BEHAVIORAL ECOLOGY OF CERCOPITHECUS LOMAMIENSIS

IN THE LOMAMI NATIONAL PARK AND BUFFER ZONE,

DEMOCRATIC REPUBLIC OF THE CONGO

by

Charlene S. Fournier Korchia

A Thesis Submitted to the Faculty of

The Charles E. Schmidt College of Science

In Partial Fulfillment of the Requirement for the Degree of

Master of Science

Florida Atlantic University

Boca Raton, FL

May 2020

Copyright 2020 by Charlene S. Fournier Korchia

ii BEHAVIORAL ECOLOGY OF CERCOPITHECUS LOMAMIENSIS IN THE LOMAMI

NATIONAL PARK AND BUFFER ZONE, DEMOCRATIC REPUBLIC OF THE

CONGO

by

Charlene S. Fournier Korchia

This thesis was prepared under the direction of the candidate’s thesis advisor, Dr. Kate M. Detwiler, Departments of Anthropology and Biological Sciences, and has been approved by all members of the supervisory committee. It was submitted to the faculty of the Charles E. Schmidt College of Science and was accepted in partial fulfillment of the requirements for the degree of Master of Science.

SUPERVISORY COMMITTEE:

______Kate M. Detwiler, PhD. Thesis Advisor

______Erik G. Noonburg, PhD.

______Rindy C. Anderson, PhD.

______Sarah L. Milton, PhD. Interim Chair, Department of Biological Sciences

______Ata Sarajedini, PhD. Dean, Charles E. Schmidt College of Science April 24th, 2020 ______Robert W. Stackman Jr., PhD. Date Dean, Graduate College

iii ACKNOWLEDGEMENTS

I would like to thank our collaborators, Drs. John and Terese Hart, and all the field assistants from the Democratic Republic of the Congo, without whom nothing would have been possible. Field work is too often regarded as a mean to an end, but I am grateful for your determination to work daily in the harsh conditions of the Congolese forests. A big thank you for your hard-work, dedication, and love for conservation.

I would like to thank all my lab mates from the FAU Primatology lab for their support and friendship. A special thanks to Daniel Alempijevic for his help on the project, from field information to data analysis in the lab, but also for his presence on a personal level, which had had many ups and downs throughout the years. Thank you for your continuous support, love, and advice.

Thank you to my committee members for their help on editing my thesis and thank you specifically to Dr. Erik Noonburg for his statistical analysis expertise and for his collaboration on my project, which will hopefully lead to publications shortly.

And finally, a special thanks to my advisor Dr. Kate Detwiler for giving me a chance in research, for welcoming me into her lab, and for giving me the opportunity to go through with my master thesis. Also, thank you for your advice, understanding, and patience for the last six years of my life. You have undoubtedly been a model and a motivation for me. I look forward to continuing our hard work through my PhD dissertation.

iv ABSTRACT

Author: Charlene S. Fournier Korchia

Title: Behavioral ecology of Cercopithecus lomamiensis in the Lomami National Park and buffer zone, Democratic Republic of the Congo

Institution: Florida Atlantic University

Thesis Advisor: Dr. Kate M. Detwiler

Degree: Master of Science

Year: 2020

In 2012, a new monkey species, Cercopithecus lomamiensis (lesula), was discovered in the Central Congo basin. Lesula is a semi-terrestrial rainforest guenon endemic to the area. Using a systematic grid approach, three terrestrial camera trap surveys

(two inside the Lomami National Park; one in the buffer zone) were conducted over three years to capture the cryptic species. The objectives of my study were to expand knowledge on the behavioral ecology of lesula and evaluate lesula’s sensitivity to hunting threats. The main findings from my study included: confirmation of terrestriality and diurnality, births clustering during the wet season, social group living of one male and multiple females, and high impact of hunting on group size. By studying the latest added species to the

Cercopithecini tribe, my thesis helps to better understand the ecological diversity occurring within this radiation of African primates and contributes to the species’ long-term conservation efforts.

v BEHAVIORAL ECOLOGY OF CERCOPITHECUS LOMAMIENSIS

IN THE LOMAMI NATIONAL PARK AND BUFFER ZONE,

DEMOCRATIC REPUBLIC OF THE CONGO

LIST OF TABLES ...... IX

LIST OF FIGURES ...... X

INTRODUCTION ...... 1

BACKGROUND ...... 2

Study area...... 2

Target species: Cercopithecus lomamiensis ...... 5

Guenon diversity ...... 6

Camera traps as a research tool ...... 8

CHAPTER 1: CAMERA TRAP SURVEYS: METHODS AND ANALYSIS ...... 9

Methods...... 9

Survey Sites ...... 9

Survey design ...... 9

Excel Database ...... 11

Sampling effort ...... 12

Events and capture rates ...... 13

Results ...... 13

vi CHAPTER 2: BEHAVIORAL ECOLOGY OF LESULA ...... 15

Introduction ...... 15

Methods...... 17

Terrestriality ...... 17

Activity pattern ...... 17

Birth seasonality ...... 17

Group size and composition...... 18

Results ...... 19

Terrestriality ...... 19

Activity pattern ...... 20

Birth Seasonality ...... 20

Group size and composition...... 22

Discussion ...... 22

CHAPTER 3: SENSITIVITY OF LESULA TO HUNTING ...... 26

Introduction ...... 26

Methods...... 27

Time-to-event analysis ...... 27

Rarefaction analysis...... 28

Hunting comparative analysis ...... 29

Group size comparative analysis ...... 29

Results ...... 29

Sampling effort needed for lesula capture ...... 29

Largest group size capture...... 30

vii Hunting: Relative abundance of lesula and game species...... 31

Variation in group size across sites ...... 32

Discussion ...... 33

CHAPTER 4: USING CAMERA TRAP DATA TO ESTIMATE PARAMETERS FOR

MODELING LESULA POPULATION SIZE ...... 36

Introduction ...... 36

Methods...... 38

Results ...... 41

Discussion ...... 42

CONCLUSION ...... 46

REFERENCES ...... 47

viii LIST OF TABLES

Page

Table 1. Sampling effort, independent events, event length, and capture rates ...... 14

Table 2. Definitions of infant age category and estimation of birth month...... 18

Table 3. Time-to-event analysis ...... 30

ix LIST OF FIGURES

Page

Figure 1. Map of survey sites ...... 4

Figure 2. First picture published of Cercopithecus lomamiensis ...... 6

Figure 3. Camera trap survey grids ...... 10

Figure 4. Screenshots of C. lomamiensis individuals ...... 12

Figure 5. Comparison between lesula and arboreal monkey species...... 19

Figure 6. Daily activity pattern of C. lomamiensis ...... 20

Figure 7. Comparison between rainfall and birth month ...... 21

Figure 8. Rarefaction analysis...... 31

Figure 9. Comparison between game species ...... 32

Figure 10. Box plot of the average and largest group size of lesula ...... 33

Figure 11. Diagram of the radius and angle of the camera detection zone ...... 37

Figure 12. Screenshot showing measurements of pixel counts ...... 40

x INTRODUCTION

Cercopithecus lomamiensis, known as lesula, is an Old-World monkey species recently discovered in the Democratic Republic of the Congo (DRC). Information currently available about the species is limited to the observational data collected during the initial discovery and documentation of the population to the scientific community. Hence, an additional field study was needed to collect finer detailed data on the ecology and behavior of the species.

The overall goal of my thesis is to increase our understanding of the natural history of the lesula species, which is the latest addition to the Cercopithecini tribe (radiation of

African monkeys known as guenons). My study is the first to carefully investigate lesula’s behavioral ecology by using data exclusively extracted from camera trap videos in two different environments (hunted versus protected forests). Specifically, I will (1) investigate the unique terrestrial niche exploited by the lesula monkey in relation to other guenon species, (2) provide supporting evidence on the impact of bushmeat hunting on the lesula population, and (3) derive parameters from camera trap data to estimate group and individual density of lesula.

Gaining information on the lesula species increases our knowledge on the ecological and behavioral diversity present within the Cercopithecus genus, as lesula adds a new branch on the phylogenetic tree of guenon evolution. In addition, new field data on lesula contributes toward conservation efforts of protecting this species and the primate fauna of the Congo Basin ecosystem, as lesula increases its endemic primate diversity.

1 BACKGROUND

Study area

The Lomami National Park and buffer zone, also known as the TL2 Landscape, named after the Tshuapa, Lualaba, and Lomami Rivers, is a 21,000 km² conservation area located in central DRC. The TL2 Project, which focuses on endemic species within the

TL2 Landscape, was established in 2007 to document and protect populations of primary conservation concern in the area. The region has a tropical climate, with stable precipitation throughout the year, outside of a brief dry season between June and July. Due to its equatorial latitude, the length of days and nights are consistently close to 12 hours year- round.

The Congo Basin is home to a little surveyed yet rich diversity of species and ecosystems, as evidenced by ongoing discoveries of new endemic species of fish

(Petrocephalus boboto and P. arnegardi), as well as new populations of bonobos (Pan paniscus) and okapis (Okapia johnstoni) (Lavoué & Sullivan, 2014; Pilbrow & Groves,

2013; Stanton et al., 2014). Despite a relatively low human population density, mammalian fauna is under threat from uncontrolled bushmeat hunting (Hart & Hart, 2011). Bushmeat hunting and trade put pressure on the survival and reproduction of many mammal species, especially non-human primates (Che et al., 2017; Cronin et al., 2017; Ziegler et al., 2016).

