea, Togo, Cameroon, and South Sudan and the Republic of Congo (Koopmans et al., 2021), and the mountain or eastern bongo (T. e. isaaci), which is endemic to Kenya.

Male and female mountain bongo have a unique set of 8 to 12 white stripes on each side of the body (S. Njuki, Mount Kenya Wildlife Conservancy, personal communication, July 18, 2022), a nose chevron, enormous ears, and spiraled horns with one twist growing up to 75 cm to 99 cm (Estes, 1999). The horns of females are thinner and more parallel, whereas the horns of males are massive. These stocky forest antelopes are deep russet red in color, with males darkening as they age until nearly black (Kingdon, 2015) ( Figure 1 ).

Primarily browsing on low-level shrubs, young trees, herbs and vines, mountain bongo habitat is found in Kenya’s highland forests, with the species preferring to range in slopes between 2000 m and 3000 m (Kingdon, 2015). Bongo are known to be inactive during the day, typically resting in thickets. They move and feed at night, moving through dense cover with horns laid back (Estes, 1999; Kingdon, 2015; Koopmans et al., 2021). Historically, healthy mountain bongo populations were found in Kenya’s mountain forests including the Cherangani Hills, Mt. Londiani, Chepalungu Hills, the Mau Complex of forests including the Mau Eburu Forest, the Aberdare ranges, and the Mt. Kenya National Forest Reserve (Gibbon et al., 2015; IUCN SSC Antelope Specialist Group, 2017; Kenya Wildlife Service, 2019) ( Figure 2 ).

Mountain bongo were first classified as critically endangered in 2008 and updated in 2016 (IUCN SSC Antelope Specialist Group, 2017). At the time of the 2016 IUCN assessment, bongo were considered to survive in four unconnected, isolated montane forests (Elkan and Smith, 2013). Though accurate population estimates of this cryptic, mostly nocturnal species inhabiting dense, montane forest have proved challenging, attempts have been made using camera trapping technology.

Since 2006, estimates have been derived from camera trapping activities of the Bongo Surveillance Project (BSP) whose cameras were installed in mountain forest locations containing secondary mountain bongo signs (Prettejohn et al., 2020). In 2015, the BSP estimated that the mountain bongo population had declined to around one hundred individuals (Shears, 2015). In 2017, this figure was re-forecasted downwards, deemed to be as low as 73 individuals (Svengren et al., 2017). By 2018 however, despite one extensive camera trapping study at Eburu Forest for example, only five bongo events were observed from the 182,781 wildlife images analyzed across a 21-month period suggesting the survival of just one or two individuals in that ecosystem (Sheppard et al., 2022). By 2019, the populations in Eburu Forest and Mt. Kenya Forest were deemed functionally extinct (Prettejohn et al., 2020; Sheppard et al., 2022), with survival hopes pinned on viable populations within the Salient section of the Aberdare National Park, and the Maasai Mau section of the Mau Forest Complex (Prettejohn et al., 2020).

The remaining wild population is joined by captive-bred populations found in Kenya at Mt. Kenya Wildlife Conservancy (MKWC), and further afield in the Americas, Europe, and Australasia. In 2017, MKWC bordering the Mt. Kenya National Forest Reserve, held 63 mountain bongo, including 25 bulls and 38 cows (Svengren et al., 2017). Five years on from that, in 2022, the population had grown to as estimated 70 individuals (S. Njuki, Mount Kenya Wildlife Conservancy, personal communication, July 18, 2022).

Kenya’s five main mountain forests – the Mau Forest complex (including Eburu Forest), the Aberdare Mountains, Mount Kenya, the Cherangani Hills, and Mount Elgon are referred to as Kenya’s water towers as they capture most of the country’s rains (Pearce, 2015; Figure 2 ). Kenya’s montane forests store and release water, ensuring the year-round flow of major rivers, including the 1000 km long Tana River. They substantially contribute to the economy, supplying more than 75% of the renewable surface water resources of Kenya (United Nations Environmental Programme (UNEP), 2012). As Mount Elgon is the only forest among Kenya’s water towers that has neither historically nor currently supported mountain bongo populations, it will not be specifically discussed.

