Endangered Zanzibar Red Colobus Piliocolobus Kirkii

U N I V E R S I T Y O F C O P E N H A G E 

FACULTY OF SCIENCE 

Master’s Thesis 

L Æ R K E N Y K J Æ R J O H A N S E N 

A Conservation Re-Assessment of the 

Endangered Zanzibar Red Colobus 

Piliocolobus Kirkii 

S C I E N T I F I C A D V I S O R : Prof. Neil Burgess 

CO- A D V I S O R : D r . Katarzyna Nowak 

S U B M I T T E D : 4 th November 2016

U N I V E R S I T Y O F C O P E N H A G E N 

Faculty: 

Institute: 

Name of department: 

Author: 

Title / Subtitle: 

Scientific advisor: 

Co – advisor: 

Faculty of Science 

Institute of Biology 

Center for Macroecology, Evolution and Climate 

Lærke Nykjær Johansen 

A Conservation Re-Assessment of the Endangered Zanzibar Red 

Colobus Piliocolobus kirkii 

Neil Burgess 

Katarzyna Nowak 

Submitted: 04 th November 2016 

Entitled pointes: 

Duration: 

Length: 

45 ECTS 

9 Months 

74 Pages 

03 November 2016 

Lærke Nykjær Johansen 

Front page illustration © (Kingdon & Happold 2013)


LÆR KE NYKJ ÆR JOH ANS EN 

According to the Chinese zodiac calendar, 2016 is the 

year of the Red Fire Monkey. 

A year were people born in the year of the snake, will 

have an exceptional connection to the monkey. 

Lærke N.J. (snake)


LÆR KE NYKJ ÆR J OH AN SEN 

PREFACE 

This thesis is the result of a 9-month Master’s project at the Center of Macroecology, Evolution 

and Climate at University of Copenhagen, Denmark. This project has been supervised by 

Professor Neil David Burgess Danish Natural History Museum, Copenhagen University 

Denmark, and co-supervisor Doctor Katarzyna Nowak, AAAS, USA. The fieldwork conducted 

in Kiwengwa – Pongwe Forest Reserve was supported by the Department of Forestry and Non- 

Renewable Natural Resources, Zanzibar. Research permit was issued by the Ministry of State 

through the Second Vice - Presidential Office of Zanzibar, Stonetown Zanzibar. 

In collaboration with my supervisors, I have been building on this project since March 2015. It 

has resulted in two trips to Zanzibar, in August 2015 and January-April 2016, subsequent four 

months’ total, spent on Zanzibar and mainland Tanganyika. 

With this thesis, I will pursue to embrace conservation of an endangered endemic species and 

what role habitat preservation has in the success of this.



ACKNOWLEDGMENTS 

First off, I want to start by thanking the people that assisted and supported me on 

Zanzibar. At Department of Forestry and Non-Renewable Natural Resources (DFNRNR) I want 

to thank the director, Mr. Sheha Hamdan Kassim for blessing me with the support from 

DFNRNR, and Mr. Hamza Madeweya for helping me with the applications for the presidential 

office. At Family Beach Bungalows in Kiwengwa, my heart goes out to Mr. Juma and his 

nephew Mr. Simai, for taking good care of me and providing me with a safe and welcoming 

place to live. Thank you for always helping me and openly answering my countless questions to 

Zanzibarian life. I want to give a big thanks to Mr. Tahir Haji for the great work with 

identification of almost 7000 trees, we did this in only five days, five very long and very hot 

days. Last but not least, my greatest thanks goes to my always loyal field assistant Mr. Mtumwa 

Simai. Thank you for following me for what felt like countless hours walking in the burning sun, 

so slow that you feel like you are losing your mind. Thank you for providing me a bike and 

spontaneously bringing me fresh fish and squids! Thank you for welcoming me in your home, to 

the fantastic juices and lunch your wife Miriam made! I only wish the best for you and your 

family. 

I am thankful for all the dear friends I made while on Zbar. In my eyes, you are not ranked, for 

this I am grateful for having crossed paths with every one of you. 

The greatest gratitudes I send to my scientific supervisors Neil, and Kate. Thank you, Neil, for 

starting this project with me, for being the minds behind it, and thanks to Kate for your 

knowledge about colobus and help on Zanzibar, I would not have been able to complete my field 

work without it. 

At Copenhagen University, my thanks go to CMEC, every one there, the fourth flour and 

“kagestuen” - no coffee, no cake, no aquarium = no thesis. 

I also sincerely thank the grants I have been awarded, Den A.P. Møllerske Støttefond, Det 

Saxild’ske Familiefond, Dr. Bøje Benzons Støttefond and ‘Nykjærs Familien Legatet’, without 

this economical support this project would never have been a possibility. Last but not least, my 

appreciation go to my father Olav Nykjær, thank you for being my extra external advisor, for 

guidance and counseling, for you and Thea visiting me on Zanzibar, thank you for helping me. 

To everyone else who has been involved in this project, to my friends, to my family 

-Ya Moya Asante Sana



ABSTRACT 

Aim: Deforestation and habitat degradation due to forest resource demand from growing human 

settlements, imposes a severe threat to conservation of endangered species worldwide. To best 

conserve our remaining wildlife more and more land becomes protected to further inhibit 

degradation. To evaluate the effects of protection management for an endangered species, we reassess 

the conservation status of the endemic Zanzibar red colobus (Procolobus kirkii), in a 

government managed forest reserve, prior and past gaining protection status. 

Location: All fieldwork was conducted in Kiwengwe-Pongwe Forest Reserve, Unguja Island 

Zanzibar, United Republic of Tanzania. 

Methods: Populations of P. kirkii were censused using line transects sampling of three transects. 

Habitat was sampled along the same transects by measuring 35 5 x 50 m vegetation plots. 

Results: I found that in total the area sampled now likely contained a higher density of colobus 

than before gazettement. Groups where encountered more often, but were in average slightly 

smaller. The habitat had undergone a radical degradation. The density of trees ≥ 2,5 m in height 

had decreased 42 % and 58 species found in 2004 were absent in in the same area sampled in 2016. 

The deforestation, species loss and human disturbance was clearly larger towards the reserve rim, 

closer to growing urban settlements. 

Main Conclusions: We found that the habitat degradation had possibly caused a population 

compression of P. kirkii. This increasing animal density in the center of the reserve, furthest from 

human disturbance from surrounding rural settlements. It is an unfortunate reality for the 

endangered species, emphasizing the need of engaging local community in conservation 

management. Conservation is a sociological matter, requiring implementation and support by the 

local community, for even the best meant management plans to have a sufficient effect. More focus 

must be turned towards the fulfilling of management goals and follow-ups on issued management 

plans, to insure implementation and effective conservation. 

i



RESUMÉ 

Vores voksende humane population truer tilværelsen for flere arter af vilde primater. Især den 

øgede urbanisering af det afrikanske kontinent har konsekvenser for flere arter heriblandt den 

Zanzibar endemiske røde colobus abe, Piliocolobus kirkii. Den årlige menneskelige 

populationstilvækst på 5% og en lokalbefolkning hvor over 80 % er afhængige af ressourcer fra 

skoven, gør at der hvert år gradvist forsvinder mere af de fragmenterede skove, Zanzibar colobus 

aben hovedsageligt lever i. 

Med denne afhandling undersøger jeg, ved transekt populationsoptælling og vegetationsanalyse, 

hvorvidt fredning af et af de mest betydningsfulde skovområde, Kiwengwa-Pongwe reservatet, har 

haft nogen effekt på den deri boende population, samt på den overordnede tilstand af skoven. Jeg 

gør dette ved at sammenligne data fra 2004, før fredning i 2007, med data indsamlet vinteren 2016. 

Jeg fandt at der, i det afgrænsede prøveområde, med al sandsynlighed var en større tæthed af aber 

i 2016, dog med en hvis usikkerhed. Det viste sig dog at, aberne generelt bevægede sig i mindre 

grupper og at der var signifikant forskelle i hvor aberne befandt sig i højere densiteter både mellem 

de to år og inden for hvert år. Vegetationsanalyse viste at, densiteten af træer var faldet signifikant 

ved alle tre transekter. Der var sket et skift i, at der nu blev fældet signifikant flere træer ved 

transekt B det nordlige transekt, samt signifikant færre træer ved transekt K3, det sydlige transekt. 

Af de 119 arter fundet under vegetations analyse i 2004, blev 58 arter ikke fundet ved samme 

analyse i 2016 hvor totalt kun 75 arter blev fundet. Der var en klar sammenhæng mellem hvor der 

var en større menneske aktivitet i skoven, med hvor flere arter var forsvundet, ved brug af flere 

parametre som proxy for menneskelig aktivitet. 