Protecting forests with high taxonomic richness has become a primary focus of the scientific community. However, conservation efforts are often met with conflict from local

2 human populations, which can increase the challenge of establishing long-term goals for the ecosystems’ health (Beyers et al., 2011).

Following discoveries of new populations, members of the TL2 Project identified an urgent need to protect endemic species from heavy hunting in their native range. After years of collaboration between local villages, government agencies, and the TL2 Project, the Congolese government created the Lomami National Park (LNP). LNP (2°0′0″S

25°2′0″E) was established on July 19th, 2016 to protect the high biodiversity of the region

(Hart, 2016). It is an area of 8,874 km² of contiguous tropical lowland rainforest within the

TL2 region (Figure 1) (Hart, 2016). Animals within the park boundaries have been protected by ranger patrols since 2013. However, bushmeat hunting still actively occurs in the buffer zone (around the park boundaries) with the implementation of hunting regulations (Hart, 2009; Hart & Hart, 2011).

3

Figure 1. Map inset: Lomami National Park (LNP) in the Democratic Republic of the Congo. Enlarged map: The three survey sites in the Lomami National Park and buffer zone.

4 Target species: Cercopithecus lomamiensis

While patrolling a village area in June 2007, a team from the TL2 Project came across a monkey specimen unknown to the scientific community. The first animal encountered was a female juvenile held captive as a pet (Hart, 2012). After months of patrolling the TL2 landscape, the first animals were seen in the wild. Although unknown to people outside the region, the monkey species was well known to local populations as lesula.

Lesula is endemic to the Lomami River Basin, which led to its scientific name of

Cercopithecus lomamiensis. C. lomamiensis is a medium-sized guenon using terrestriality as a common mode of locomotion. However, as being a cryptic animal, physical encounters in the wild are rare and have been of only few individuals at a time (Hart et al, 2012). Its presence is most commonly recorded by vocalization surveys (Hart et al, 2012). Its pelage color varies from black limbs to yellowish buff back fur and darkens from juvenile to adult.

It is described as having a distinctive elongated nasal ridge and large orbits (Figure 2) (Hart et al., 2012). The elongated nasal ridge is a possible characteristic of scent-marking behaviors in the long-faced monkey group, composed of only C. hamlyni before the addition of C. lomamiensis in 2012 (Kingdon, 2015). Phylogenetic analysis confirmed lesula clusters within the Cercopithecus clade, and a sister-species to C. hamlyni.

Molecular divergence dates suggest the last common ancestor of the two species to be between 1.7 to 2.8 Ma (Hart et al., 2012). C. hamlyni lives on the east bank of the Lualaba

(or Congo) River whereas C. lomamiensis lives on the west bank of the Lomami River.

Between these two rivers, no closely related species has been documented. This suggests that an allopatric speciation event occurred in the past, separating the ancestral population.

5

Figure 2. First picture published of a Cercopithecus lomamiensis individual. It is a captive adult male in Yawende, DRC. Photo by Maurice Emetshu.

Guenon diversity

The Cercopithecini tribe (commonly called guenons) is the most diverse and understudied radiation of the Old-World monkeys. It consists of a total of 34 identified species, of which 23 are from the Cercopithecus genus (Mittermeier et al., 2013). The TL2

Landscape represents high guenon diversity with six distinct Cercopithecus monkey 6 species encountered: C. wolfi, C. ascanius, C. mitis, C. neglectus, C. dryas, and C. lomamiensis. Recent discoveries of C. dryas and C. lomamiensis increase the endemic primate diversity within the Lomami National Park, which highlights the importance of continuous protection of its ecosystem (Hart et al., 2012; Hart et al, 2019).

Among the tribe, guenons exhibit high phenotypic, genetic, and behavioral diversity. Members of the Cercopithecus genus are characterized as arboreal guenons; however, the hamlyni species-group (C. hamlyni and C. lomamiensis) represents a unique phylogenetic lineage with an ecological niche that is thought to be mostly terrestrial

(Arenson et al., 2020). The incongruent pattern of lesula and hamlyni being terrestrial within a clade of arboreal monkeys raises questions about the evolution of terrestriality in guenons and the distinctiveness of the ecology and behavior of this species-group (Arenson et al., 2020). Unfortunately, it is difficult to address these questions, as to date there are very few studies about the natural history of either C. hamlyni or C. lomamiensis (Colyn &

Rahm, 1987; Easton et al., 2011; Hart et al., 2012; 2013; Rowe & Myers, 2013; Thomas,

1991). This lack of comprehensive ecological data warranted an intensive study on lesula to complement initial field observations and learn more about the behavioral ecology of one of the only two species comprising the hamlyni species-group.

However, the ability to make direct observations in the TL2 forest is restricted by the cryptic behaviors exhibited by lesula, reduced visibility in dense ground vegetation, and the remoteness of the environment. These factors limit the practicality of traditional methods of field-based observations, such as focal follows and line transects.

Consequently, lesula seems to be a good candidate for remote camera trapping methods in

7 order to increase the quantity and quality of information collected on this recently discovered species (Caravaggi et al, 2017; Rowcliffe, 2017).

Camera traps as a research tool

The use of camera traps to collect ecological and behavioral data on rare, non- habituated species has proven to be cost and time effective in previous studies (Boyer-Ontl

& Pruetz, 2014). Camera trap studies are non-invasive and are successfully used by primatologists as a research tool for conservation, management, and monitoring of primate populations (Galvis et al. 2014; Pebsworth & LaFleur, 2014).

Camera trap footage can be used for numerous research purposes in situations where it was previously difficult to obtain accurate information, such as monitoring human impact on wildlife. For example, Hegerl et al. (2017) used camera traps to assess bushmeat hunting by comparing occupancy of large forest mammals inside and outside a national park in Tanzania. They observed a significant decrease in large carnivores and omnivores in degraded forests outside the park.

When habituation is not feasible or potentially detrimental to animals’ health and wellbeing, camera traps can capture behaviors that are impacted by human presence

(Boyer-Ontl & Pruetz, 2014). Loken et al. (2013) used camera traps to document terrestriality in a species of orangutans (Pongo pygmaeus morio) as a locomotion strategy to overcome loss of canopy connectivity in Borneo. In addition, camera trap studies have become increasingly popular over the last decade to measure population parameters, such as animal abundance or density (Rowcliffe et al., 2008).

8 CHAPTER 1: CAMERA TRAP SURVEYS: METHODS AND ANALYSIS

Methods

Survey Sites

The TL2 Project established three research areas where the target species

Cercopithecus lomamiensis had been reported. Two of the survey sites (Losekola: 1° 23'

27.564'' S 24° 58' 27.732'' E and E15: 1° 38' 30.516'' S 25° 1' 58.26'' E) were within the newly formed Lomami National Park. The third survey site was in the buffer zone (Okulu:

1° 6' 22.608'' S 24° 31' 25.572'' E) where animals are intensively hunted (Figure 1). Okulu village is inside the lesula range and villagers were willing to collaborate with field assistants from the TL2 Project. The Lomami National Park is divided into management areas that are regularly patrolled by members of the TL2 Project. Losekola and E15 are sites where lesula is known to occur and patrol teams frequently report morning dawn calls.

The three sites are linked by contiguous, closed-canopy, tropical rainforest with no major geographical barriers separating them. In addition, they are of similar elevation and possess comparable fauna and flora (McPhee, 2015).

Survey design

Three terrestrial camera trap surveys were conducted over a three-year period.

Static grids were generated on ArcGIS software and the cameras were placed systematically at each location using a GPS. Forty-one cameras were set on a 4.1 km² grid between October and December 2013 in Okulu, 41 cameras on a 3.8 km² grid between

January and November 2014 in Losekola, and 20 cameras on a 12.0 km² grid between 9 August 2015 and January 2016 in E15 (Figure 3). Moultrie and Bushnell cameras of similar technical specifications were used throughout the surveys. All cameras had infrared sensors to detect animal movement during the day and night.

Figure 3. Camera trap survey grids of the three surveys: Okulu, Losekola, and E15 showing functioning camera traps with or without lesula captures, as well as nonfunctioning camera traps. The scale is consistent over the three maps.

Based on field observations, lesula was thought to be a semi-terrestrial species, which led to the choice of positioning ground cameras on trees facing animal trails. No bait was used, and sites were not selected by tree size or species, presence of fruits, animal prints or any other criteria (Bowler et al., 2016). The surveys were designed to capture C. lomamiensis behaviors; therefore, cameras were set to capture videos rather than photos.

Video length varied among camera traps and surveys from 20 to 90 seconds, but it is

10 unlikely that the difference resulted in capture bias. Date and time of capture were stamped on each video and all videos were set at the highest resolution (1280 x 720 pixels).

Excel Database

After receiving videos from the field to the lab, two independent teams composed of undergraduate research students and volunteers were trained to enter data. One team focused on lesula sightings whereas the second team recorded all species captured. We watched the entire set of videos (n = 1,885 in Okulu; n = 9,179 in Losekola; n = 3,570 in

E15) to identify the target species using MPEG Streamclip 1.9.2

(http://www.squared5.com/svideo/mpeg-streamclip-mac.html). The software allowed us to watch videos frame by frame for a precise recording of observation time. We used Excel

2016 to maintain a uniform database over the three surveys. I compared lesula data from both teams to control for coding errors and I corrected all discrepancies (n = 27) (Scotson et al, 2017).

Metadata were recorded, as well as the number of lesula individuals per video with information on age, sex, and any unique markings that might serve for individual identification. Unique markings included scars, kinked tails, and unusual pelage color.