The climate of these montane forests, occupying elevations above 1500 m AMSL and into moorland and subalpine habitat in some cases, is wetter than the surrounding lowlands. Nighttime temperatures fall below 10°C in the cold season (June through August) and daytime temperatures rise above 30°C during the warm season (January through March) (Martin and Burgess, 2022). Frosts are possible at higher elevations, and freezing rain occurs.

Within these water towers, various vegetation zones can be distinguished including the closed-canopy forest, the bamboo zone, the sub-alpine and alpine vegetation. Within the forest belt, at the lower attitudes (between 1500 m and 3350 m), common hardwood tree species include camphor (Ocotea usambarensis), cedar (Juniperus procera), and podo (Podocarpus milanjianus) (Lambrechts et al., 2003; Luke and Beentje, 2020). These species give way to mountain bamboo (Oldeania alpina; also known as Arundinaria alpina, or four other synonyms) (POWO, 2024) and trees such as East African rosewood (Hygenia abyssinica in the higher attitudes up to 3,600 m (Luke and Beentje, 2020).

Kenyan mountain forests offer shelter to faunal species uniquely adapted to high elevation forest ecosystems. In addition to the endemic, critically endangered mountain bongo, Kenya’s mountain forests support the Jackson’s mongoose (Bdeogale jacksoni) (Lambrechts et al., 2003), and many amphibians and birds are also endemic to the montane areas (BirdLife International, 2022). Other large, threatened mammals of international conservation interest include the giant forest hog (Hylochoerus meinertzhageni), black rhinoceros (D. bicornis), forest elephant (L. africana), and leopard (P. pardus) (Lambrechts et al., 2003). The presence of mountain bongo in these forests could be viewed as an indicator of a healthy mountain ecosystem, suggesting not only that the faunal assemblage is complete but also may play a role in balancing the abundance or distribution of other species in the ecosystem.

Kenya’s montane forests face decline through regularly occurring human activities such as deforestation, cattle grazing, infrastructural development (Ayuyo and Sweta, 2014), selective logging, charcoal burning, Cannabis sativa fields, landslides, quarries, and human encroachment (Lambrechts et al., 2005), and, in some cases, sand harvesting (Olang and Kundu, 2011). Kenya mountain forests with current and/or historical mountain bongo populations include Mau Forest Complex including Eburu Forest, the Aberdare ecosystem, Mt. Kenya National Park and National Reserve, and the Cherangani Hills Forest ( Figure 2 ).

Historical mountain bongo hunters, known through mountain forest conservation networks in Kenya, were invited to share their knowledge on successful bongo hunting and live-trapping practices, and the behavioral ecology of the species. The selection of respondents was based on two main criteria. First, individuals with confirmed knowledge of the mountain bongo and former hunting practices were identified through snowball sampling. This technique, useful when potential participants are hard to find, involves research participants recruiting others for a study, allowing researchers to use their judgement in choose participants with specific expertise (Atkinson and Flint, 2001; Nikolopoulou, 2022). Second, individuals from multiple mountain ecosystems where the mountain bongo historically existed were selected. This approach provided diverse perspectives and allowed for comparisons across different ecosystems, enhancing the reliability and depth of the information gathered.

Interviews were conducted in the respondent’s language of choice. Interviews conducted in languages other than English were interpreted with the use of one expert translator, who was known to the interviewer. A total of eight men – between the ages of 60 and 90 years – were interviewed with the use of a questionnaire, and all responses written down in English. As documented in other exploratory, qualitative studies, small, purposeful samples can provide valuable insights on targeted themes (Rust et al., 2017; Poor et al., 2021). Each interview took two to five hours to complete. All respondents signed a written consent, adapted from the ethics protocol of the University of Guelph, Ontario, Canada, providing their information-sharing permission, prior to the commencement of the interview.

The interviews used a pre-determined list of questions and themes; however, by using the semi-structured approach, each interviewee had the flexibility to respond more broadly, and were invited to recount stories and memories spontaneously. Semi-structured interviews provide relevancy to the topic while remaining responsive to the participant (Bartholomew et al., 2000). The interviewer was able to learn more about each respondent’s specific information and to probe further for answers. By having a consistent structure within each questionnaire, coupled with an open-ended questioning format, respondents were able to provide responses rich with additional information.