Ud fra den forringede tilstand af skoven, skiftet i hvor der flest aber blev observeret, samt den 

formindskede observerede gruppestørrelse antager jeg, at det forøgede estimerede totale antal aber 

skyldes populations kompression, hvor individer fra andre mere forstyrret dele af skoven er 

indvandrede til dette undersøgte område, resulterende i en kunstig forøgelse af populationstallet. 

Dette er en kedelig realitet for den truede aber og viser hvor vigtig, inkludering af 

lokalbefolkningens behov og tilgængelige ressourcer er, i naturbevaring og forvaltning. 

ii



iv



TABLE OF CONTENT 

LIST OF FIGURES ......................................................................................................................................................... V 

LIST OF TABLES .......................................................................................................................................................... VI 

LIST OF ABBREVIATIONS AND ACRONYMS................................................................................................................ VII 

INTRODUCTION .......................................................................................................................................................... 1 

BACKGROUND ............................................................................................................................................................ 2 

PILIOCOLOBUS GENUS .......................................................................................................................................................... 3 

PILIOCOLOBUS KIRKII ............................................................................................................................................................ 5 

Habitat management ................................................................................................................................................ 6 

Main threats of Piliocolobus kirkii ............................................................................................................................. 9 

Consequences of endangerment ............................................................................................................................... 9 

METHODS ..................................................................................................................................................................12 

Study Site ................................................................................................................................................................. 12 

Monkey census ........................................................................................................................................................ 13 

Disturbance ............................................................................................................................................................. 14 

Vegetation sampling ............................................................................................................................................... 15 

DATA ANALYSIS ................................................................................................................................................................. 16 

Distance sampling (DS)............................................................................................................................................ 16 

Whiteside method (WM)......................................................................................................................................... 17 

Population structure ................................................................................................................................................ 18 

Vegetation analysis ................................................................................................................................................. 18 

Correlations ............................................................................................................................................................. 18 

Mapping .................................................................................................................................................................. 18 

RESULTS .....................................................................................................................................................................19 

Population analysis ................................................................................................................................................. 19 

Encounter rate ......................................................................................................................................................... 20 

Group size means .................................................................................................................................................... 21 

VEGETATION ANALYSIS ....................................................................................................................................................... 22 

Species composition and abundance ...................................................................................................................... 23 

CORRELATIONS ................................................................................................................................................................. 26 

iii



HUMAN DISTURBANCE ....................................................................................................................................................... 28 

Linear Regression .................................................................................................................................................... 28 

DISCUSSION ...............................................................................................................................................................31 

DISTANCE SAMPLING ......................................................................................................................................................... 31 

GROUP SIZE ..................................................................................................................................................................... 33 

VEGETATIVE OUTCOMES ..................................................................................................................................................... 35 

CONSEQUENTIAL RESULTS ................................................................................................................................................... 37 

MANAGEMENT STATUS ...................................................................................................................................................... 39 

CONCLUDING REMARKS ............................................................................................................................................40 

FURTHER RESEARCH ..................................................................................................................................................41 

REFERENCES ..............................................................................................................................................................42 

APPENDIX ..................................................................................................................................................................47 

Appendix 1 – Study site map ................................................................................................................................... 48 

Appendix 2 - Colobus and food species maps .......................................................................................................... 49 

Appendix 3 – Detection functions g(x) ..................................................................................................................... 51 

Appendix 4 – Cluster size distribution...................................................................................................................... 53 

Appendix 5 - Statistics ............................................................................................................................................. 54 

Appendix 6 – Plot specific species decline ............................................................................................................... 56 

Appendix 7 - Species list .......................................................................................................................................... 57 

Appendix 8 - Correlations ........................................................................................................................................ 59 

iv


LÆR KE NYKJ ÆR JOH AN S EN


LÆR KE NYKJ ÆR JOH AN S EN 

LIST OF FIGURES 

Figure 1 The location of The Eastern Arc and Coastal Forest of Tanzania/Kenya hotspot………...2 

Figure 2 Distribution of the 18 presently recognized taxa of red colobus monkeys………………..3 

Figure 3 Zanzibar red colobus.......…………………………………………………………..…….5 

Figure 4 Map of the different protected/unprotected forests of Unguja Island…………...…..……8 

Figure 5 Chimpanzee with colobus…………………………...…………………………….……11 

Figure 6 Location of transects…………………...………………………………………….……12 

Figure 7 Examples of human disturbance observed under transect walks…………………...…..15 

Figure 8 Techniques for line transect observation………………………………………………..17 

Figure 9 Encounter rates………………………………………………………………………….20 

Figure 10 Group size……………………………………………………………………………..22 

Figure 11 Results of vegetation analysis...………………………………………………………..25 

Figure 12 Correlation map………………………………………………………………………..26 

Figure 13 Correlation map significance levels…………………………………………..…….…27 

Figure 14 Frequency of human disturbances recorded along transect………………………….. 28 

Figure 15 Linear regression on located observed colobus and human disturbances……...……….29 

Figure 16 Linear relationships of parameters proxy for human disturbances……………..………30 

Figure 17 Map of area surrounding The Kiwengwa – Pongwe Forest Reserve…………………...48 

Figure 18 Location of colobus encounters and group size……………………………………….49 

Figure 19 Proportion of colobus food species within each vegetation plot………………………50 

Figure 20 2004 observation distances and detection function g(x) Distance sampling method…..51 

Figure 21 2016 observation distances and detection function g(x) Distance sampling method…51 

Figure 22 2004 observation distances and detection function g(x) Whiteside method……………52 

Figure 23 2016 observation distances and detection function g(x) Whiteside method……………52 

Figure 24 Frequencies and cumulative frequencies distribution of group observations....…...…53 

Figure 25 Species lost per plot.......................................................................................................56 

v



LIST OF TABLES 

Table 1 Population estimates from DISTANCE sampling ................................................................ 19 

Table 2 Population estimates from Whiteside method ...................................................................... 19 

Table 3 Location of observation ........................................................................................................ 21 

Table 4 Pared students t-test for paired samples on vegetative differences ....................................... 54 

Table 5 Pared students t-test for paired samles on wood harvest at different transects. .................... 54 

Table 6 Kruskal – Wallis and Dunn’s multiple comparison test summarized. .................................. 55 

Tabel 7 Correlation analysis statistical parameters ............................................................................ 59 

Table 8 List of all positive correlations ............................................................................................. 60 

Tabel 9 List of all negative correlations ............................................................................................. 61 

vi



LIST OF ABBREVIATIONS AND ACRONYMS 

AOD Animal-to-Observer Distance 

AIC Akaike Information Criterion 

CF Community Forest 

CFMG Community Forest Management Groups 

CITES the Convention on International Trade in Endangered Species 

CoFMA Community Forest Management Agreement 

DBH Diameter at breast height 

DFNRNR Department of Non-renewable Natural Resources 

DS Distance Sampling 

EACF Eastern Arc Mountains and Coastal Forests of Tanzania a nd Kenya 

Hotspot 

ER Encounter Rate 

FR Forest Reserve 

ICDP Intergraded Conservation Development Projects 

JCNP Jozani - Chwaka Bay National Park 

KP Kiwengwa – Pongwe Forest Reserve 

KPFR Kiwengwa - Pongwe Forest Reserve 

NGO Non-Governmental Organization 

NT Near Threatened 

PA Protected Area 

PD Perpendicular Distance 

VCC Village Conservation Councils 

WM Whiteside Method 

vii



INTRODUCTION 

Of the 25 wild primates listed as the world’s most endangered, 10 are from central Africa, with 

ome genera being highly represented on this list(Schwitzer et al. 2015). In the case of the red 

colobus genus (Piliocolobus sp.) several species are found in populations with less than 5000 

individuals remaining. The Piliocolobus genus holds many endemic species often restricted to very 

small patchy habitats, opposing a large threat of extinction to several of these unique species 

(Struhsaker 2005). Piliocolobus kirkii (commonly known as Zanzibar red colobus) is endemic to 

Zanzibar and threatened of extinction by factors imposed by a rapidly increasing human population 

(Struhsaker & Siex 2016). Zanzibar holds one national park where the occurrence of P. kirkii is 

well studied and protected. But other, less prosperous, protected areas lack the same engagement 

and follow-up on implementation and efforts in conservation management, despite near equivalent 

importance for Zanzibar wildlife conservation. 

In the period of 2004-2005 Dr. Katarzyna Nowak studied the behavioral and demographic 

flexibility of P. kirkii in Kiwengwa – Pongwe (Nowak 2007). Since then the Kiwengwa – Pongwe 

forest has gained status as forest reserve. The main purpose of this study is, to investigate how/if 

the status as protected area has had an effect on the population of P. kirkii and it’s habitat in 

Kiwengwa-Pongwe Forest Reserve. 

I will do this by: 

1) Comparing population and habitat data prior to protection of Kiwengwa-Pongwe Forest 

Reserve, with data collected in 2016, almost 10 years past gazettement. I will investigate 

both between- and within sampling year variations. 