Lesula age-class categories were defined as follows: (1) infant – clinging to mother, may move away for short periods of time, not weaned, (2) juvenile – immature individual who is fully weaned, travels independently of mother, (3) large juvenile – individual not fully mature but almost fully grown, (4) big – fully grown individual where we were unable to determine sex (Figure 4). Sex categories were determined as follows: adult male – visible testicles and adult female – visible nipples or a clinging infant (Figure 4).

11 Figure 4. Cropped screenshots of camera trap videos of C. lomamiensis individuals: (A) an adult male, (B) an adult female with a juvenile, (C) an adult female with a clinging infant, and (D) an adult male and adult female in close proximity showing sexual dimorphism in body size.

Sampling effort

Several rotations were done at each camera trap site, where cameras were checked several times, keeping their positions the same throughout the study. Field assistants recorded set-up and pick-up dates for each rotation. However, I marked the beginning and the end of each rotation as the date of set-up and the date of last video respectively.

Information on camera failure at the pick-up time was not available. I considered low battery and/or full SD card as significant factors for the camera to stop triggering.

12 Therefore, I recorded the total sampling effort at each camera trap site as each rotation end- date minus set-up date. In some cases, C. lomamiensis events were captured during the first and last days of rotations. These events were included in the dataset for analysis. I assumed a half-day effort for the day of the camera set-up and a half-day effort for the day of the camera pick-up, regardless of recorded times.

Events and capture rates

I defined an event as a series of chronological videos where a cluster of activity was observed. A new event was recorded after a 30-minute period without any detected activity.

I averaged event length and recorded the longest event captured at each site. To determine capture rate, I divided the number of events by the number of camera trap days and multiplied it by 100 for analysis of event rates (Gregory et al., 2015). There was only one month (October) where the three surveys were all active, so I used the capture rate in

October for comparison purposes.

Results

Data were recorded from 39, 33, and 20 cameras from Okulu, Losekola, and E15 respectively (Figure 3). I excluded 11 cameras from the analysis due to malfunctions. I obtained a total of 598 independent lesula events from 5,960 camera trap days over a three- year period (Table 1). More events were recorded in Losekola than in either Okulu or E15 during the full survey length as well as in the month of October (Table 1). The average event length is similar at the three sites and the longest event was recorded in Losekola

(Table 1). Also, I recorded similar capture rates at the three survey sites (Table 1).

13 Table 1. Sampling effort (days), number of independent events, mean event length ±SE (minutes), longest event recorded (minutes), capture rate for full-length survey and October only, for the three different surveys.

Capture Mean Longest Capture Site Effort # events rate event length event rate (October)

Okulu 1,551 143 2.27 (±0.32) 28 9.22 8.76 Losekola 2,430 282 3.80 (±0.35) 69 11.61 24.59 E15 1,979 173 2.88 (±0.31) 35 8.74 8.40 Total 5,960 598 - - - -

14 CHAPTER 2: BEHAVIORAL ECOLOGY OF LESULA

Introduction

In the evolutionary history of the Cercopithecini tribe, a leading hypothesis is that terrestriality originates from a single evolutionary transition, for patas (Erythrocebus patas), l’Hoesti (Cercopithecus lhoesti), and savannah monkeys (Chlorocebus sp.) (Tosi et al., 2004, 2005; Xing et al., 2007). However, a recent study looking at the morphology of lesula skeletons shows a strong adaptation to terrestriality, with a clear ancestral preference for arboreality (Arenson et al., 2020). These results corroborate behavioral observations of lesula as a semi-terrestrial to terrestrial primate and imply multiple evolutionary transitions in substrate use among the guenon radiation (Arenson et al., 2020;

Hart et al., 2012). This transition may be an adaptation of the lesula monkey as a response to ecological pressures in the higher substrates. The high species diversity in the Congo

Basin may have driven lesula to explore the forest floor and find an underutilized niche as a way to avoid competition with sympatric species (Kautt et al., 2018). Using a behavioral approach, I examined the substrate use of lesula through terrestrial camera traps as a way to determine its primary mode of locomotion.

As an Old-World monkey species, the prediction is that lesula exhibits strictly diurnal behaviors; however, the morphological analysis of lesula skulls illustrates a unique shape with orbits larger than other guenon species (Hart et al, 2012). As morphology is linked to function, a study looking at 60 extant primate species reveals that nocturnal species exhibit proportionally larger orbits than diurnal species (Kay & Kirk, 2000). 15 Consequently, the large orbits of lesula could be potentially linked to a new adaptation and demonstrate a divergent activity pattern. Looking at the recorded videos over a three-year period, I determined the daily activity pattern of the lesula monkey.

Guenon species are also known to exhibit reproductive and birth seasonality

(Butynski, 1988; Cords, 1988). When no marked annual variation in daylength or temperature are observed, rainfall and food availability are likely to determine the timing of reproduction (Butynski, 1988). In environments that reflect two wet seasons per year, such as the DRC, guenon birth seasons are often estimated to last between four and six months and to be correlated with the first half of the long, wet season (Butynski, 1988).

My study looked into a possible birth seasonality in the lesula species, with a potential positive relationship between birth season and precipitation of the Congo Basin.

Guenons usually live in medium sized social groups, from 10 to 40 individuals, composed of a single male and multiple females with their juveniles (Cords, 1987; 1988).

Females are philopatric and males disperse when they reach sexual maturity (Kingdon &

Groves, 2013). As a Cercopithecus species, lesula should follow the same social system; however, there are few data available on lesula group size and composition. Field observations from the initial study on lesula provide a maximum number of five individuals per group and no information on group composition (Hart et al., 2012). Lesula’s small group size may be attributed to an ecological adaptation, which would give the individuals an advantage of living in groups of reduced size or it may be the results of limited surveys in a challenging environment. The species’ cryptic behaviors make it difficult to conduct group counts in a forest environment. The use of camera traps, as an observational tool, is essential to capture data from non-habituated animals without disturbing individual

16 behaviors or the dynamics of a group. My study investigated the behavioral ecology of the lesula species for the first time relative to other guenon species.

Methods

Terrestriality

Arboreal monkeys are known to rarely use the forest floor. Thus, I compared capture rates of all lesula events to those of arboreal monkey species detected by the terrestrial camera traps at the three sites, which include Cercopithecus ascanius, C. mitis, and C. wolfi. I also calculated the percentage of lesula terrestrial events, which I defined as at least one video per event, where a minimum of one lesula was seen with all four limbs on the ground for at least one second.

Activity pattern

By organizing all the video events by time of day, I determined hourly capture rates to reveal the activity pattern of lesula over a 24-hour period.

Birth seasonality

I recorded infant captures in an Excel database, with an estimate of their age and birth month using the features listed in Table 2 (Förster & Cords, 2002; Krige & Lucas,

1975; Lee, 1984). To increase accuracy, two other observers viewed all infant videos and estimated their age following the same categorization. I discarded events for which no agreement could be made.

Precipitation data were collected between 2008 and 2013 in Obenge (1° 24' 24.588''

S 24° 59' 22.128'' E), which was a former village located within the Lomami National Park boundaries (mean annual precipitation = 2021 mm; rainfall days = 145 days) (J. Hart, unpublished data). 17 Table 2. Definitions of infant age category and estimation of birth month.

Birth month Category Definition estimate

Newborn infant; Clinging to mother; <1 month Infant 1 Skin can be seen through fine textured hair; before event Infant body size is relatively small compared to mother body size (range from 1/4 to 1/3).

Hair fully grown; Can be clinging to the mother or walking by itself where mother is in proximity (both individuals 3 months Infant 2 captured on the same screen); before event If clinging, infant covers the full ventral portion of the mother.

Independent from mother; Still possesses infant physical features, 6 months Infant 3 behaviors, and coloring; before event Transitioning from infant to juvenile.

Group size and composition

I defined group size by using a conservative estimate where individuals are identified and counted within an event using age, sex, and/or physical features, as well as their position and directionality of movement in the videos. Each group size estimate was recorded by at least two observers to ensure reliability.

For calculation of mean group size, I defined a group as three or more individuals with at least one female or juvenile present (Glenn, 1997). I averaged group size for the

18 three survey sites and noted the largest group captured. I determined group composition by assessing the number of individuals per event in each age-sex category.

Results

Terrestriality

Lesula events captured by the terrestrial camera traps were higher than those of arboreal monkey species (Figure 5). Lesula events were terrestrial 98.8% of the time, strongly suggesting terrestriality as its main mode of locomotion. In contrast, I recorded only 11 events for the three other Cercopithecus species observed (C. ascanius, n = 1; C. mitis, n = 9; C. wolfi, n = 1) with only one non-terrestrial event for C. ascanius.

12 10.03

8

4 Events/100 days CT Events/100 0.02 0.15 0.02 0 C. ascanius C. mitis C. wolfi C. lomamiensis Monkey Species

Figure 5. Comparison of capture rates between lesula and arboreal monkey species (Cercopithecus sp.) in the Lomami National Park area, DRC.

19 Activity pattern

All recorded lesula events occurred between 6:00 and 19:00 hours, confirming lesula’s diurnal behavior (Figure 6). Data show a reasonably constant activity for 11 hours across the day with a slight peak at 14:00 hours.

1.2 Night Time 1.0 0.8 0.6 0.4 0.2

Events/100 days CT Events/100 0.0

Hour of the day

Figure 6. Daily activity pattern of C. lomamiensis estimated by averaging capture rates over the three surveys.