Responses provided by the eight former hunters led to additional data collection following the interviews. Bongo location site names and details were documented where possible (further detailed in Section 2.4.2), as were details on observed bongo browse. With the help of local guides, browse species named by interviewees in the local languages of Ogiek, Kiswahili and Kikuyu were later identified and photographed. Samples were collected and stored in a plant press for further identification by a trained Kenyan botanist, Francis Norman Muruga Gachathi, Kenya Forestry Research Institute (KEFRI).

Interviews completed with eight former bongo hunters resulted in the provision of information from four different mountain forest ecosystems including four accounts from the Aberdare, three from the Mau Complex, and two each from Eburu and Mt. Kenya ( Figure 3 ).

Respondents reported on experiences over a roughly 40-year timeframe, from the early 1950s to 1992, including both legal and illegal hunting experiences. Prior to the nation-wide hunting ban in Kenya in 1977, under Legal Notice 120, the Wildlife (Conservation and Management) (Prohibition on Hunting of Game Animals) Regulations 1977 (Attorney-General, Republic of Kenya, 2012), hunting was legal and widespread. During this time, trophy hunting safaris were a leading form of tourism. Six of the eight former bongo hunters were involved with this side of the industry; two of the six also hunted for meat using local methods during this time. With the Government of Kenya’s policy making hunting illegal in the 1970s, the other two key respondent interviews were done with former poachers, active during the years after the wildlife hunting ban came into effect.

Respondents possessed a diverse hunting experience, with one, for example, retaining hunting licenses for seven African countries, and active across 50 years of trophy hunting in East and Central Africa, while another played the role of scout in advance hunting parties in the South West Mau, and never shot bongo. Combined, the total years of hunting experience documented was at least 120 years for the eight men.

General consensus among the eight historical bongo hunters indicates that mountain bongo live in single male, multi-female social groups in which the male stays slightly separated from the main herd (mentioned by 7 of the 8 hunters) (see Supplementary Table S1 for a summary of bongo behavioral ecology information shared by bongo hunters). Females were known to separate from the herd during estrous periods to mate with a breeding bull for three to four days (4/8). Females also separate from the herd when giving birth, returning several days later with the calf (2/8). Old males were chased out of the herd, and other out-group males could be found alone or in pairs (5/8).

The earliest memories from the eight respondents took place in the 1950s, when herd sizes were reportedly observed up to 50 strong, particularly in the Aberdare and Mau ecosystems. During those times, large bongo herds of 20 to 30 were common in the Aberdare, Mau, and Mt. Kenya forests, interspersed with smaller herds of 4–7 individuals. In Eburu Forest, group sizes were not as big, with hunters reportedly tracking groups of 4–6 individuals at most.

Over the following decades, single male, multi-female social structures remained intact and home range sizes remained relatively consistent. However, the herd sizes gradually declined over time. By the 1960s and 1970s, observed group sizes dropped to between 10 and 20 individuals, and by the present day, remnant populations are known to be surviving in narrowly defined locations within wider ecosystems. Figure 4 offers a change-over-time diagram of declining bongo populations in two documented ecosystems: the South Western Mau and the southeastern slopes of the Aberdare. The columns illustrate the observed typical herd composition and size over the various time windows from the oral reports. The main herds, comprised of females and calves, declined over time through over-hunting, while the single breeding males continued to be represented in the declining populations, and continued to be observed on the periphery of the main herds of females and calves.

Respondents suggested that bongo densities were ecosystem dependent. Predictions surmised that in the 1970s, when bongo populations were still healthy, the eastern Aberdares and the Mau Forest comprised densities in the hundreds, or densities that were too many to estimate (5/8). The three hunters who did not report on bongo density knowledge across different ecosystems were limited in exposure to a singular ecosystem and lacked travel experience. Figure 5 illustrates historical mountain bongo ranges across Kenya, as reported by bongo hunters.

Bongo populations were not found across the entirety of these mountain forests but favored specific forest sections with bongo-preferred features. Respondents indicated that bongo were found in the same habitat as giant forest hog (H. meinertzhageni) (7/8), with bongo reportedly eating similar foods to giant forest hog but at different heights; in the vicinity of salt licks or sources of salty water (6/8); and in the same forest zone where mountain bamboo was dominant (5/8), between 2,150 and 3,300 m (Luke and Beentje, 2020).