2) Investigating influences on habitat and population, with proxies for human disturbance, 

to assess the human involvement in the conservation of this species. 

3) Assess the possible correlations between investigated parameters, to pursue a qualified 

estimation and overall picture, of the conservation status of this endangered species and 

the forest reserve, to understand the effectiveness of conservation management in a 

developing country. 

1



BACKGROUND 

The fragmented forests of Zanzibar are part of a strip of costal forest mosaics, from southern 

Somalia to Mozambique, known as The Eastern Arc and Coastal Forest of Tanzania/Kenya 

hotspot (EACF) (Burgess et al. 1998). This area is one of 25 worldwide biodiversity hotspots 

which are characterized by being, places of conservation top priority, for having exceptional 

concentrations of endemic species, undergoing exceptional habitat loss (Myers et al. 2000). 

The EACF hotspot is by far the hotspot with the highest species to area ratio, both concerning 

endemic plants and vertebrate species, and is among the top eight hottest hotspots in terms of five 

priority factors: no. endemic plants, no. endemic vertebrates, endemic plats to area ratio, endemic 

vertebrates to area ratio and remaining primary vegetation as % of original extent (Myers et al. 

2000). The Zanzibar archipelago is included in the hotspot due to high levels of strictly endemic 

plans, butterflies, bird, and the endangered primate, the Zanzibar Red Colobus, Piliocolobus kirkii 

(Gray 1868) (Burgess et al. 1998). 

Figure 1 The location of The Eastern Arc and Coastal Forest of Tanzania/Kenya hotspot. The red line 

shows the area considered within the hotspot. Map modified from: (Gereau et al. 2016; Gaba 2010). 

2



PILIOCOLOBUS GENUS 

The Zanzibar red colobus is endemic to the Ugunja, Uzi and Vundwe Islands of the Zanzibar 

archipelago. It is of the African genus of colobi monkeys known as red colobus (Piliocolobus). 

They are of the old-world monkey family (Cercopithecidae), known for their classic monkey 

looks, with long tails, limbs and functional hands and feet (Groves 2007). All African colobuses 

can be recognized by having their thumbs totally reduced or apparent as small stumps, as they are 

on the Zanzibar red colobus (Struhsaker 1975; Groves 2007). 

All Piliocolobus species are distributed around equatorial Africa from Senegal to Zanzibar, with 

species ranges being allopatrically divided (with the exception of a putative hybrid zone in central 

African region) (Struhsaker 1975; Struhsaker 2005; Davies & Oates 1994; Groves 2007; Oates & 

Ting 2015). 

Figure 2 Distribution of the 18 presently recognized taxa of red colobus monkeys. 

1: P. temminckii, 2: P. badius, 3: P. waldroni, 4: P. epieni, 5: P. pennantii, 6: P. preussi, 7: P. bouvieri, 8: 

P. tholloni, 9: P. parmientieri, 10: P. lulindicus, 11: P. foai, 12: P. oustaleti, 13: P. langi, 14: P. 

semlikiensis, 15: P. tephrosceles, 16: P. rufomitratus, 17: P. gordonorum, 18: P. kirkii. ‘H’ is the putative 

hybrid population in the eastern Democratic Republic of Congo. Map: Oates & Ting 2015. 

Taxonomy of Piliocolobus species has undergone several changes over time. It has over the last 

40 years changed several times ranging from olive, red and black-and-white colobus in one genus, 

red colobus only holding one taxa, to the current recognized deviation into three separate genera, 

with 18 taxa of red colobus (Grubb et al. 2003; Oates & Ting 2015). Several of these taxa are 

3



restricted to very small patchy habitats (see Figure 2) (Oates & Ting 2015; Groves 2007). In 

general, the colobus taxonomy is somewhat a gray area severely lacking a consensus in the field 

of genus/species classification especially for the red colobus (N. Ting 2016, personal comment, 2 

November, e-mail correspondence). Six taxa of Piliocolobus have repeatedly been included in the 

IUCN SSC Primate Specialist Group’s lists of the world 25 most endangered primates from 2000 

- 2016. (Mittermeier et al. 2007; Mittermeier et al. 2009; Mittermeier et al. 2012; Schwitzer et al. 

2014; Schwitzer et al. 2015). 

The Miss Waldron’s red colobus (Piliocolobus badius waldroni) was by 2000 already announced 

extinct (Struhsaker 2005; Mittermeier et al. 2007). In 2007, IUCN added three Piliocolobus to 

their list of top endangered species from 2006 – 2008 (Mittermeier et al. 2007). A total of three 

Piliocolobus species are ranked ‘critically endangered’, seven species ranked ‘endangered’ and 

two species as ‘near threatened’ (IUCN 2016). Only one species, the Oustalet’s Red Colobus 

(Proclobus rufomitratus oustaleti) is fairly common, illustrating a rapid decline of several 

Piliocolobus species within very few years (Mittermeier et al. 2009). The IUCN primate specialist 

group underline the importance of bringing more focus to this genus, as they are in urgent need of 

attention from conservationist and researchers: 

“It is significant that there are three red colobus monkeys on the 2006 – 2008 list — there could 

(should) undoubtedly be more… need for further research and urgent conservation measures for 

the entire genus” (Mittermeier et al. 2007). 

Struhsaker (2005) investigated the conservation status of all endangered red colobus and 

concluded that hunting, habitat degradation, fragmentation and loss, and possible intrinsic factors 

following as an aftermath, are the greatest risks to survival of red colobus monkeys (Struhsaker 

2005). 

Piliocolobus are in general known to by a rather shy, arboreal living species normally habituating 

tropical and lowland forests (Struhsaker 1975; Davies & Oates 1994). Though the colobus is most 

commonly restricted to wooded habitats, the different species have shown a wide variety of 

adaptation to other habitats (Davies & Oates 1994). They have been found to also inhabit other 

less forest like habitats and more “open habitat”, like gallery forests with interrupted canopy, and 

even wooded savannahs, mangrove swamps and farmlands (Davies & Oates 1994; Struhsaker 

1975; Galat-Luong & Galat 2005). 

4



PILIOCOLOBUS KIRKII 

Being endemic to Zanzibar separated from mainland Tanzania (Tanganyika) by the approximately 

40 km wide Zanzibar Channel, P. kirkii has a restricted ability of deviations in distribution range. 

P. kirkii can be distinguished from its nearest relatives, the Udzungwa red colobus (P. 

gordonorum) by a distinct pelage color and pattern, the slightly different acoustics of male calls, 

and reduced size in accordance to the effects of the island rule, on island insular mammals (Nowak 

et al. 2008; Groves 2007). 

Figure 3 Zanzibar red colobus, Piliocolobus kirkii, adult male photographed in Jozani – Chawaka bay 

National park. They can be recognized by their characteristic chestnut red backside, crown and 

exceptional long tale with color lightened towards tip. They have white head, limbs and ventral side, with 

black face, shoulder region, lower part of arms and legs, hands and feet. Males can be distinguished by 

their brooder skulls and slight sexual dimorphism. The shoulder area and face is lined with long white 

hairs sometimes resembling the classic look of a mad scientist. Photo: Lærke Nykjær Johansen. 

5



The latest estimations declared less than 2000 individuals remaining, with population trends still 

declining (Struhsaker & Siex 2016). This should qualify P. kirkii as one of Tanzania’s primates of 

greatest conservatory concern (Davenport et al. 2013; Struhsaker 2005). 

The highest numbers of P. kirkii is found in and around the combined tropical ground water, coral 

rag and mangrove forests of Jozani - Chwaka Bay National Park (abbreviated JCNP) (see Figure 

4)(Struhsaker & Siex 2016). The Kiwengwa - Pongwe Forest Reserve (abbreviated KPFR or KP) 

is the second most important forest area for sustainable colobus populations. It is the second largest 

continuous forest area on the island and simultaneously the northern border of their distribution 

range. 

South of Jozani - Chwaka Bay National Park the colobus populations inhabit both protected and 

unprotected species-supportable habitat mosaics, scattered to the Kungwi Community Forest, 

secondary forests, shrubs, shambas (shambas are areas of agricultural purpose) and the mangrove 

swamp forests of Uzi and Vendwe Island (Siex & Struhsaker 1999b; Siex & Struhsaker 1999a; 

Nowak et al. 2009). 

The 2013 list of Priority Primate Areas mentions both Jozani - Chwaka Bay National Park, 

Kiwengwa – Pongwe Forest Reserve and the unprotected/unmanaged forest and mangrove 

swamps of Uzi and Vundwe Islands, as to be of special interest for the protection of the endangered 

Zanzibar red colobus (Davenport et al. 2013). These forests are under different levels of 

management, whereof no official management plans for Uzi and Vundwe Islands are currently 

present. The area has several times been proposed as an area worthy of gazetting and is highly 

threatened by the nearby growing human settlements. (Nowak 2013; Nowak & Lee 2013; 

Davenport et al. 2013; Nowak et al. 2009). 

