Camera Traps Uncover the Behavioral Ecology of an Endemic, Cryptic Monkey Species in the Congo Basin

Cercopithecus lomamiensis (lesula) was described in 2012 and is endemic to the TL2 (Tshuapa–Lomami–Lualaba) landscape in the central Congo Basin. It is closely related to C. hamlyni, likely via an allopatric speciation event, and is listed as Vulnerable due to bushmeat hunting. Members of Cercopithecus are typically arboreal, but the hamlyni species-group is thought to be more terrestrial, raising questions about the evolution of terrestriality within guenons. The study aimed to fill gaps in ecological and behavioral knowledge for lesula using camera traps to: (1) quantify degree of terrestriality, (2) assess daily activity pattern (confirm diurnality vs possible crepuscular/cathemeral activity suggested by large orbits and pre-dawn calls), (3) investigate birth seasonality relative to rainfall, and (4) document group size and composition, thereby informing conservation planning for this cryptic species.

Background literature notes high guenon diversity in the TL2 region and understudy of core Congo Basin species. Prior genetic work identified lesula as sister to C. hamlyni with non-overlapping ranges. Morphological analyses suggest terrestrial adaptations in lesula’s postcranium despite the predominantly arboreal nature of Cercopithecus. Diurnality is typical for Old World monkeys, though larger orbits can be associated with low-light environments and some primates show dawn/dusk or nocturnal activity under certain ecological pressures. Guenons typically exhibit birth seasonality correlated with rainfall/food availability and live in one-male, multi-female groups with female philopatry and male dispersal. Initial field observations for lesula reported small apparent group sizes, possibly due to detection bias in dense forest.

Ethics: Research conducted under permission from ICCN; buffer zone surveys authorized locally. Study area: TL2 Landscape (∼40,000 km²) including Lomami National Park (8874 km²; tropical lowland forest with two wet seasons; near-equatorial 12:12 h photoperiod). Sites: Two within LNP (Losekola; E15) and one in buffer zone (Okulu, hunted area). Survey design: Three terrestrial, systematic camera-trap (CT) surveys over three years: Okulu (Oct–Dec 2013), Losekola (Jan–Nov 2014), E15 (Aug 2015–Jan 2016). Systematic grids in ArcGIS with sampling points spaced 500 m (Okulu, Losekola) and 1 km (E15). Cameras: 41 CTs at Okulu on 4.1 km² grid, 41 at Losekola on 3.8 km² grid, 20 at E15 on 12.0 km² grid; Moultrie and Bushnell units with IR sensors, no bait; positioned on trees facing animal trails; video mode, 1280×720 resolution; date/time stamped. Data processing: Sampling effort = total functional CT days; videos viewed (Okulu n=1885; Losekola n=9179; E15 n=3570). Two independent data-entry teams reconciled discrepancies (n=27). Event definition: cluster of chronological videos from same social group; new event if no lesula detected within 30 min. Capture rate = events / CT-days ×100. Analyses: (1) Degree of terrestriality by comparing lesula vs sympatric arboreal guenons’ capture rates on ground CTs and scoring an event as terrestrial if ≥1 video showed ≥1 animal with all four limbs on ground ≥1 s. (2) Activity pattern via hourly capture rates over 24 h. (3) Birth synchronicity by infant age classification (Infant 1 <1 mo; Infant 2 ≈3 mo; Infant 3 ≈6 mo before event) using physical traits; three observers, discard disagreements; estimate birth months; compare with monthly rainfall from Obenge rain gauge (2008–2013; mean annual precipitation 2021 mm; 145 rainy days). (4) Group size/composition by minimum number of unique individuals per event using age/sex traits, unique markings, positions, movement direction; group defined as ≥3 individuals with ≥1 female or juvenile; at least two observers per count; calculate mean group size across sites.