Prior to the 1950s, mountain bongo stretched from Limuru (50 km from Nairobi) in the south to the Cherangani Hills in the north (MPJ). In the 1950s, within the Aberdare, bongo were primarily found along the eastern flanks, with large populations in the northeastern and southeastern Aberdare (Maganjo, Thuti, and North Mathioya), and a third population in the Salient. Bongo were found on these eastern slopes throughout the year. Seasonally, small bongo populations were detected in the western Aberdare below the moorland at Chania River/Gura Falls (PMK).

At Mt. Kenya, bongo were only found in the southwestern slopes from Ragati to Embu. One hunter reported tracking bongo regularly in the Biruiru Hills of Ragati, where each herd established a home range around a particular salt lick (MM). Another hunter, based at Embu in the 1950s, witnessed a bongo coming down to the river to drink right beside their camp (TSS). A third respondent went on hunting expeditions in the 1960s at Meru but learned that bongo were not known there (KS).

Formerly the entire Mau Complex was thought to house mountain bongo, and the population so ubiquitous that it was impossible to estimate size (TSS). In the South Western Mau Forest, bongo were found primarily from Bosta to Kabbilat on one side, and from Emitik and Chemosit on the other. By the 1970s however, respondents claimed they could go a year without seeing a live bongo, only locating dead, snared individuals at salt licks (KS).

Within living memory, Eburu Forest was not thought to have a large bongo population. Total number was estimated to be between 20 and 30 during the 1960s and 1970s (PMK; SMN; TSS) and dropping to less than 15 individuals by the 1990s (SMN).

Respondents uniformly reported on known features of bongo activity patterns, with consensus scored on features including nocturnality, with a heightened period of activity at dawn and dusk (8/8); relatively small home ranges, of 1 km to 12 km suggested (8/8), and staying close to their food patches; the habit of spending several days in a single area feeding on a preferred browse species (6/8); and the behavior of following the same predictable trails and tracks (4/8). Similar to elephants (L. africana), bongo were observed to repeatedly follow the same trails through the forest.

Though differences in features of bongo ranging patterns were observed between rainy and dry seasons, these behaviors were inconsistently reported by former bongo hunters. Some felt that bongo ranged more during the dry seasons (2/8), and some felt that ranging increased during the rainy seasons (2/8). Others reported that food choices differed with the changing seasons (3/8), such as, ‘bongo moved with the season, chasing the food they eat’ (PMK).

Bongo were believed to have no competitors for browse. Though bushbuck (Tragelaphus scriptus) and giant forest hog (H. meinertzhageni) were observed feeding from the same climbers and other plants, they would be using resources at different levels. Bongo are thought to have more limited diets and spend longer periods feeding from one source than the bushbuck or the giant forest hog.

Predation, beyond human hunting activities, was thought to be rare, but it was witnessed. Leopard (P. pardus) and large groups of spotted hyena (Crocuta crocuta) opportunistically snatch youngsters. In the 1960s and 70s, it was still possible to see larger-sized bongo cows up in trees after being killed and dragged up there by the oversized leopard of the Mau Forest (TSS), and African crowned eagle (Stephanoaetus coronatus) have been known to collect days-old baby calves (MG).

Prior to 1977, when wildlife hunting was legal and widespread in Kenya, mountain bongo, like other prized faunal trophies, were on a quota for licensed hunters only. For example, in 1957 professional hunters were able to purchase a license to kill one male mountain bongo per year for 20 Kenya shillings ( Supplementary Figure S2 ; TSS). To do this, foreign hunters hired the services of local guides and trackers.