Habitat management 

Zanzibar has four forest types listed after level of protection; National park (NP), Forest Reserve 

(FR), Community forest (CF) and Unprotected forest / plantations (Figure 4). 

Largest is Jozani – Chwaka Bay National Park, a 50 km 2 area protected in 2004 and managed by 

Ministry of Agriculture through the Department of Non- Renewable Natural Resources 

(DFNRNR) (formerly known as Department of Commercial Crops Fruit-trees and Forests 

(DCCFF)) (Ministry of Natural Resources and Tourism 2014). Conservation management on 

Zanzibar, is based on the holistic community integrating conservation approach, known as the 

6



“New Conservation Debate” (Minteer & Miller 2011). JCNP is surrounded by a buffer zone where 

villagers and farmers from the surrounding nine villages, are allowed to collect their needed natural 

resources, and continue agricultural farming, managed by local community based natural resource 

management comities. This approach to allow locals a sustainable use of needed natural resources 

surrounding the national park, and to ensure the best possible conservation of endangered flora 

and fauna biodiversity (Saunders 2011). Within the boundaries of the national park there is 

complete protection of all flora and fauna and no resource extraction of any kind is allowed 

(hunting, charcoal and limestone excavation, timber and firewood collection and so on). 

Forest reserves as Kiwengwa – Pongwe Forest Reserve are protected areas of special interest due 

to conservation, biodiversity or other interests, and are also under government management. They 

have no entrance limitation; you are allowed to use the forest for recreational purposes and for 

extraction of natural resources in agreement with the local community council. 

The smaller community forests fragment on the southern part of Unguja island, are based on the 

ideas behind new generation alternative conservation approach called Integrated Conservation 

Development Projects (ICDP). They try to integrate local communities in conservation of their 

natural surroundings, having a developmental and beneficial payoff for the community. An 

agreement known as the Community Forest Management Agreement (CoFMA) between the 

government, NGO’s of interest and the local villagers, establishing local community based 

organizations (Community Forest Management Groups (CFMG)) and Village Conservation 

Councils (VCC)) responsible for forest- and natural resource management. This gives the local 

community the inclusive rights to the forest management, forest resource utilization and shared 

benefits accrued from forest resources at community level. Simultaneously fulfilling conservation 

goals of NGO’s, conservation advocates and others interest groups (Rabe & Saunders 2014; 

Hassan & Said 2011; Ministry of Natural Resources and Tourism 2014). 

7



Figure 4 Map of the different protected/unprotected forests of Unguja Island. The largest population of 

Zanzibar red colobus is found in JCBNP followed by KPFR, which is also the northern barrier of their 

distribution range. To the south colobuses live in patchy forest fragment, and other habitat types outside 

of government protection. The government manages Jozani – Chwaka Bay National Park (JCBNP), 

situated in the center of the island and five other protected areas ranging from JCBNP and northwards. 

The patches of community managed forest areas south of JCBNP are managed by local Community 

Forest Management Groups (CFMG) and Village Conservation Councils (VCC). Map: Wildlife 

Conservation Society (WCS). 

8



Main threats of Piliocolobus kirkii 

Loss of habitat and habitat degradation is recognized as the greatest cause to the decreasing P. 

Kirkii population (Siex and Struhsaker 2016). 

The majority of Zanzibar’s human population still live a somewhat simplified lifestyle, dependent 

on forest products (Siex 2011). Up to 92% being depend on wood- or charcoal fires as only source 

of energy for cooking (National Bureau of Statistics 2014). This is estimated to cause a yearly 

demand of fuelwood exceeding 1,5 million m 3 , triggering an estimated yearly over harvest of wood 

approximately 800.000 m 3 (Ministry of Natural Resources and Tourism 2014). These estimates 

are based solely on demands for fuelwood, not including the wooden furniture industry and wood 

for housing constructions etc., which certainly contribute a great deal, as the majority of Zanzibar’s 

human population live in houses made using a wooden frame or palm thatch roofs on wooden 

roofing beams. In addition to use of trees for fuel, houses and furniture it has been estimated that 

more than 500 ha coral rag forest was cleared to make room for agricultural fields in 2007 alone 

(Ministry of Natural Resources and Tourism 2014). With a yearly population growth of ~3 % and 

a 2% yearly immigration rate, the agricultural needs are likely also steadily increasing every year 

(National Bureau of Statistics 2014). During interviews of local civilians in 2011 over 66 % of 

respondents answered that the rate of deforestation in their area was high / very high and that the 

majority of the needed forest products in their community came from government protected areas 

(Hassan & Said 2011). 

Consequences of endangerment 

It is widely believed that the Zanzibar red colobus has embraced using secondary habitat types, as 

a necessity, due to the lack of primary habitat, or due to habitat insufficiency (Nowak 2013; Nowak 

& Lee 2013). Some populations have been shown to spend up to 85% their time in the mangrove 

forest (Rhizophoraceae sp.) as a place of refugee from the frequent human disturbances and forest 

degradation of the adjacent coral rag forest, where the populations previously roamed (Nowak 

2013). 

Like other red colobus under the pressure of habitat disruption, the P. kirkii has also adapted it’s 

folivorous diet to include secondary plant species and some fruits, possible because of the 

insufficient amounts of favored foods available (Nowak 2008; Siex & Struhsaker 1999a). 

Piliocolobus normally get their supply of water through their foliage diet, but the embrace of 

mangrove leaf with a higher salinity, to their diet has resulted in a frequent water drinking behavior 

9



(Nowak 2008). Groups with more primary forest available, use mangrove forest less for foraging 

and equally drink water less frequently (Nowak 2008). This emphasizing that the consumption of 

mangrove leaves and therefore necessity of drinking water, not being a favored food source but a 

bearable adaption to the refuge life. 

Groups living in close proximity to human settlements, in shambas or near plantations, have 

acquired the behavior of eating charcoal, due to phenolic acids and other organic compounds in 

leaves of species non-native to Zanzibar (Struhsaker et al. 1997). The charcoal apparently absorbs 

these compounds in exotic species like mango (Mangifera sp.), compounds which are otherwise 

toxic in higher concentrations (Struhsaker et al. 1997). 

Adaptations towards a change in habitat is also seen in their social structure and foraging strategies. 

Piliocolobus generally live, forage and travel in large multi-male/female troops of 15-80 

individuals, but in habitats of poorer quality and fewer food species present, these large groups are 

not sustainable (Struhsaker 1975; Struhsaker et al. 2004). P. kirkii and other colobuses have coped 

with this by converting their social structure to a fission-fusion behavior, where the main group 

splits into two or more subgroups when foraging (Struhsaker 2000; Struhsaker et al. 2004; Siex & 

Struhsaker 1999b; Galat-Luong & Galat 2005). This intergroup fragmentation happens as a 

reaction to a low density or patchy distribution of available foods, presumably to reduce 

interspecific competition and increase foraging efficiency (Nowak 2007). On the down side the 

protection from predators is lower in small groups, but the near absence of any predators on 

Zanzibar imposes a very little predation pressure (Nowak et al. 2008). 

There is very little knowledge about any instances of human poaching of Zanzibar red colobus. 

Possibly due to the human population being 98% Muslim, and therefore generally not being prone 

pursuers of bush meat (In personal conversation with locals, February 2016). A research project 

running from 2010–2014 regarding the conservation and management of Eastern African costal 

forest, listed hunting as the second largest threat to the Zanzibar wildlife after need for agricultural 

lands and wood fuel (Ministry of Natural Resources and Tourism 2014). The report implies that 

there is hunting on monkeys, but does not elaborate on the extent of this (Ministry of Natural 

Resources and Tourism 2014). It has been suggested that the monkeys have become subjects to 

hunting by immigrants from the mainland and other countries, and not by the native Zanzibarians. 

Colobus confiding to larger groups not embracing fission-fusion behavior would presumably not 

10



be protected against human hunting as larger populations are noisier and therefore easier for people 

to detect. P. kirkii has been a protected species since 1919 and listed in Appendix 1 in CITES, 

which means all hunting and trading of the species is illegal (Nowak et al. 2008; CITES 2016). 

This protection status possibly hesitating locals in sharing knowledge of any illegal handling of 

the species. 

Figure 5 Other colobine species are very threatened due to hunting by humans and other predators. 

Chimpanzee (Pan troglodytes scchweinfurthii) feeding on a Ugandan red colobus (Piliocolobus 

rufomitratus tephrosceles). In Kibale Forest National Park local populations of red colobus are going 

extinct, estimating a total population drop of 89% mainly due to over predation by chimpanzees (Lwanga 

et al. 2011). Photo: Alain Houle. 