Birth Seasonality

I recorded 77 events involving infants with 104 independent infant sightings, representing 12.9% of all lesula events. Only infant sightings that could be assigned to age categories by at least two of the three observers were used to estimate the timing of the birth season, yielding 60 events with 80 infant sightings. I discarded a total of 17 events

(22%) and 24 infant sightings (23%). Of the three categories used for infant classification,

20 infants 1 (n = 38) were separated from infants 2 and 3 (n = 42) in the analysis (Table 2). A major birth peak occurred from July through October (Figure 7). An overlap was detected between the number of infant 1 births and the amount of rainfall during these months

(Figure 7).

12 300

10 250

8 200

6 150

4 100 (mm) Rainfall Number of birth of Number

2 50

0 0

Infants 1 Infants 2-3

Figure 7. Comparison of rainfall and births. Average monthly rainfall in millimeters in Obenge, DRC (S1.40683 E24.98948) inside the Lomami National Park between 2008 and 2013 (J. Hart, unpub, data). Estimated number of births per month of C. lomamiensis over the three surveys. Infant age categories are labeled regarding estimated age (1: <1-month- old; 2: about 3-month-old; 3: about 6-month-old). I grouped infants 2 and 3 for analysis purposes.

21 Group size and composition

The mean group size from the three surveys combined was 6.81 individuals (SE =

.27). The largest group captured was 32 individuals in Losekola.

Groups mostly comprised females and juveniles, and adult males were observed only in larger groups. The mean group size where two group males were captured was 14 individuals per group (n = 9 events). Even though I identified approximately the same number of male and female individuals in all surveys combined (males = 147, females =

140), I recorded 58 lone male events compared to only 12 lone female events. In addition,

I observed small groups of two and three males on two occasions.

The ratio of mature to immature individuals was about 5:3 (mature = 1,163, immature = 696). Individuals for which no age or sex could be determined, classified as unknowns, represented approximatively 19% and were excluded from ratio calculations. I decided not to include male to female ratio due to a high number of adults with no sex identification (males = 147, females = 140, bigs = 876).

Discussion

I successfully used camera traps to increase our knowledge on the behavioral ecology of the little-known C. lomamiensis guenon species. I was able to confirm terrestriality and diurnality, estimate the timing of the birth season, and estimate group size and composition of the lesula monkey.

This behavioral study confirmed terrestrial quadrupedalism as lesula’s main mode of locomotion. The transition to terrestriality in the lesula species is possibly an adaptation to minimize interspecific competition in the forest canopy (Kautt et al., 2018; Stroud &

Losos, 2016). Given the different types of available resources in a rainforest, the ancestral

22 population may have taken advantage of a new ecological opportunity of using the forest floor (Stroud & Losos, 2016). The divergence in substrate use and foraging resources from sympatric guenon species could have promoted the adaptive radiation of the hamlyni species-group (Yoder et al., 2010). The next steps will be to select sites where lesula is abundant and set camera traps at different canopy levels (camera trap “column”) to assess the frequency of substrate use and the full locomotor profile of the lesula species

(Alempijevic et al., in press).

Lesula activity pattern was confined to the daylight hours, consistent with the diurnality of all Cercopithecus species. Spending most of its time on the forest floor, lesula’s large orbits are probably an adaptation for dim forest environment rather than an adaptation for a divergent activity pattern.

The birth season, based on infant category 1, occurred in July through October. The equatorial LNP region does not experience marked monthly variation in day length or temperature and possesses two wet seasons per year. Lesula was predicted to experience a birth seasonality during the second wet season, which would last between four and six months (Butynski, 1988). Birth season of C. lomamiensis conforms to the expected pattern and lasts about four months in the second wet season of the year (Figure 7). In this case, rainfall was used as a proxy for food availability and fruiting season. Fruits are high energy food items for primates and are preferred by nursing females to successfully raise their offspring (Nowell & Fletcher, 2007). As little is known on the feeding ecology of lesula, this initial observation could be used for designing future behavioral studies. In the analysis of birth seasonality, I decided to separate infants 1 from infants 2 and 3. The reason is that it was difficult to assign older infants to categories due to lack of visibility or hard-to-assess

23 criteria. Thus, agreement between all three observers for infant 2 and 3 categories was challenging to achieve. However, all observers agreed 100% of the time on the infant 1 categorization, which represented the most objective approach in estimating birth season.

My study provided preliminary results of the birth seasonality exhibited by the lesula monkey, but a full study is required to increase the reliability of the results.

Based on field observations, it was thought that lesula lived in small social groups

(Hart et al., 2012). However, my study reveals that lesula can live in larger groups, up to

32 individuals. Medium group size is observed in other closely related terrestrial species, such as Allochrocebus solatus and Cercopithecus hamlyni, which live in groups of 8-25 and 2-22 individuals respectively (Rowe & Myers, 2016). However, data suggest that the mean group size (m = 6.81 individuals; SE = .27) for the whole dataset was on the low range and probably underestimated. Despite being a semi-terrestrial to terrestrial species, individuals climb into the canopy and space themselves out when traveling, so it is difficult for ground camera traps to capture and record complete social groups. By having a limited

50-degree ground vision, the camera detection zone will capture partial group count more often than complete group counts.

Findings reveal that lesula groups are mostly composed of multiple females with their juveniles, which suggests a polygynous mating system. In addition, I recorded all- male groups comprising up to three males. In other field studies of guenons, groups of lone males (one to three in this study) are identified as bachelor groups. Therefore, it is expected that lesula exhibits a female philopatry and male dispersal social system, similar to most

Cercopithecus species (Cords, 1987;1988; Kingdon & Groves, 2013). In addition, results indicate that lesula groups mostly contain single males; however, a few larger groups

24 contained two mature males. There are three possible explanations for these video observations. First, one of the males may have been in the transition of becoming sexually mature (from subadult to adult) and at the edge of dispersing. Second, in some

Cercopithecus species, it is common to observe multimale influxes and promiscuous mating during the breeding season (Cords, 1987; Glenn, 1997; Pazol, 2003). In the case of lesula, one of the two males observed may have been an immigrant male that joined the group for a limited period of time (part of the mating season), when the resident male’s position in the group was being challenged by bachelor males (Cords, 1987). Finally, multiple males in female dominant groups in lesula may be the norm when groups become large enough. It has been noted in some Cercopithecus species that several sexually mature males can reside in groups for several months (Bourlière et al., 1970; Galat-Luong, 1975).

25 CHAPTER 3: SENSITIVITY OF LESULA TO HUNTING

Introduction

Wildlife populations in the DRC have been largely depleted by uncontrolled and unsustainable bushmeat hunting (Hart & Hart, 2011). For the last seven years, the deep conservation commitment of the Congolese government and the TL2 Project led to the creation of the LNP and the strong enforcement of laws that prohibit hunting inside the boundaries of the park for all species (Hart, 2016). However, hunting regulations differ between species in the buffer zone. Despite hunting seasons and limited animal quota regulations being in place, wildlife still faces important hunting threats.

Ungulate and primate species are actively hunted in the buffer zone around the LNP

(Amboko et al., 2018). A recent assessment of C. lomamiensis’ conservation status estimates that the species faces hunting pressures in 80% of its range during the hunting season (December to August) (Detwiler & Hart, in press). Lesula’s conservation status is listed as Vulnerable, so it is important to assess the impact of hunting on its population

(Detwiler & Hart, in press). Determining game species abundance between hunted and protected sites can provide useful information about the effectiveness of hunting regulations implemented in the LNP (where hunting is strictly prohibited) and the buffer zone (where hunting seasons apply).

Bushmeat hunting in the buffer zone may be an important factor influencing the lesula population. As a highly social species C. lomamiensis may face a significant impact on its group size due to high predation risk. Two major factors account for group living in 26 primates: food resources and predator avoidance. As no information on feeding ecology was available from the camera trap study, I focused on investigating predation risk on the lesula population. Living in social groups confers two important benefits to primate populations based on: (1) the detection effect – grouping increases the likelihood of detecting predators before being detected and (2) the diffusion effect – grouping decreases the chances of each individual of being preyed upon (Dehn, 1990). Differences in group size can then be explained by the disadvantages of group living, such as reduction in foraging efficiency (Wrangham et al., 1993). Primates must balance costs and benefits of group living to increase their chances of survival. For instance, a comprehensive study looking at different baboon species (Papio sp.) throughout Africa shows an increase in group size when predator density becomes higher, which refers to the two first benefits of living in large social groups for predator avoidance (Bettridge & Dunbar, 2013).

Comparing capture rates and group size in hunted and protected sites gives information on a possible strategy used by lesula individuals to limit their sensitivity to hunting threats. My study provides the first supporting evidence on the impact of hunting on the lesula population and consequently encourages the longitudinal conservation efforts of the Congo Basin area.

Methods

Time-to-event analysis

Uneven sampling effort can bias comparisons between sites, so I used statistical analysis to control for sampling bias on capture rates. This was necessary since Okulu, the hunted site, was the shortest camera trap survey.

27 In collaboration with my committee member, Dr. Erik Noonburg, I ran a time-to- event analysis by modeling lesula captures as randomly distributed in time. Under this assumption, inter-event times follow an exponential distribution, which was approximately true in my data. The number of days needed for 95% confidence of capturing lesula on camera is therefore the 95th quantile of an exponential distribution with rate parameter equal to the capture rate for each site. I conducted the same analysis using only events of

15 or more individuals for Losekola and E15 sites. No group size of 15 individuals or more was recorded in Okulu. This way, I estimated the sampling effort (camera trap days) needed at each site to capture a large group count of lesula. All analyses were conducted in R statistical software version 3.5.1 (https://www.r-project.org).

Rarefaction analysis

The Okulu survey in the buffer zone investigates the lesula population where it is actively hunted but has also the smallest sampling effort of the three surveys. To control for bias on group size findings, Dr. Erik Noonburg and I used rarefaction curves to assess if the sampling effort was sufficient to capture the largest group size of lesula at each site.