When working with international clients, Kenyan trackers developed techniques to improve bongo hunting success by going out in advance parties in preparation for the arrival of international trophy hunters ( Table 1 ). They would start at a naturally occurring salt or mineral lick (7/8), close to water sources (4/8) within the mountain bamboo (Oldeania alpina) vegetation layer, and follow the largest tracks (4/8), assuming that these tracks were made by males. As trophy hunters only shot the largest males, the fact that male bongo had a habit of separating themselves from the rest of the herd made them prime hunting targets. Reportedly, hunters would also seek out signs of giant forest hog (2/8), or locate a favorite bongo browse species, such as the locally named ‘setiot’ or ‘thunguya’ (7/8). Table 2 provides a list of species routinely named by hunters as important bongo browse and photos of those browse species are presented in Supplementary Figure S3 .

Bongo were difficult to locate without the help of dogs, and therefore these advance parties would use dogs to ease the work. Even if hunters did not use dogs, everyone reported that hunting with dogs was the easiest way to locate bongo (8/8). Those who did not use dogs at all experienced diminished success. For example, in 1960 there were 30 hunting licenses issued and just two bongo shot, and in 1961 there were 41 licenses taken out with five mountain bongo shot (Anderson, 1963). Unlike other antelope species, when being chased by dogs, bongo would eventually turn to face the dog pack and aggressively confront them. While the trophy hunters were there for the impressive horns of the biggest breeding bulls, illegal hunters were eating the meat with their families, and selling the magnificent, and much-sought, skins to curios traders in Nairobi (MG). Instead of guns, poachers used dogs and spears.

None of the eight men interviewed engaged in nighttime hunting, when bongo are active. Hunting was only done at dawn and dusk. Bongo rest during the day and are known to be excellent at hiding in dense forest. One prolific hunter claimed that in all his 50 years in the bush, he had never seen a bongo in the day (PMK), and could only find them with dogs. When asked, former hunters were able to report on the total number of bongo they had killed. One prolific poaching team, active in the Eastern Aberdares until 1992, killed 27 bongo over 23 years of hunting. The majority of hunters reported on killing far fewer bongo, ranging between 3 and 12 individuals across a span of 5 to 20 years (see Table 1 for a complete listing).

After the 1977 hunting ban in Kenya, the mountain bongo became a target for live export to international zoos, primarily through the Mount Kenya Game Ranch (now Mt. Kenya Wildlife Conservancy, MKWC) (Powers, 2010). Pitfall traps, bamboo fencing and net traps were three of the live-trapping methods employed for bongo, with pitfall traps proving most effective; detailed descriptions and diagrams of these methods and their success rates can be found in Data Sheet 1 .

The eight former hunters who lived in diverse parts of Kenya and once hunted in various forest ecosystems, were largely in agreement over a number of features of mountain bongo behavior. Key highlights include the fact that mountain bongo ranged over small areas (1–12 km); repeatedly used the same game trails; were nocturnal and crepuscular; demonstrated a preference for naturally occurring salt licks and ranged within the elevation layer dominated by mountain bamboo (2,150 m to 3,300 m). Furthermore, bongo spent multiple days repeatedly feeding on preferred browse in single male, multi-female social groups with the male preferring to separate slightly from the herd. Although small herds of four to six individuals were observed, mountain bongo of the 1950s and 1960s were also comfortable existing in much larger herds. Group sightings of 20–50 animals were once a common occurrence. These massive herds did not range throughout the Aberdare, the Mau Complex nor Mt. Kenya forests, but were existing in sub-sections of those larger ecosystems including, for example, primarily the eastern/southeastern Aberdares, and the south/southwestern Mt. Kenya forests.

Also emerging from the data is evidence of the gradual decline into endangerment of a species left unchecked over a 50-year period. Key influences across this period included a major effort to harvest live bongo from the Aberdare for export to international zoos in the 1970s; the heavy price paid for over a decade of unregulated and illegal shooting in the 1970s and 1980s when the elephant population was decimated from 165,000 to 15,000 (TSS); and the fact that poachers knew the bongo behavior well, aware that animals resting quietly in the dense undergrowth were difficult to locate during daylight hours but easy to find with hunting dogs.

TEK/LEK as a knowledge base has been criticized for its lack of testability (Usher, 2000). However, the repetition of similar details provided across informants in this study adds strength to the individual claims, increasing confidence levels in the most consistent facts provided. It is clear from the duplication of responses, that historical bongo hunter ecological knowledge provides a useful source of information about wild bongo behavior and offers clues to the decline of the species over time. It is recommended therefore, in this specific case with mountain bongo in Kenya, that knowledge shared by these bongo experts be applied to the design of any future captive-bred mountain bongo re-introduction plans.