P. kirkii groups living in proximity of shambas and feeding on unripe plantation coconuts (Cocos 

nucifera), thereby creating conflicts with local farmers, has caused chasing, trapping and poisoning 

of the monkeys to keep them out of crops. Research from 1999 by Siex and Struhsaker showed 

that the red colobus’ consumption of coconut actually promoted the net coconut harvest, and this 

should have put an end to this pursuing threat, and farmers demanding economic compensation 

for nonexistent colobus crop raids (Siex & Struhsaker 1999a; Rabe & Saunders 2014). 

11



METHODS 

In the period 2004 – 2005, Dr. Katarzyna Nowak conducted fieldwork in Kiwngwa – Pongwe 

forest, before FR status. To simplify comparison possibilities, data collection methods have largely 

followed Dr. Nowak’s methods. Prior to engaging fieldwork, the forest reserve was visited in 

August 2015 to judge the state of the transects and assess which transects could be reused. 

Study Site 

Kiwengwa – Pongwe Forest Reserve (Lat.: 06°00’43” S, Lon.: 039°22’01” E) is located in the 

Northeastern district of the main island Unguja in the Zanzibar archipelago. It is a natural forest 

situated only a few hundred meters from the coast, following the coastline from Pongwe to Cairo, 

covering a total of 33 km 2 . The vegetation type ranges from high coral rag with a canopy height 

of up to 30 m to shrubs and cultivated grounds (see Appendix 1). 

Figure 6 Location of the three transects used during transects walks. Transect K3 and K2 are placed 

parallel 2 km apart running from edge to edge of the forest reserve. Transect B is located close to the 

Mchekeni Caves visitor center, a place of higher core forest due to the water catchments in the caves. 

Transect walks were conducted in the direction Start - End. See Appendix 1 for additional map of study 

area. 

12



Monkey census 

Census of monkeys in Kiwengwa-Pongwe Forest Reserve was completed using transect walks. 

Prior to initiation of fieldwork I attended primate observation training in Udzungwa Mountains 

National Park, mainland Tanzania. Observation techniques were trained by census of P. kirkii’s 

nearest relative P. gordonorum, C. mitis (Sykes’ monkeys who frequent associate of Piliocolobus) 

and Colobus angolensis (black and white colobus, sister genus to red colobus). Training was led 

by Dr. Francesco Rovero. 

Transect walks in KPFR were conducted from January - April 2016, during the short winter dry 

season. Three transects B, K2 and K3 were traversed during census. 

- Transect B (S5° 59.974' E39° 21.596' - S5° 59.981' E39° 21.983') north transect, 0,7 km 

long, located close to the Mchekeni caves visitor center. 

- Transect K2 (S6° 00.513' E39° 22.901' - S6° 00.559' E39° 21.537') middle transect, runs 

parallel 2 km north of K3 and has a length of 2,5 km. 

- Transect K3 (S6° 01.559' E39° 23.510' - S6° 01.606' E39° 21.884'), is the most southern 

transect with a length of 3 km, located 4 km from the southern edge of the reserve. 

All transects run in an East – West direction at 169°. 

Walks were initiated at 06:30h. (SD 0.006). This start time was chosen in order to be methodically 

consistent with the data collection from 2004-2005 and guidelines for observation of diurnal 

primates. (National Research Council 1981; Whitesides et al. 1988). The transects were traversed 

at a pace of 1 – 1½ km h -1 starting at forest rim and moving inwards (Transects K3 and K2 in an 

East – West direction and transect B in a West – East direction). 

During transect walks all audial and visually detected encounters with humans, dogs, human 

disturbances, P. kirkii and sykes’ monkeys (Cercopithecus mitis ssp. albogularis) were notated. 

All human-related encounters, audio and visual detections, and monkeys detected by audially, 

were allowed 1 min stationary. Sightings of C. albogularis groups were allowed 5 minutes 

stationary and encounters of P.kirkii groups or mixed P.kirkii and C. albogularis groups were 

allowed 10 minutes stationary. Detection angels and sighting distances were detected using a field 

compass and a Berger & Schröter Range Finder, or in some cases visually estimated. Locations 

were defined as meters from transect start and all sightings of monkeys were also marked with a 

GPS waypoint on a Garmin ETREX 10 GPS. 

13



Most traverses of transect B were conducted by me alone, all other walks where accompanied by 

field assistant Mtumwa Simai. Mtumwa is an experienced monkey observer, with a great 

knowledge of the terrain and safety in KPFR. Therefore, we concluded it would benefit 

detectability being two observes despite the possible slightly enhanced noise factor. 

In cases of shorter rains, census was paused maximumly 30 minutes per walk. In case of rainfalls 

longer than 30 minutes total, the transect was abandoned until the following day. An individual 

transect was given a rest period of 72 hours between two repetition, to avoid monkeys being 

influenced by our presence, (Minimum recommended rest period of 36 hours (Whitesides et al. 

1988)). A total of 12 census walks of each transect were completed. 

Disturbance 

During census walks all human disturbances detected were denoted. Locations were noted as 

meters from transect start and all disturbances except audio detected disturbances (like with 

monkey encounters) were also marked with a GPS waypoint. Detected disturbances include: 

Fresh cut trees 

Wood bundles 

Human encounters 

Wood piles, firewood, poles etc. 

Encounters with dogs (with or without human accompaniers) 

Manmade forest clearings 

Wood cutting stations 

Trash and waste dumping 

Audio detected woodcutting ex. ax, saw or chainsaw 

Detection of humans talking or walking in forest 

See Figure 7 for examples of disturbances found under transect walks. 

Some areas had no encountered human disturbances, because audio detected disturbances ex. 

hearing use of axe or chainsaw, the precise location of the disturbance can be flexible. The location 

where the disturbance was heard most clearly was noted as the location of the disturbance. This 

also applies for audial detected monkeys. 

14



Figure 7 Pictures of human disturbance observed under transect walks. A: woodcutting station at transect 

K2. B: Cycad (Encephalartos hildebrandtii) cut for access to trees on transect B. C: Area of newly cut 

trees at transect B. D: poles of wood laying on transect K2. E: wood bundle found at K3. F: Waste 

dumped at transect K2. Photos by: Lærke Nykjær Johansen. 

Vegetation sampling 

Data for vegetation analysis was collected in 5×50 meter plots. Vegetation plots where placed at 

the start of each transect (0 m) and each 200 meters for transect K2 and K3, and each 100 meters 

for transect B, following locations from 2004. For highest similarity between vegetation data 

collected in 2004 and 2016, speciation was conducted by local botanist Tahir Abbas Haji, who 

also assisted in 2004. 

Within each plot species and Diameter at Breast Height (DBH = 130 cm) was registered of all 

trees, shrubs, bushes and lianas with a height ≥ 2,5 meters. Vegetative disturbances where 

registered, by measuring DBH and species of all trees cut by human activity. 

15



Data was collected from a total of 35 plots. 7 plots on transect B, 13 plots on transect K2 and 15 

plots on transect K3. The plot area 400 meters down transect K3 was not examined in 2004 because 

of a previous wildfire. It has because of this been excluded from further data analysis. Resulting 

in a total of 34 plots used in data analysis. 

DATA ANALYSIS 

Data were analyzed using Microsoft Excel with XLSTAT and Analysis ToolPak add-ons. Graphs 

and statistic tester were calculated using Graphpad Prism version 6. DISTANCE version 6.2 was 

used to calculate P. kirkii population estimations. 

Distance sampling (DS) 

The program DISTANCE was used to estimates population densities, group sizes and number of 

animals within sampling area, based on perpendicular distances. DISTANCE makes these 

estimates based on a detection function g(x), modeled to best fit the distribution of perpendicular 

distance data entered. By incorporating an observer’s decrease in ability to detect a given animal 

over distance, DISTANCE estimates a density within the sampled area (Thomas et al. 2010). A 

half-normal key function and a half-normal key function with a cosine adjustment were selected 

based on lowest Akaike Information Criterion (AIC), to best describe observation distance 

distributions (Chosen model 2004 AIC 394,93 < 355,81 and for 2016 AIC: 372,72 < 395,32; 

396,36). Distances entered were perpendicular distances to estimated center of group, if no 

estimation to group center was possible, perpendicular distance to first observed animal was used. 

Group sizes were number of monkeys observed including lowest estimated other individuals in 

group, based on movement etc. 

The perpendicular distance (PD) is the shortest distance from transect to observed animal, 

calculated by basic trigonometry. As few animals are observed in an angle to the transects, 

elongating the measured distance from observer to animal, the following mathematical formula is 

used to calculate perpendicular distance: 

16



Figure 8 Techniques for observing primates during transect sampling explaining relationship between 

sighting angle and distance of observed animal in group, and shortest distance to transect sampled. P = 

perpendicular distance from transect to first observed animal, ȓ = ½ mean group spread, ɵ = sighting angle 

and S = sighting distance. 

P = S× sin(θ) 

S = Sighting distance 

Ɵ = Sighting angle 

Whiteside method (WM) 

For alternative estimations of population densities, sighting distances were also calculated using 

Whiteside method. The Whiteside method estimates an adjusted perpendicular distance P’, to 

calculate perpendicular distance including average group spread as a variable (Whitesides et al. 