The camera trapping process was simulated for each site by drawing a sequence of 100 encounter events at random, where each event was an observed group size. At each step in the sequence, I recorded the maximum group size that occurred up to that step. The process was repeated for 1000 simulated sequences and the maximum group size was averaged at each step. Under the assumption that events are distributed at random over time, the average inter-event time (i.e. sampling effort) is equal to the inverse of the total capture rate for each site. I plotted the average maximum group size versus the total expected

28 sampling effort for 1,2,…,100 captures. All analyses were conducted in R statistical software version 3.5.1 (https://www.r-project.org).

Hunting comparative analysis

I compared capture rates of lesula between hunted versus protected sites and compared lesula to the five mammal species most captured on cameras, which are also the most targeted by hunters: four species of duikers (Philantomba monticola, Cephalophus weynsi, Cephalophus dorsalis, and Cephalophus sivicultor) and one species of wild hog

(Potamochoerus porcus).

Group size comparative analysis

Using the methods of group size estimates described in Chapter 2, I conducted a one-way ANOVA and post-hoc comparisons using the Tukey HSD test to compare mean group size at each of the three sites. I conducted the analyses in R statistical software version 3.5.1 (https://www.r-project.org). I also compared the largest group size captured in hunted versus protected sites.

Results

Sampling effort needed for lesula capture

The number of days required to observe (CL = 95%) a lesula event was less than

35 days at any site (Table 3). Rare observations of group count of 15 or more individuals required more than 700 camera trap days (Table 3).

29 Table 3. Time-to-event analysis (in days) for all events at each site and for events with only 15 or more individuals captured in Losekola and E15.

Time-to-event Site Time-to-event (15+ individuals) Okulu 32.49 - Losekola 25.81 727.96

E15 34.27 846.94

Largest group size capture

The largest lesula group counts were captured early in the three surveys (< 300 camera trap days), confirming that the sampling effort (> 1,500 camera trap days in any survey) was sufficient to capture the largest group at each site (Figure 9).

30 40

30

20 Group size Group

10

0 0 300 600 900 Camera trap days

Okulu Losekola E15

Figure 8. Rarefaction curves of lesula average group size in function of camera trap days for each survey site: in the buffer zone (Okulu) and in the Lomami National Park (Losekola and E15).

Hunting: Relative abundance of lesula and game species

Of the five ungulate species compared, capture rates were lower at the hunted site than the two protected sites (Figure 8). However, lesula capture rates were similar across all three surveys (Figure 8).

31

40.9 40 Okulu Losekola E15

30 24.9

20

11.6 9.2 9.2

Events/100 days CT Events/100 8.7 10 7.0 5.3 5.5 5.1 2.6 1.5 1.9 0.9 0.1 0.3 1.2 0.1 0 P. monticola C. weynsi C. dorsalis C. sivicultor P. porcus C. lomamiensis

Mammal species

Figure 9. Comparison of the six highest capture rates (events per 100 camera trap days) of game species at the hunted site (Okulu) and the protected sites (Losekola and E15). P. monticola, C. weynsi, C. dorsalis, and C. sivicultor are duiker species and C. porcus is a hog species.

Variation in group size across sites

The one-way ANOVA test showed a significant difference in group size at the p<.05 level for the three conditions [F(2) = 5.96, p = .003]. Post-hoc comparisons using the Tukey HSD test indicated that mean group size at Okulu (M = 4.88, SD = 0.34) was significantly smaller than mean group size at Losekola (M = 7.11, SD = 0.38) and E15 (M

= 7.48, SD = 0.57) respectively (Figure 10). As well, the largest group recorded in Losekola and E15 was more than twice the size as in Okulu (Figure 10).

32

** **

Figure 10. Box plot showing the average and largest group size of lesula at each survey site. One-way ANOVA: F(2) = 5.96, p = .003; Tukey HSD: Losekola-E15, p = .826; Okulu- E15, p = .004; Okulu-Losekola, p = .007.

Discussion

By comparing the capture rates of game species events at the three sites, I can conclude that ungulate species at Okulu are strongly affected by bushmeat hunting compared to the two sites where hunting is strictly prohibited. In the future, another survey in hunted forests would be desirable to rule out any ecological differences at the different 33 sites. However, capture rates of lesula were similar at the three sites regardless of hunting activity, suggesting a high density of lesula groups throughout their range.

In addition, findings suggest that lesula group size does differ between hunted and protected sites. Rarefaction analysis ruled out the possibility of less sampling effort at

Okulu affecting the observed differences. Hunting threats in the buffer zone may then be the cause of the smaller group size observations. When dealing with high predation risk, animals rapidly adopt behavioral responses that will increase their individual chance of survival (Bettridge & Dunbar, 2013; Tutin & White, 1999). Two forms of antipredator strategies exist: (1) avoiding being detected by predators and (2) dealing with the predators once detected (Dugatkin, 2004). In baboon species, the positive correlation between predator density and group size provides a survival advantage (Bettridge & Dunbar, 2013).

By increasing their vigilance, baboons are able to reduce the effect of surprise and deter an ambush predator from attacking (Bettridge & Dunbar, 2013). However, each species must balance their own costs and benefits of group living based on their ecological pressures.

Lesula individuals seem to use crypticity as an adaption to reduce their chances of detection by predators. As well, the type of predators may have an influence on the behavioral response of the animals. When predators are human hunters and lesula individuals have limited possible outcomes once detected, decreasing detection by reducing their group size may be the best long-term strategy to increase their chances of survival.

The slow reproductive rate of primates makes it difficult for their populations to recover when human hunting is extreme and continuous (Strier, 2017). If the harvest rates become higher than the replacement rates, hunting becomes unsustainable, which makes the population highly vulnerable to decline (Refischt & Koné, 2005; Strier, 2017).

34 Consequently, understanding the human impact on the lesula population is an important step to evaluate the benefits of the implementation of hunting regulations in the area.

The time-to-event analysis helped determine the sampling effort needed to capture single events and group count of more than 15 individuals on videos. This information will be used to design future camera trap survey protocols (number of cameras needed or length of survey) to answer specific behavioral questions.

My study contributes new information on the consequences of intensive hunting on game species populations outside LNP. Decline in wildlife communities in the buffer zone could lead to ecological decay and imbalance in the whole LNP ecosystem (Tagg et al.,

2020). The results from this study provide baseline data for the TL2 Project to use in future surveys to monitor remote and endangered species for the long-term success of the implementation of hunting laws. Monitoring lesula populations over time within the national park and buffer zone will actively contribute to the conservation of the Congo

Basin ecosystem and its endemic species, such as the lesula monkey.

35 CHAPTER 4: USING CAMERA TRAP DATA TO ESTIMATE PARAMETERS FOR

MODELING LESULA POPULATION SIZE

Introduction

Density is challenging to assess in the field when dealing with non-habituated, cryptic primates (Pebsworth & LaFleur, 2014). In recent years, camera traps have been a popular tool for statistical modeling using a method called capture-recapture, which estimates density of individually identifiable species (Burton et al., 2015). However, it has not been feasible to identify enough lesula individuals on camera trap videos to use capture- recapture methods. Many species are not easily identifiable using videography data, so new methods are being created to bypass this limitation (Campos-Candela et al., 2017; Evans

& Rittenhouse, 2018; Rowcliffe et al., 2008)).

The random encounter model (REM) is a new method that was created to estimate group and individual density without the need for individual recognition (Rowcliffe et al.,

2008). The following REM equation can be used with parameters exclusively extracted from camera trap videos (Rowcliffe et al., 2008: equation 4):

푦 휋 퐷 = . 푡 푣푟(2 + 휃) in which 푦 is the number of independent events, 푡 is the total sampling effort (days), 푣 is the average speed of animal movement (km/day), and 푟 and 휃 are the radius (km) and angle

(radians) of the camera trap detection zone respectively (Figure 11). The REM equation above measures group density, but when multiplied by group size, it can derive an estimate

36 of individual density. The individual density calculated from the REM equation and the known geographic range provides a method to estimate total population size for lesula, a first for this species (Hart et al., 2012).

Angle (degrees)

Radius (m) Radius

Figure 11. Diagram showing two parameters used in the REM equation: the radius 푟 and the angle 휃 of the camera detection zone. The radius and angle were calculated in meters and degrees in the field and then converted to kilometers and radians respectively for use in the REM equation.

37

Census surveys of primate populations are fundamental to create management plans for long-term conservation efforts, as well as to understand sociality in different primate species (Fashing & Cords, 2000). Different census survey methods exist in the field to estimate density values of non-human primates. One the most reliable methods is based on home range size and area occupancy (Fashing & Cords, 2000). However, this method requires the process of habituation, which is the relatively persistent waning of a response as a result of repeated stimulation that is not followed by any kind of reinforcement

(Thorpe, 1963). In primate studies, it means that the repeated neutral contacts between humans and monkeys can lead to a reduction in fear, and ultimately to the ignoring of an observer (Williamson & Feistner, 2003). Primatologists can use habituation in the field to be able to recognize different social groups and follow them in their natural environment for hours at a time to collect data required for modeling methods. Until habituation of a lesula group becomes possible, extracting data from camera trap videos provides a viable option to estimating density of this species. An important goal for the conservation of lesula is to determine a reliable method for estimating a total population size. Applying this method would help to monitor the lesula population over time and detect population trends, such as a possible decline. As a first step toward establishing baseline data for estimating the lesula population size, my study determines the parameters needed to use the REM method for the first time on a primate species.