Sampling: 92 cameras deployed (data from 39 Okulu, 33 Losekola, 20 E15; 11 excluded for malfunctions). Total 5960 CT-days and 598 independent lesula events; capture rates: Okulu 9.22, Losekola 11.61, E15 8.74 events/100 CT-days (overall 10.03). Terrestriality: Lesula capture rates were 67–500× higher than arboreal sympatric guenons (C. ascanius, C. mitis, C. wolfi) on ground CTs. Lesula events met terrestrial criteria 98.8% of the time (591/598 events). Only 10 terrestrial events across arboreal species (C. mitis n=9; C. wolfi n=1); one C. ascanius event not terrestrial by definition. Activity: All lesula events occurred between 06:00–19:00, confirming strict diurnality; few at dawn (n=10; 06:00–07:00) and dusk (n=2; 18:00–19:00); consistent daytime activity for ∼11 h. Birth seasonality: 77 infant events with 104 infant sightings (12.9% of all events); used 60 events (80 sightings) with ≥2/3 observer agreement. Infant 1 (n=38) analyzed separately from Infant 2+3 (n=42). Peak births estimated August–October, coinciding with onset of the second wet season. Group size: Considering qualifying group events (≥3 individuals with ≥1 female/juvenile; 40% of events, n=238), mean group size was 7.17 ± 4.68. Singletons or pairs were 54% (n=323) of all events; only 15% (n=87) exceeded the mean. Largest observed counts: 32, 31, and 25 individuals (Losekola, E15). Group composition: Total individuals tallied across events = 2298; mature:immature ratio 1.67 (matures=1163; immatures=696); adult male:female ratio 1.05 (♂=147, ♀=140), though sex was unknown for 75% of matures (n=876). Unknown age/sex comprised 19% (n=439) and were excluded from ratios. Adult males present in 24% of group events (n=58), with nine events showing two adult males (mean group size in those = 14). Among 188 singleton events (excluding unknowns): 49% subadult/adult unknown sex (n=93), 32% adult males (n=61), 10% juveniles (n=19), 8% adult females (n=15) (χ²=45.32, df=3, p<0.001). Overall, 45% of identified males were lone or in male-only groups (n=66).

Camera trapping proved highly effective for detecting and studying the cryptic, threatened lesula across protected and hunted sites. Findings confirmed strong terrestriality, exclusive diurnal activity, a defined birth season aligning with the wet season, and social organization consistent with Cercopithecus: medium-to-large multi-female groups with typically one resident adult male (occasionally two), female philopatry, and male dispersal, including lone and bachelor males. Terrestriality aligns with morphological evidence and may reflect adaptive divergence to reduce competition with sympatric arboreal guenons by exploiting the forest floor niche. The diurnal pattern, despite large orbits, suggests orbit size may be an adaptation to dim understorey light rather than nocturnality; arboreal surveys could clarify links to pre-dawn calling. Birth seasonality likely tracks rainfall-driven food availability, supporting energy needs during lactation. Group size estimates are likely conservative given limited ground camera fields of view and vertical spacing when groups travel. The results provide species-specific behavioral ecology essential for conservation planning, including timing of protection efforts (e.g., during birth peaks) and mitigation of snare risks for terrestrial primates.

The study expands knowledge of lesula behavioral ecology, demonstrating high terrestriality, strict diurnality, wet-season birth peaks, and multi-female group structure with male dispersal. Camera traps are validated as a non-invasive, cost- and time-effective tool for cryptic primates in remote, hunted landscapes, generating actionable data for conservation management. Future research should deploy cameras at multiple canopy levels to quantify full locomotor profiles, use paired camera designs to improve group size estimation, investigate reproductive behavior to confirm birth seasonality, and explore the relationship between pre-dawn vocalizations and early-morning activity.

• Ground-based camera placement likely underestimates true group sizes due to limited field of view (∼50°) and vertical/behind-camera spacing of individuals.
• Video quality occasionally limited infant age classification; only events with high inter-observer agreement were retained, reducing sample size.
• Arboreal activity was not directly sampled; inferences about full locomotor profile are constrained to terrestrial detections.
• Some cameras malfunctioned (11 units excluded), slightly reducing sampling effort.
• Sex identification was often not possible (75% of mature individuals), limiting precision of sex ratio estimates.