1988). 

P ′ = P(1 + r̅ 

S ) 

P = perpendicular distance to first observed animal 

ȓ = ½ mean group spread 

S = sighting distance to first observed animal 

17



Using the Whiteside method, the detection function to best fit the distribution of data was, for 2004 

a half normal key function (ACI 371,53



RESULTS 

Population analysis 

AIC was lower for 2004 results generating a better fit detection function g(x), for the observation 

data using both perpendicular distance calculation methods (ACI; 2004 DS: 353,84 WM: 371,53, 

2016 DS: 394,92, WM: 423,96). Both methods estimated densities (groups/km 2 ) larger in 2016 

than in 2004, but also with higher standard errors (Density; 2004 DS: 6,49±1,48, WM: 4,60±1,17, 

2016 DS: 12,95±5,36, WM: 6,42±3,42). The estimated average group sizes vary very little 

between years of the same method, but groups were calculated to be approximately one individual 

larger using Whiteside method (DS: 2004: 6,56±0,49, 2016: 7,66±1,15, WM: 2004: 7,66±1,15, 

2016: 7,10±0,87) (table 1 and 2). 

Table 1 Population estimates from Distance sampling with standard error 

AIC 

DENSITY 

GROUPS/KM 2 

INDIVIDUAS 

/KM 2 

2004 353,84 6,49 ± 1,48 40,03 ± 

10,16 

2016 394,92 12,95 ± 5,36 87,66 ± 

37,88 

GROUP 

SIZE 

6,56 ± 

0,49 

6,77 ± 

0,84 

N COLOBUS IN 

AREA 

240 ± 60,92 

526 ± 227,31 

Table 2 Population estimates from Whiteside method with standard error 

AIC 

DENSITY 

GROUPS/KM 2 

INDIVIDUAL 

S / KM 2 

2004 371,53 4,60 ± 1,17 35,30 ± 

10,41 

2016 423,96 6,42 ± 3,42 45,62 ± 

24,92 

GROUP 

SIZE 

7,66 ± 

1,15 

7,10 ± 

0,87 

N COLOBUS IN 

AREA 

212 ± 65,52 

274 ± 149,63 

19



Encounter rate 

Encounter rates for 2004 and 2016 were calculated as average numbers of colobus groups 

encountered per kilometer traversed. The average encounter rates for 2016 (2016 B: 1,79 ± 0,65, 

K2: 0,80 ± 0,74, K3: 0,42 ± 0,25) are in general higher on all transects than in 2004 (2004 B: 0,83 

± 1,29, K2: 0,46 ± 0,15, K3: 0,30 ± 0,30), 

ER REPETITIONS MAX. MEAN ± P-VALUE SIGNIFICANT 

VARIABLE 

SD 

2004 49 4,29 0,49 ± 0,68 0,001 ** 

2016 36 2,86 1,00 ± 0,82 

B 2004 12 4,29 0,83 ± 1,29 0,01 * 

B 2016 12 2,86 1,79 ± 0,65 

K2 2004 19 0,80 0,46 ± 0,15 0,31 no 

K2 2016 12 2,40 0,80 ± 0,74 

K3 2004 18 1,00 0,30 ± 0,30 0,12 no 

K3 2016 12 0,67 0,42 ± 0,25 

Figure 9 Encounter rates for all three transect in 2004 and 2016 

several of the transects also showing a larger variance. A Mann-Whitney U test showed a 

significant difference between the overall encounter rate of 2004 and 2016 (Mann-Whitney 

U=527, p



Table 3 Comparison of encounter rates within each sampling year. 

KRUSKAL-WALLIS TEST ON ENCOUNTER RATES 2004 SIGNIFICANT SUMMARY 

P - VALUE 0,0027 Yes ** 

DUNN'S MULTIPLE COMPARISONS TEST Mean rank diff. Significant Summary 

K2 2004 VS. K3 2004 15,56 Yes ** 

K2 2004 VS. B 2004 9,333 No ns 

K3 2004 VS. B 2004 -6,222 No ns 

KRUSKAL-WALLIS TEST ON ENCOUNTER RATES 2016 Significant Summary 

P - VALUE < 0,0001 Yes **** 

DUNN'S MULTIPLE COMPARISONS TEST Mean rank diff. Significant Summary 

K2 2016 VS. K3 2016 5,5 No ns 

K2 2016 VS. B 2016 -13 Yes ** 

K3 2016 VS. B 2016 -18,5 Yes **** 

Group size means 

The larges mean size of groups observed classified by transect, was in 2004 found at transect K2 

(M=7,3, SD±3,4) and in 2016 on transect B (M=6,1, SD±4,7). Average cluster sizes were larger 

at both transect K2 and K3 in 2004 (2004 M±SD: K2=7,3±3,4, K3=6,2±2,6) than in 2016 (2016 

M±SD K2=5,5±5,2, K3=4,7±2,8). The smallest average cluster size was in 2004 at transect B 

(M=5,0, SD±3,7) and in 2016 at K3 (M=4,7, SD±2,8). A Mann-Whitney U test confirmed a 

significant difference in mean cluster size, being lower in 2016 (M=5,46, SD±4,48) than in 2004 

(M=6,43, SD±3,32) (Mann-Whitney U =949, p = 0,04). A Mann-Whitney U test also showed that 

the average cluster size has dropped significantly on transect K2 (Mann-Whitney U =163,5, 

p=0,03) but had not changed significantly at transect B (Mann-Whitney U=47, p=0,73) or K3 

(Mann-Whitney U=79, p=0,11). Within each sampling year, there was not found a significant 

difference in cluster size between the transects (Kruskal-Wallis test; DF (2004) = 45, DF (2016) = 

53, P (2004) = 0,29, P (2016) = 0,8). 

A total of 45 P. kirkii groups were observed in 2004 versus 54 in 2016. Singletons and smaller 

groups were observed more frequently in 2016 than during census in 2004 (Appendix 4). In 2004 

6 observations (13%) were singletons or doubletons. In 2016 15 observations (28%) were 

singletons or doubletons. Singleton observations were most frequent at transect K2 both years. 

Transect K2 was overall the transect with most observations, holding 22 observations in 2004 

(48% of all observations in 2004) and 24 observations in 2016 (44% of all observations in 2016). 

The largest group observed in 2004 (N = 15) was observed at transect K2, likewise the largest 

single group of colobus observed in 2016 (N = 24) was also observed at transect K2. 

21



Figure 10 Cluster size means + 1 SD of each transect and overall mean for 2004 and 2016 and Mann - 

Whitney U test. 

CLUSTER 

SIZE GROUPS MAX. MEAN ± SD 

P- 

VALUE SIGNIFICANT 

2004 46 15 6,4 ± 3,3 0,042 * 

2016 54 24 5,5 ± 4,5 

B 2004 7 10 5,0 ± 3,7 0,73 no 

B 2016 15 15 6,1 ± 4,7 

K2 2004 22 15 7,3 ± 3,4 0,026 * 

K2 2016 24 24 5,5 ± 5,2 

K3 2004 16 10 6,2 ± 2,6 0,11 no 

K3 2016 15 12 4,7 ± 2,8 

VEGETATION ANALYSIS 

A total of 9293 and 5428 live stems were measured within the 34 analyzed vegetation plots in 

2004 and 2016. This is a reduction of 41,6% in forest density over the 12 years between samplings. 

The average number of stems in each plot has dropped significantly between the two sampling 

years (Pared T-test=6,83, DF=33, P



The diameter sum of cut trees follow the same tendencies as number of cut stems in each plot 

being significantly larger at B 2016 (pared T-test=6,42, DF=6, P=0,0007) then anywhere ells 

(Figure 11D). 

Because the stem densities vary so significantly between the two years, comparing proportions of 

cut trees of plot total stem density may be more descriptive to illustrate vegetative disturbance 

levels. Here we find no significant difference between the two years in proportion of cut stems in 

plot, though 2016 is a fraction higher (Wilcoxon=177, P=0,13) (Figure 11E). The number of cut 

stems removed from each plot compared to how many live stems are found within the same plot 

is on average significantly higher on transect B in 2016 than in 2004 (Wilcoxon=28, P=0,016). At 

transect K2 the proportion is higher in 2016 and at K3 it is lower, but her neither are significant 

(Wilcoxon: K2: W=49, P=0,09, K3=-17, P=0,63). Comparing diameter proportions, also here 

there is a significant difference in proportions at transect B (Wilcoxon=28, P=0,016). The mean is 

larger in 2016 at K2 (M=0,19 SD=0,13) and K3 (M=0,16 SD=0,15), but the difference is not 

significant (Wilcoxon: K2 W=31, P=0,3, K3 W=15, P=0,67). Overall the proportion is larger in 

2016 than in 2004. The variance at the different transects is here also fairly large (Figure 11F). 