Methods

I extracted data from camera trap videos to estimate parameters that will be used in the REM equation described previously:

38 (a) Independent events: See Chapter 1, Methods, section ‘Events and capture rates.’

(b) Sampling effort: See Chapter 1, Methods, section ‘Sampling effort.’

(c) Speed of animal movement: The first step toward the estimate of day range was to estimate lesula travel speed using video recordings. Travel speed is defined as the average speed at which an active animal moves across its environment (Rowcliffe et al.,

2016). I analyzed videos (n = 40) based on two criteria: (1) the animal travels perpendicularly to the camera trap detection zone and (2) the sex of the animal is clearly determined. Adult males (n = 21) are identified based on visible testicles and adult females

(n = 19) based on visible nipples or clinging infants. Foraging and vigilance are integral parts of daily activity, so I included these behaviors in the estimate of travel speed.

However, I discarded sequences where animals react to the camera traps to reduce bias in the analysis. I used pixel counts to measure animal body size (from top of the head to base of the tail) and the distance traveled on videos using the ImageJ software (Figure 12). I applied an average lesula body size from field measurements of 53.50 cm for adult males

(n = 2) and 44.55cm for adult females (n = 4) to calibrate the pixel count and then estimate the distance traveled in meters (Hart et al., 2012; D. Alempijevic unpublished data adult female specimen, 2017). The distance travelled was divided by the duration of the sequence in seconds to obtain a travel speed (in meters per second) (Rowcliffe et al., 2016). The travel speed of adult male and female individuals was averaged for further analysis.

39

Figure 12. Screenshot of a lesula video showing measurements of pixel counts for the animal body size (from top of the head to base of the tail) and for the distance traveled on video. Pixel counts were measured using ImageJ software and were used for travel speed calculations.

Using the average travel speed calculated, I extrapolated an average day range. Day range is defined as the rate of movement over a 24-hour period (Rowcliffe et al., 2016):

푣 = 휇푝 in which 푣 is the day range (km/day), 휇 is the travel speed (km/h), and 푝 is the proportion of time spent traveling (h/day). No data on lesula’s activity budget are available; therefore,

I used a travel activity budget from two semi-terrestrial sister species, Cercopithecus hamlyni and Allochrocebus lhoesti (Goodwin & Kaplin; Ukizintambara). To obtain the proportion of time spent traveling, I multiplied the percentage of travel activity to the

40 number of hours lesula was active per day, which I calculated from the daily activity pattern

(Chapter 2, Methods, section ‘Activity pattern’). I removed the earliest and latest hours of the day, as the event capture rates were below 0.2. By multiplying the result to the travel speed, I obtained an estimate of day range, one of the key parameters in the REM equation.

(d) Radius of camera detection zone: The camera radius 푟 was determined by conducting field trials. The camera traps were approached from the front at different distances to trigger videos and the furthest distance the camera was able to capture videos was recorded as the radius.

(e) Angle of camera detection zone: I extracted values for the angle of camera detection zone from three sources: the camera trap manufacturer’s guides (Moultrie and

Bushnell), Cusack et al. (2015), and McPhee (2015).

(f) Group size: See Chapter 2, Methods, section ‘Group size and composition’.

Results

The parameters estimated, such as travel speed, day range, and group size, are in line with closely related Cercopithecus species. Results of each parameter are described below:

(a) Independent events: see Table 1

(b) Sampling effort: see Table 1

(c) Speed of animal movement: I calculated a travel speed of 0.29 m/s and 0.37 m/s for males and females respectively. By combining both sexes, I estimated an average travel speed of 0.32 m/s.

Traveling budget from semi-terrestrial sister species, Cercopithecus hamlyni and

Allochrocebus lhoesti was averaged to 27% per day. To obtain the proportion of time spent

41 traveling, I multiplied the percentage of travel activity to the number of hours lesula was active per day that I calculated from the daily activity pattern (11 hours) (Figure 6). By multiplying the result to the travel speed, I obtained an average day range of 3.4 km/day.

(d) Radius of camera detection zone: There was a variation in radius measurements in the field for the different cameras placed at different locations. The field measurements ranged from 6m to 14m.

(e) Angle of camera detection zone: The angle described by the camera trap manufacturer’s guides, the calculated angle in Cusack et al. (2015), and McPhee (2015), all had very similar values for this parameter, approximately 50 degrees.

(f) Group size: See Chapter 2, Results, section ‘Group size and composition’ for average group size at the three sites combined and Chapter 3, Results, section ‘Variation in group size across sites’ for average group size at each site.

Discussion

The ability to use remote cameras to estimate the status of wildlife populations significantly expands the contexts in which these data can be applied (Evans &

Rittenhouse, 2018; Nakashima et al., 2017). Rowcliffe et al. (2008) created the random encounter model to calculate animal density without the need for individual recognition.

In my study, I successfully extracted the different parameters for further use in the density model equation.

One pitfall of the REM method is that the parameters used are proxies for true measurements. Missing actual parameters derived from direct field measurements may impact the accuracy of the results. However, until lesula social groups are habituated to use

42 more accurate census methods, it would be a valuable first attempt in documenting density and population size of the lesula species.

Based on Rowcliffe et al. (2008), the model follows four assumptions, which I address here.

(1) Animals move independently of the camera traps, which is true in my study.

The camera traps were placed on systematic grids not following any bias. Even though cameras were facing animal trails, lesula was actively seen traveling perpendicular to trails, which suggests that trail usage for camera trap placement was not a limiting factor.

(2) Camera traps are randomly assigned for each event to be independent from each other, which is true. I defined independent events as being separated by 30-minute intervals. The only issue would be if two independent events occurred within 30 minutes and were counted as one. However, this possibility seems to be unlikely or rare as most events had tightly clustered videos.

(3) Population is closed, which means that no individuals could enter or leave the social group during the study period. Wild primate populations are dynamic due to births or deaths of individuals as well as immigration/emigration of males in male dispersal species. However, primate life history is relatively long compared to the camera trap survey length in my study (between 3 and 11 months). Therefore, it is unlikely that the numbers of these ecological events were high enough to drastically impact the REM density estimates.

(4) Animals move independently to each other like gas particles. For primate species, this assumption seems to be violated as lesula individuals live in social groups.

However, the assumption may be extrapolated to groups moving independently to each

43 other. Rowcliffe et al. (2008) state that the REM method is still reasonably robust despite this assumption being violated, but it is a factor that needs to be further investigated.

Recent research has identified concerns with biases in using the REM model. An example of possible bias in interpreting REM results is found in Cusack et al. (2015).

Authors explored a lion population (Panthera leo) in a savannah setting where camera traps were set up on trees. Trees provided the only shady spots available to the animals in the savannah, so this choice for camera placement imposed a strong bias on the analysis. The density estimates using daylight events were strongly overestimated compared to nighttime events.

Recently, researchers have been refining the REM model in order to fit their research environment and target species. A new REM study by Garrote et al. (2019) accounted for bias of putting cameras on roads that were used more often by the Iberian lynx (Lynx pardinus) for travel purposes. The model was successfully corrected to fit the limitations of their population.

In my study, multiple parameters need to be reconsidered to increase reliability of the results. By collaborating with Dr. Erik Noonburg, I am working on adding group spread into the model to remove the violation that animals move independently to each other, as this is not the case for lesula, which lives in social groups. The standard REM model described by Rowcliffe et al. (2008) treats individuals or groups as infinitesimally small points. If group spread (the distance between members of a group) is small relative to the radius of the camera trap detection zone, this approximation is acceptable for animals that travel in groups. In our case, the estimate of group spread is about an order of magnitude larger than r, which makes the approximation unreasonable. Incorporating group spread (s)

44 into the original derivation of the REM results should be appropriate to increase the accuracy of the lesula density estimates.

In addition, a new field study is needed to estimate day range by collecting data on the activity budget of lesula, which will take into account time spent traveling arboreally.

The lesula travel speed estimation in my study is only for ground traveling, as is the day range. My results do not account for canopy traveling. However, the traveling activity budget that I used from sister species includes traveling activity as a whole (terrestrial and arboreal). The use of camera trap “columns” in a future study would provide a means to quantitatively document the locomotor behavior exhibited by lesula and then refine its day range to increase accuracy of the REM results (Alempijevic et al., in press). Additionally,

I will add a confidence interval for each parameter in the REM model to determine a range of density values that will more likely be representative of the true density.

Ultimately, I plan on using the REM equation with the refined parameters to estimate group and individual densities of lesula in the LNP and buffer zone. It would allow me to quantitatively assess population changes over time and address the impact of hunting in the area. Finally, extrapolation of these results would allow me to estimate a total population size of C. lomamiensis for the first time in order to update the conservation status of the species through the IUCN Red List of Threatened Species.

45 CONCLUSION

My study was the first to investigate the behavioral ecology of the lesula species since its documentation in 2012. The goal was to learn more about the latest member of the Cercopithecini tribe and its place in the evolutionary history of guenons. Both C. lomamiensis and C. hamlyni are part of their own clade separated from the other

Cercopithecus species and very little data are currently available. By studying the lesula species, my project increased our understanding of the ecological diversity occurring within the guenon radiation as a whole.

Using data collected exclusively from camera traps, I successfully gained new scientific knowledge on the behavioral ecology of lesula in two different environments

(hunted versus protected forests). I was able to document its terrestriality, diurnality, birth seasonality, group size, and female philopatry/male dispersal social system. In addition, I

(1) confirmed that lesula possesses a unique ecological niche regarding locomotor behavior, but conforms with other Cercopithecus species regarding all other parameters investigated, (2) provided supporting evidence on the impact of bushmeat hunting on lesula group size, and (3) derived parameters from video footage to estimate density of lesula in a future study.