The value to bongo recovery of the information shared in this study illustrates the potential application of TEK/LEK in situations where scientific inference is limited. When a species decline occurs before their behavioral ecology has been categorically documented, historical ecological knowledge offers a glimpse into a world that is no longer accessible, and one that can inform decisions for the present and future survival of a species. Furthermore, along with countering knowledge gaps in animal behavior, population size and ranging patterns, TEK, LEK and historic observations can, in part, surmount a shifting baseline in what we know about these factors.

Yet, TEK and LEK appear to be environmental decision-making considerations not widely utilized. At the federal level in Canada for example, TEK is officially recognized as a legitimate information source for wildlife management legislation, yet it is not widely used by administrators when writing regulations (Beaulieu-Guay, 2022).

Integrating different types of knowledge in environmental management projects has reportedly been inherently complex (Raymond et al., 2010), even though Indigenous knowledge systems are considered to be essential for sustainable development and environmental management (Athayde et al., 2017). A developing trend in current studies is hybridization, where traditional knowledge, practices, and beliefs are merged with novel forms of knowledge and technologies to create new knowledge systems (Gomez-Baggethun et al., 2013).

The challenges with TEK/LEK usage appear considerable, and include communication barriers, arising from the use of different languages; conceptual barriers, stemming from differing worldviews and relationships with nature; and political barriers, resulting from an unwillingness to acknowledge TEK that may conflict with industry or government agendas (Ellis, 2005). Others have noted that there is a lack of appropriate methods for organizing and presenting it for assessment and management purposes (Usher, 2000).

While it may be true that interviewing people successfully can prove difficult – especially with interviewees using different languages from one’s own, or coming from challenging sub-groups of society such as the elderly, researchers who hope to work with traditional keepers of ecological knowledge must shed the perception of intellectual superiority and seek ways to collaborate for the benefit of everyone involved. Allotting sufficient time and personnel for meetings and interviews can go a long way in addressing these challenges. Allowing the conversation to be structured in a way best suited to the respondent will also help keep thoughts and memories clear. Respondents may have challenges with hearing and seeing properly and are not always able to grasp the intentions of the interviewer. Semi-structured interviewing enables the tailoring of questions for each respondent’s understanding (Berg, 2008). Also, where possible, return meetings should be arranged to limit the length of time required at any given session. Confidence in the information gathered can be buttressed through cross-referencing measures such as interviews with additional informants and/or the completion of ground truthing surveys, and other field data collection methods.

There are social processes underpinning different perspectives on nature, and these value systems shift over time leading to tensions within a society over natural resource management, usage, and markets. These differing perspectives come about when the meaning of items and subjects change over time. Perspectives on trophy hunting have shifted considerably in the last century. Big game hunters, who once captured the imagination of many with their African hunting exploits, have been replaced by tourists shooting photos instead. In modern times, there is a great debate over the role and value of trophy hunting and hunters in conservation (Sas-Rolfes, 2017). Yet, in a recent systemic review of 49 studies interweaving TEK into quantitative wildlife analysis, the most common descriptor of knowledge holders was hunters/trappers (Stern and Humphries, 2022).

For the most part, conservation challenges are complex in character and there is a need to widen perspectives to embrace an extensive variety of collaborators, both conventional and unconventional. It is sometimes in the unexpected that the solutions are found. Reformed poachers, for example, make exceptional camera trapping team members (Sheppard et al., 2022). Having spent countless hours traveling in it, they know the terrain and the routes the animals take, and therefore have insight into the prime locations for camera traps (pers. obs.). By incorporating a multitude of perspectives, it may be possible to better tackle some of the conservation challenges plaguing the planet. Though the inclusion and integration of local, expert, or hunter knowledge about wildlife populations and their habitats can be problematic, it is possible to co-produce research to meaningfully include TEK/LEK into analysis (Castello, 2004; Berkes, 2012). Moreover, when ecological information about critically endangered species is difficult to acquire, traditional and historical knowledge holders provide a reservoir of experiential knowledge and a vital source of additional data.