Significant results of multiple comparison of the above-mentioned parameters within sampling 

years are summarized in Appendix 5, including results and further analysis with Dunn’s multiple 

comparison test, if Kruskal-Wallis test showed a significant result. 

In 2016 there is only found a within year difference in total diameter of cut trees. The total diameter 

of cut trees at transect B is significantly bigger than the total diameter of cut trees at transect K3. 

In 2004 transect B is significantly different from one or both transects at several parameters. 

Transect B has significantly less stems per plot than transect K2 and K3. K3 has significantly 

more cut stems per plot than transect B, and the total diameter of cut trees in plot is significantly 

lower at transect B than at transect K2. The average DBH proportion of cut trees of plot total DBH 

was significantly lower at transect B in 2004. Average proportion did not vary between transect 

K2 and K3. 

Species composition and abundance 

A total of 141 species of trees, shrubs and lianas were found in the two sampling years. 119 species 

in 2004 and 75 species in 2016, totaling 61 species shared between the two sampling years. 58 

species found in 2004 were not found in 2016, and 14 new species were found in 2016 (a total 

23



species list can be found in Appendix 7). Of the 58 species not found in 2016, 6 have only been 

identified to genus level. Five species are only known by local Swahili name and do not yet have 

a scientific name. 14 stems or stumps, out of a total of 6226 live and dead stems measured where 

unknown or unidentifiable in 2016, and 63 out of 10440 in 2004. In both cases an identification 

rate of >99%. In average 5,9 fewer species were found in each vegetation plot. Some species more 

common in 2004 have been lost from several plots. One of the more common species Euclea 

schimperi has disappeared from 22 of 34 plots. 30 of the 58 species not found in 2016 were locally 

rare only found in a single vegetation plot. The plot that had undergone the highest species decline 

is the rim plot (B0000) from 0-50 m at transect B (N= 11) followed rim of transect K2 (K22400) 

(N=9). 

24

p ro p o rtio n 

p ro p o rtio n 

S te m s 

C m 

S te m s 

C m 



A 

S te m s in p lo t 

B 

P lo t m e a n D B H 

5 0 0 

4 0 0 

**** 

** 

*** 

1 5 0 0 ** **** 

** *** 

3 0 0 

* 

1 0 0 0 

2 0 0 

5 0 0 

1 0 0 

0 

0 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 0 4 

K 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 0 4 

K 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 

T ra n s e c t 


C 

M e a n n o c u t s te m s / p lo t 

D 

c u t D B H 

8 0 

* 

4 0 0 

** 

6 0 

* 

3 0 0 

4 0 

2 0 0 

2 0 

1 0 0 

0 

0 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 1 6 

K 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 0 4 

K 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 



E 

0 .4 

P r o p o r tio n c u t s te m s o f p lo t to ta l 

F 

0 .4 

* 

p r o p o r tio n c u t D B H 

* 

0 .3 

* 

0 .3 

0 .2 

0 .2 

0 .1 

0 .1 

0 .0 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 0 4 

K 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 

0 .0 

B 2 0 0 4 

B 2 0 1 6 

K 2 2 0 0 4 

k 2 2 0 1 6 

K 3 2 0 0 4 

K 3 2 0 1 6 

2 0 0 4 

2 0 1 6 



Figure 11 Results of vegetation analysis. A: analysis of average number of stems in plot B: average total 

DBH for each plot C: Mean number of cut stems per plot D: Average DBH of total cut stems in plot E: 

average no of cut stems in plot of plot total number of stems F: average DBH of cut stems compared with 

plot total DBH. 

25



CORRELATIONS 

At total of 19 parameters of interest covering, P. kirkii abundance (P. kirkii abbreviated as 

colobus), vegetation quality, disturbance indicators, food species availability, geographical 

parameters and DBH measurements where analyzed for correlating effects in the spearman’s 

correlation map in Figure 12. Correlation range from positive correlation to negative correlation in 

a red – blue scale. Of the 361 comparison of parameters, 47 significant positive and 14 significant 

negative correlations were found. Correlations summary of correlation parameters and significant 

correlations can be found in appendix 8. 

Correlation map 

N food species 2016 

DBH of live stems 2016 

N Colobus 2004 

N trees DBH > 10 cm 2016 

Food species Prop. 2016 

Meters from rim 





N stems 2016 


N stems 2004 


N cut sems 2016 

Fewer stems 

Disturbance 

DBH of cut stems 2016 

Species lost 

Figure 12 Correlation map on 19 variables. Correlation range +1 to -1 in a red to blue color scale. White 

is correlations between -0,1 - +0,1, yellow and green are respectively +0,1 - +0,2 and -0,1 - 0,2. 

26



Correlation map (P-values) 











N stems 2016 


N stems 2004 


N cut sems 2016 

Fewer stems 

Disturbance 

DBH of cut stems 2016 

Species lost 

Figure 13 Significance levels for correlation mad. Black squares represent a significant relationship 

between the two variables. 

27

Frequency 

100 

200 

300 

400 

500 

600 

700 

800 

900 

1000 

1100 

1200 

1300 

1400 

1500 

1600 

1700 

1800 

1900 

2000 

2100 

2200 

2300 

2400 

2500 

2600 

2700 

2800 

2900 

3000 



HUMAN DISTURBANCE 

Human disturbances recorded along transect while traversed during census are showed in Figure 

14. There has been an increase in level of disturbance along forest rim at approximately 0 m and 

3000 m. 

10 

Human Disturbance 

2004 2016 

5 

0 

Meters from transect start 

Figure 14 Frequency of human disturbances recorded along transect walks. 

Linear Regression 

The location of observed human disturbances and P. kirkii monkeys can best be described as 

distance 0-1500 m from forest rim, 1500 m being furthest from any forest border/rim. The linear 

regression fit to best describe the tendencies observed in locations of humans and P.kirkii 

monkeys, shows that humans were more frequent closer to forest border and colobus being more 

frequent the further you get from forest rim. The linear regression has a better fit description of the 

tendencies in location of human disturbances, describing 32% of the data whereof only 10% of the 

large variation in colobus data can be describe by a linear relationship to location. 

28

Frequency 



Linear Regression 

10 

8 

Disturbance 

R² = 0,3249 

6 

Colobus 

R² = 0,0951 

4 

2 

0 

0 250 500 750 1000 1250 1500 


Figure 15 Linear regression on located observed colobus and human disturbances at K2 and K3. 

The relationship between human disturbance and vegetative changes were further analyzed with 

linear regression tests. All four linear regressions are fairly scattered, but some assumptions can 

be made on the found. There was not found a relationship between the number of species that had 

been lost from a plot and the fluctuations in the number of stems in plot. Comparing the number 

of stems cut in a plot with number species lost from plot, there is a relatively clear tendency 

towards more species not being found where plots had been subjected to a higher level of 

woodcutting. There was a linear negative relation between how deep in the forest reserve the 

vegetation plot was (e.g. meters from rim or forest edge) and how many species where not refound 

in the plot. More species have been lost at the forest edges than at the core of the forest. 

Also, when comparing disturbances encountered during transect walks and species lost in same 

area there was found a positive tendency describing the relationship. Encountered human 

disturbances are only registrations from 2016 and include all transects. Some areas had no 

encountered human disturbances, but because audio detected disturbances ex. hearing use of axe 

or chainsaw, humans talking and dog barking, the precise location of the disturbance can be 

flexible. The location where the disturbance was heard most clearly was noted as the location of 

the disturbance. This contributes some of the variation in the regression relationship which 

explains approximately 9 % of the variation in the data. The regression between cut stems, meters 

from rim and species loss describe respectively 12 % and 24 % of the variation. 

29



1 

1 

Figure 16 Linear relationships between the different parameters that are used as a proxy for human 

disturbance. 

30



DISCUSSION 

In this thesis, I studied P. kirkii along three transects. I estimated population density and structures, 

and will in this discussion summarize and outline how these findings correlate to habitat and 

human disturbances, in the two sampling years. Furthermore, I will discuss what alternatives 

approaches could improve this work and also relate my findings to other relevant studies. 

Both methods for estimation of P. kirkii density, showed a higher density in 2016 than in 2004. 

This is consistent with the significantly higher encounter rate also in 2016. The overall significant 

difference in encounter rate is largely due to the large difference in ER at transect B. 

The within year comparison of transects ER, showed that the colobuses (P. kirkii implied) in 2004 

were encountered more often in the area around transect B over transect K3, and transect K2 over 

transect B, but there was only found a significant difference between K2 and K3. In 2016 colobuses 

were much more frequently observed at transect B than anywhere else in the studied area. ER was 

still higher at K2 than at K3 but statistically not significant. Finding collectively emphasizes a shift 

in where colobus roam in higher numbers and prioritizing of area from north to south, transect B 

to transect K3, becoming more pronounced in 2016. 