New information on the lesula species helps understand the evolutionary history and diversity of the Cercopithecus genus, but also contributes to the continuous conservation efforts of protecting the high primate endemism of the TL2 landscape.

46 REFERENCES

Alempijevic, D., Hart, J. A., Hart, T. B., & Detwiler, K. M. (in press). Local knowledge and camera traps reveal habitat preference and distribution of the endangered dryas monkey (Cercopithecus dryas) in the Lomami National Park and buffer zone, Democratic Republic of the Congo. Oryx.

Amboko, J., Hart, J. A., Silegowa, H., Bofenda, M., Mirambo, M., & Hart, T. B. (2018). Monkey abundance and hunting strategies across a protection gradient in the Lomami Landscape, D.R. Congo. 27th Congress of the International Primatology Society (IPS). Nairobi, Kenya.

Arenson, J. L., Sargis, E. J., Hart, J. A., Hart, T. B., Detwiler, K. M., & Gilbert, C. C. (2020). Skeleton morphology of the lesula (Cercopithecus lomamiensis) and the evolution of the guenon locomotor behavior. American Journal of Physical Anthropology. 1-22. DOI:101002/ajpa.24025

Bettridge, C. M. & Dunbar, R. I. M. (2013). Predation as a determinant of minimum group size in baboons. Folia Primatologica. 83, 332-352. DOI:10.1159/000339808

Beyers, R. L., Hart, J. A., Sinclair, A. R. E., Grossmann, F., Klinkenberg, B., & Dino, S. (2011). Resource wars and conflict ivory: The impact of civil conflict on elephants in the Democratic Republic of Congo – The case of the Okapi Reserve. PlosOne. 6(11). e27129. DOI:10.1371/journal.pone.0027129

Bourlière, F., Hunkeler, C., & Bertrand, M. (1970). Ecology and behaviour of Lowe’s guenon (Cercopithecus campbelli lowei) in the Ivory Coast. In J.R. Napier and P.H. Napier. (Eds.) Old World Monkey. New York Academic Press.

Bowler, M. T., Tobler, M. W., Endress, B. A., Gilmore, M. P., & Anderson, M. J. (2016). Estimating mammalian species richness and occupancy in tropical rainforest canopies with arboreal camera traps. Remote Sensing in Ecology and Conservation. 1-12. DOI:10.1002/rse2.35

Boyer-Ontl, K. M. & Pruetz, J. D. (2014). Giving the forest eyes: The benefits of using camera traps to study unhabituated Chimpanzees (Pan troglodytes verus) in Southeastern Senegal. International Journal of Primatology. 35, 881-894. DOI:10.1007/s10764-014-9783-3

47 Burton, A., Neilson, E., Moreira, D., Ladle, A., Steenweg, R., Fisher, J., …& Boutin, S. (2015). Wildlife camera trapping: A review and recommendations for linking surveys to ecological processes. Journal of Applied Ecology. 52(3), 675-685. DOI:10.1111/1365-2664.12432

Butynski, T. M. (1988). Guenon birth seasons and correlates with rainfall and food; pp 284-322. In Gauthier-Hion, A., Bourlière, F., & Gautier, J-P. (Eds.) A primate radiation: evolutionary biology of the African guenons. University Press, Cambridge, Great Britain.

Campos-Candela, A., Palmer, M., Balle, S., & Alós, J. (2018). A camera-based method for estimating absolute density in animals displaying home range behaviour. Journal of Animal Ecology. 87(3), 825-837. DOI:10.1111/1365-2656.12787

Caravaggi, A., Banks, P. B., Burton, A. C., Finlay, C. M. V., Haswell, P. M., Hayward, M. W., …& Wood, M. D. (2017). A review of camera trapping for conservation behavior research. Remote Sensing in Ecology and Conservation. 3(3), 109-122. DOI:10.1001/rse2.48

Che, N. B., Nkemnyi, F., Atem, E. T., & Giliba, R. (2017). The correlation between bushmeat harvesting and wildlife abundance in the Tofala-Mone forest corridor, Cameroon. International Journal of Conservation Science. 8(3), 465-474.

Colyn, M. & Rahm, U. (1987). Cercopithecus hamlyni kahuziensis (Primates, Cercopithecidae): une nouvelle sous-espèce de la forêt de bambous du parc national “Kahuzi-Biega” (Zaïre). Folia primatologica. 49(3-4), 203-208.

Cords, M. (1987). Forest guenons and patas monkeys: male-male competition in one- male groups; pp 98-111. In B. Smuts, D. Cheney, R. Seyfarth, R. Wrangham, & T. Struhsaker. (Eds.) Primates Societies. The University of Chicago Press, Chicago and London.

Cords, M. (1988). Mating systems of forest guenons: a preliminary review; pp 323-339. In Gauthier-Hion, A., Bourlière, F., & Gautier, J-P. (Eds.) A primate radiation: evolutionary biology of the African guenons. University Press, Cambridge, Great Britain.

Cronin, D. T, Sesink Clee, P. R., Mitchell, M. W., Bocuma Meñe, D., Fernández, D., Riaco, C., …& Gonder, M. K. (2017). Conservation strategies for understanding and combating the primate bushmeat trade on Bioko Island, Equatorial Guinea. American Journal of Primatology. 79. e22663. DOI:10.1002/apj.22663

Dehn, M. (1990). Vigilence for predators: detection and dilution effects. Behavioral Ecology and Sociobiology. 26(5), 337-342.

48 Detwiler, K. M. & Hart, J. A. (In press). Cercopithecus lomamiensis. The IUCN Red List of Threatened Species. https://www.iucnredlist.org.

Dugatkin, L. A. (2004). Principles of animal behavior (3rd ed.). Chapter 12: Antipredator Behavior. (pp 384-411). New York: W. W. Norton & Co. Print.

Easton, J., Chao, N., Mulindahabi, F., Ntare, N., Rugyerinyange, L., & Ndikunwimana, I. (2011). Status and conservation of the only population of the vulnerable owl-faced monkey Cercopithecus hamlyni in Rwanda. Oryx. 45(3), 435-438. DOI:10.1017/S0030605310001468

Evans, M. J. & Rittenhouse, T. A. G. (2018). Evaluating spatially explicit density estimates of unmarked wildlife detected by remote cameras. Journal of Applied Ecology. 55, 2565-2574. DOI:10.1111/1365-2664.13194

Fashing, P. J. & Cords, M. (2000). Diurnal primate densities and biomass in the Kakamega forest: An evaluation of census methods and a comparison with other forests. American Journal of Primatology. 50, 139-152.

Förster, S. & Cords, M. (2002). Development of mother-infant relationships and infant behavior in wild Blue monkeys; pp. 245-272. In M. E. Glenn & M. Cords. (Eds.) The guenons: Diversity and adaptation in African monkeys. New York, NY: Kluwer Academic.

Galat-Luong, A. (1975). Notes préliminaires sur l’écologie de Cercopithecus ascanius schmidti dans les environs de Bangui (R.C.A.). Terre et Vie. 122: 288-297.

Galvis, N., Link, A., & Di Fiore, A. (2014). A novel use of camera traps to study the demography and life history in wild animals: A case study of spider monkeys (Ateles belzebuth). International Journal of Primatology. 35, 908-918. DOI:10.1007/s10764- 014-9791-3

Garrote, G., Perez de Ayala, R., Alvarez, A., Martin, J. M., Ruiz, M., De Lillo, S., & Simon, M. A. (2019). Improving the random encounter model method to estimate carnivore densities using data generated by conventional camera-trap design. Oryx. DOI:10.1017/S0030605318001618

Glenn, M. E. (1997). Group size and group composition of the mona monkey (Cercopithecus mona) on the island of Grenada, West Indies. American Journal of Primatology. 43, 167-173.

Goodwin, R. & Kaplin, B. Cercopithecus hamlyni. In Rowe, N. & Myers, M. All the World’s Primates. (Eds.). www.alltheworldprimates.org. Downloaded on 13 Mar. 2020.

49 Gregory, T., Lunde, D., Zamora-Meza, H. T., & Carrasco-Rueda, F. (2015). Records of Coendou ichillus (Rodentia, Erethizontidae) from the Lower Urubamba Region of Peru. ZooKeys. 509, 109-121. DOI:10.3897/zookeys.509.9821

Hart, T. B. (2009). Bushmeat 9: A Congo chronology of bushmeat. Searching for Bonobo in Congo. https://www.bonoboincongo.com/2009/11/15/bushmeat-9-a-congo- chronology-of-bushmeat/ Hart, T. B. (2012). A new species of monkey from the center of Congo. Searching for Bonobo in Congo. https://www.bonoboincongo.com/2012/09/12/a-new-species-of- monkey-from-the-center-of-congo/

Hart, T. B. (2016). Lomami National Park: A new protected area in D. R. Congo. Searching for Bonobo in Congo. https://www.bonoboincongo.com/2016/07/13/lomami-national-park-a-new-protected- area-in-d-r-congo/

Hart, J. A., Butynski, T. M., Sarmiento, E. E., & DeJong, Y. A. (2013). Cercopithecus hamlyni Owl-faced monkey (Hamlyn’s monkey). In T. M. Butynski, J. Kingdon, & J. Kalina (Eds.), Mammals of Africa. Volume II: Primates (pp. 339-344). Bloomsbury Publishing.

Hart, J. A., Detwiler, K. M., Alempijevic, D., Lokasola, A. & Rylands, A. B. (2019). Cercopithecus dryas. The IUCN Red List of Threatened Species 2019:e.T4216A17947691. http://dx.doi.org/10.2305/IUCN.UK.2019- 1.RLTS.T4216A17947691.en. Downloaded on 4 Oct. 2019.

Hart, J. A., Detwiler, K. M., Gilbert, C. C., Burrell, A. S., Fuller, J. S., Emetshu, M., … & Tosi, A. J. (2012). Lesula: a new species of Cercopithecus monkey endemic to the Democratic Republic of Congo and implications for conservation of Congo’s Central Basin. PlosOne. 7(9), e44271. DOI:10.1371/journal.pone.0044271

Hart, T. & Hart, J. A. (2011). Breaking the bushmeat cycle in Congo: A good news story. Swara, East African Wildlife Society. 1, 16-19.

Hegerl, C., Burgess, N. D., Nielsen, M. R., Martin, E., Ciolli, M., & Rovero, F. (2017). Using camera trap data to assess the impact of bushmeat hunting on forest mammals in Tanzania. Fauna and Flora International. 51(1), 87-97. DOI:10.1017/S0030605315000836

Kautt, A., Machado-Schiaffino, G., & Meyer, A. (2018). Lessons from a natural experiment: Allopatric morphological divergence and sympatric diversification in the Midas cichlid species complex are largely influenced by ecology in a deterministic way. Evolution Letters. 2(4), 323-340. DOI:10.1002/evl3.64

Kay, R. F. & Kirk, C. (2000). Osteology evidence for the evolution of activity pattern and visual acuity in primates. American Journal of Physical Anthropology. 113, 235-262.

50 Kingdon, J. (2015). The kingdom field guide to African mammals. Second Edition. Princeton University Press, Princeton & Oxford, USA.

Kingdon, J. & Groves, C. P. (2013). Tribe Cercopithecini; pp 245-387. In Butynski, T. M., Kingdon, J. & Kalina, J. (Eds.) Mammals of Africa: Volume II: Primates. Bloomsbury Publishing, London.

Krige, P. D. & Lucas, J. W. (1975). Behavioural development of the infant Vervet monkey in a free ranging troop. Behavioural Sciences. 2(3), 151-160.

Lavoué, S. & Sullivan, J. P. (2014). Petrocephalus boboto and Petrocephalus arnegardi, two new species of African electric fish (Osteoglossomorpha, Mormyridae) from the Congo River basin. ZooKeys. 400, 43-65. DOI:10.3897/zookeys.400.6743

Lee, P. C. (1984). Early infant development and maternal care in free-ranging Vervet monkeys. Primates. 25(1), 36-47.

Loken, B., Spehar, S. & Rayadin, Y. (2013). Terrestriality in the Bornean orangutan (Pongo pygmaeus morio) and implications for their ecology and conservation. American Journal of Primatology. 75, 1129-1138. DOI:10.1002/ajp.22174

McPhee, S. (2015). A camera trap study of the cryptic, terrestrial guenon Cercopithecus lomamiensis in Central Democratic Republic of the Congo. Master thesis. Florida Atlantic University. Boca Raton, Florida.

Mittermeier, R. A., Rylands, A. B., & Wilson, D. E. (2013). Handbook of the mammals of the world. Vol. 3. Primates. Lynx Edicions, Barcelona.

Nakashima, Y., Fukasawa, K., & Samejima, H. (2017). Estimating animal density without individual recognition using information derivable exclusively from camera traps. Journal of Applied Ecology. 55, 735-744. DOI:10.1111/1365-2664.13059

Nowell, A. A. & Fletcher, A. W. (2007). Development of independence from the mother in Gorilla gorilla gorilla. International Journal of Primatology. 28, 441-455. DOI:10.1007/s10764-007-9128-6

Pazol, K. (2003). Mating in the Kakamega forest blue monkeys (Cercopithecus mitis): Does female sexual behavior function to manipulate paternity assessment? Behaviour. 140, 473-499.

Pebsworth, P. A. & LaFleur, M. (2014). Advancing primate research and conservation through the use of camera traps: Introduction to the special issue. International Journal of Primatology. 35, 825-840. DOI:10.1007/s10764-014-9802-4

Pilbrow, V. & Groves, C. (2013). Evidence for divergence in populations of Bonobos (Pan paniscus) in the Lomami-Lualaba and Kasai-Sankuru regions based on

51 preliminary analysis of craniodental variation. International Journal of Primatology. 34, 1244-1260. DOI:10.1007/s10764-013-9737-1

Refischt, J. & Koné, I. (2005). Market hunting in the Taï region, Côte d’Ivoire, and implications for monkey populations. International Journal of Primatology. 26, 7874- 7876.

Rowcliffe, J. M. (2017). Key frontiers in camera trapping research. Remote Sensing in Ecology and Conservation. 3(3), 107-108. DOI:10.1002/rse2.65

Rowcliffe, J. M., Field, J., Turvey, S. T., & Carbone, C. (2008). Estimating animal density using camera traps without the need for individual recognition. Journal of Applied Ecology. 45, 1228-1236. DOI:10.1111/j.1365-2664.2008.01473.x

Rowcliffe, J. M., Jansen, P. A., Kays, R., Kranstauber, B., & Carbone, C. (2016). Wildlife speed cameras: Measuring animal travel speed and day range using camera traps. Remote Sensing in Ecology and Conservation. 84-94. DOI:10.1002/rse2.17

Rowe, N. & Myers, M. (2013). All the world’s primates. Primate Conservation Inc.

Rowe, N. & Myers, M. (2016). All the world’s primates. Pogonias Press, Charlestown, RI. Print.

Scotson, L., Johnston, L. R., Iannarilli, F., Wearn, O. R, Mohd-Azlan, J., Wong, W. M., …& Fieberg, J. (2017). Best practices and software for the management and sharing of camera trap data for small and large scales studies. Remote Sensing in Ecology and Conservation. 3(3), 158-172.

Stanton, D. W. G., Hart, J., Vosper, A., Kümpel, N. F., Wang, J., Ewen, J. G., & Bruford, W. (2016). Non-invasive genetic identification confirms the presence of the endangered okapi Okapia johnstoni south-west of the Congo River. Oryx. 50(1), 134- 137. DOI:10.1017/S0030605314000593

Strier, K. B. (2017). Primate Behavioral Ecology. 5th ed. Routledge, London and New York, NY.

Stroud, J. & Losos, J. (2016). Ecological opportunity and adaptive radiation. Annual Review of Ecology, Evolution, and Systematics. 47(1), 507-532. DOI:10.1146/annurev-ecolsys-121415-032254

Tagg, N., Kuenbou, J. K., Laméris, D. W., Kamkeng Meigang, F. M., Kekeunou, S., Epanda, M. A., …& Willie, J. (2020). Long-term trends in wildlife community structure and functional diversity in a village hunting zone in southeast Cameroon. Biodiversity and Conservation. 29, 571-590. https://doi.org/10.1007/s10531-019- 01899-1

52 Thomas, S. C. (1991). Population densities and patterns of habitat use among anthropoid primates of the Ituri Forest, Zaïre. Biotropica. 23(1), 68-83.

Thorpe, W. H. (1963). Learning and Instinct in Animals. London: Methuen.

Tosi, A. J., Detwiler, K. M. & Disotell, T. R. (2005). X-chromosomal window into the evolutionary history of the guenons (Primates: Cercopithecini). Molecular Phylogenetics and Evolution. 36, 58-66.

Tosi, A. J., Melnick, D. J., & Disotell, T. R. (2004). Sex chromosome phylogenetics indicate a single transition to terrestriality in the guenons (tribe Cercopithecini). Journal of Human Evolution. 46, 223-237. DOI:10.1016/j.jhevol.2003.11.006

Tutin, C. E. G. & White, L. (1999). The recent evolutionary past of primate communities, likely environmental impacts during the past three millennia; pp 220-236. In Fleagle, J. G., Janson, C. H. & Reed, K. E. (Eds.) Primate Communities. Cambridge University Press, Cambridge.

Ukizintambara, T. Allochrocebus lhoesti. In Rowe, N. & Myers, M. All the World’s Primates. (Eds.). www.alltheworldprimates.org. Downloaded on 13 Mar. 2020.

Williamson, E. A. & Feistner, A. T. C. (2003). Habituation primates: Processes, techniques, variables and ethics. In Setchell, J. M. & Curtis, D. J. (Eds.) Field and laboratory methods in primatology: A practical guide. Cambridge University Press.

Wrangham, R., Gittleman, J., & Chapman, C. (1993). Constraints on group size in primates and carnivores: population density and day-range as assays of exploitation competition. Behavioral Ecology and Sociobiology. 32(3), 199-209.

Xing, J., Wang, H., Zhang, Y., Ray, D. A., Tosi, A. J., Disotell, T. R., & Batzer, M. A. (2007). A mobile element-based evolutionary history of guenons (tribe Cercopithecini). BMC Biology. 5(5). DOI:10.1186/1741-7007-5-5

Yoder, J., Clancey, E., Des Roches, S., Eastman, J., Gentry, L., Godsoe W., …& Harmon, L. (2010). Ecological opportunity and the origin of adaptive radiation. Journal of Evolutionary Biology. 23(8), 1581-1596. DOI:10.1111/j.1420- 9101.2010.02029.x

Ziegler, S., Fa, J. E., Wohlfart, C., Streit, B., Jacob, S., & Wegmann, M. (2016). Mapping bushmeat hunting pressure in Central Africa. Biotropica. 48(3), 405-412. DOI:10.1111/btp.12286

53