Together with population estimates, this implies that there has been an overall increase in colobus 

abundance, and the increase is most pronounced at transect B in the Mchekeni area. 

DISTANCE SAMPLING 

The two calculation methods used to estimate population density, generated quite difference 

population estimations, though with uncertainties both methods estimated a higher colobus density 

in 2016 than in 2004. This is consistent with the significantly higher encounter rate also in 2016, 

but the encountered groups were also significantly smaller, on average by one individual. 

It is generally accepted by primatologist that a minimum 60-80 observations are required for 

proper populations estimations (Marshall et al. 2005). The total observations for both years fall far 

shorter then general requirement, making population estimates based on an inadequately small 

dataset. 

31



Secondly there is in both sampling years a very low detection within the first 5 m from the transect. 

This could imply violation of the two first assumptions in line transect surveys 1) that all animals 

on transect (0m distance) are detected, 2) that animals do not move before detection (National 

Research Council 1981). I cannot argue against these possible violations, but as this lack in 

observation has occurred equally in both sampling years, it should not be consequential for the 

comparison of the data. This neglect in observations in close proximity to the transect could be 

because of misted or bad observation skills, but it could also plausibly be because the colobus in 

general stay further away from the transects. This because the transects in 2016, obviously were 

used relatively often by locals as a common path for entering and exiting the forest (personal 

observation). Only short parts of the transects had to be reopened at the start of the field work. 

This mostly being parts of lower vegetation and shrubby areas, probably of less interest for the 

locals to gain access to. To decrease accessibility for local use, the path width had been kept to a 

minimum. This also mean that it can be difficult to survey without generating any sound when, 

due to dens understory brushing against legs. As transect walks were conducted during dry period, 

where forest floor leaf litter was extremely dry generating noise despite an effort tiptoeing on more 

solid rocky underlay. Colobus could possibly have detected our sound and fled to a further distance 

why using animal-to-observer distance (AOD) in this case possibly could have been a better 

approach for calculation of detection functions. There is an ongoing scientific discussion on which 

method should be used to estimate population densities (Hassel-Finnegan et al. 2008). Both 

methods used in this study were based on perpendicular distance, but animal-to-observer distance 

have been used when surveying populations in JCNP (Siex & Struhsaker 1999b). Due to the low 

detection, nearest to the transect, of that ever reason this is caused, it could in the future be a 

possible better approach using AOD also when sampling in KP. 

The densities calculated for 2016, showed a large variance, why the results should not be 

interpretation as definite populations sizes, but implied estimations. As the results are higher in 

2016 using both calculations, I have chosen to interpret the results suggesting that the sampled 

area in KPFR holds an undefined, but slightly density of colobus now then in 2004. The 

inaccuracies are also likely due to the reduced data set. 

32



GROUP SIZE 

Singleton or doubleton observations had become more common in 2016. Observed groups where 

overall significantly smaller in 2016 than earlier observed. Even with the observation of the 

exceptionally large group at transect K2 (N=24) during 2016 survey, the average group size at 

transect K2 was significantly smaller than in 2004. 

Earlier studies have shown large fluctuations in cluster size between forest core and edge, and have 

therefore been estimated separately (Nowak 2007). The smallest ever reported population size for 

P. kirkii of average 5,5 individuals, was from observation of edge groups in KPFR (Nowak 2007). 

The fact that the 2016 overall mean population size (core and edge forest of all transects together), 

is equal to the lowest earlier reported, is somewhat worrying for the conservation status of P. kirkii. 

To why this is worrying I will elaborate on later in the discussion. 

Methodically the data from 2004 was collected over a 12-month span covering all seasonal 

changes, where 2016 data was collected over a 3-month span within one season, weakens the 

reliability of comparisons. And indeed, the collection over a longer time span would give a more 

precise population estimate despite seasonal changes. The 2016 data were as mentioned collected 

during the winter dry season, where infant recordings in Kiwengwa peak (Nowak & Lee 2011). 

This means 2016 population estimates and cluster size averages in fact may be overestimated due 

to a possibly higher infant rate. Earlier studies also imply a lower infant survival rate in KPFR and 

other more disturbed habitats, compared to population living in stable habitats in JCNP, or with 

access to refugee in mangrove forests, also emphasizing that the very low average group size still 

possible could be an overestimation (Nowak & Lee 2011; Siex & Struhsaker 1999b). 

Group size is largely determined as a compromise between foraging investment, resource 

availability and protection from predators in larger numbers (Struhsaker 2000). As we know there 

are hardly any larger predators on Zanzibar, and hunting upon colobus is restricted, this therefore 

not imposing a necessary limitation to minimal population sizes. Investigation of P. gordonorum 

group size showed a significant decrease in population size in human disturbed areas (Marshall et 

al. 2005). Human encounters in KP during census has increased at forest rim, but using N cut stems 

per plot as a proxy for disturbance, there was not found a significant difference between the two 

sampling years, though there has been a shift from transect K3 to B in location of greatest number 

33



of stems harvested. As colobus are highly social animals, where population sizes have earlier been 

reported this low, it has been addressed as the possible minimum bearable limit of red colobuses 

essential ecology and requirement for social interaction (Marshall et al. 2005). 

Variation in group size has also increased at all locations which could indicate a home range 

overlap between populations of different sizes. As this could compromise safety of smaller groups, 

I suspect the local variation in group size may be caused by smaller groups representing foraging 

parties that have split from larger social groups. 

Fission-fusion behavior has earlier been observed in KP, where groups in core forest showed to 

split into ≥2 foraging subgroups in 72% of observations (Nowak & Lee 2011). Adaptation to 

fission-fusion behavior has been documented in several Piliocolobus species, to increase foraging 

yield and decrease intergroup competition, in habitats with clumped food resources, low species 

diversity and large home ranges needed to cover dietary requirements (Marshall et al. 2005; Nowak 

2007; Struhsaker 2000; Struhsaker et al. 2004) P. gordonorum display fission-fusion behavior in 

heavily human disturbed areas, as such a displayed flexible group structure may be a necessary 

adaptation to living in an inadequate human dominated habitat (Nowak & Lee 2011; Marshall et 

al. 2005). 

Where colobuses were observed more frequently in KP, cluster size also tend to be larger, 

indicating a habitat able to sustain more colobus, as mentioned group size is simultaneously largely 

determined by habitat quality (e.g. food available ect.) (Siex and Struhsaker 1999b; Struhsaker 

1975). Arguing that the decrease in population size observed in KPFR could be an indicator of a 

reduced habitat quality and increase in disturbance. 

Also, this could imply that P. kirkii in KPFR indeed do prefer higher coral rag forest as main 

habitat, but can embrace other habitat types, if conditions are right and if there is a sustainable 

diversity in available food sources. This is supported by an investigation by Siex (2011) where 

five different habitat types between KPFR and JCNP, where examined only finding one sign of P. 

kirkii presence outside high coral rag forest, opposed to 27 within (Siex 2011). These tendencies 

were furthermore consistent throughout investigated areas on the whole island. 

34



VEGETATIVE OUTCOMES 

Vegetation analysis showed that the forest density and likewise mean DBH has dropped 

significantly most at transect K2. In 2004 there was a significant higher amount of woodcutting at 

transect K3 than anywhere ells. This has decreased significantly and there has instead been a 

significant increase in wood harvest at transect B. In most places of the forest, the stems harvested 

by woodcutters are thin, rarely with a diameter above >5 cm, purposely to make small household 

fires for cooking. During walks at transect B we observed harvest of very large trees, much more 

frequent than at other transects, which is also indicated by the bias between number of cut stems 

and the very high DBH of cut stems at transect B. This is why several factors for vegetative human 

disturbance have been included in the vegetation analysis, and to investigate if the was a bias in 

tree harvested in comparison to mass and density available the two proportional differences 

(proportion of cut stems of total stems in plot and proportion cut DBH of plot total DBH). The 

results stowed bias at transect B possibly because of the easier transportation of larger trunks by 

accessing to the forest from the road to the Mchekeni caves visitor center. 

Of the species found and not re-found not much is to be said about their conservation state as very 

few species have been evaluated by any conservation agency. Of the eight species evaluated by 

IUCN or CITES, Encephalartos hildebrandtii is the only one under concern, listed as near 

threatened (NT) with declining population size (Bösenberg 2010). Special notice has been made 

to sub-population rapidly being destroyed on Zanzibar, due to the growing demand for agricultural 

grounds, tourism and local urban development (Bösenberg 2010). During census, we experienced 

several cases of E. hildebrandtii that had been cut to gain access to trees behind the cycad. 

Of other evaluated species, Erythrococca berberidea is listed as of least concern, due to great 

protection in South Africa. Subpopulations of E. berberidea in Tanzania are considered threatened 

due to continuous degradation of habitat in protected areas, but this concern seems to be focused 

on two forest reserves in proximity of Dar es Salaam and does not list any detail on Zanzibar 

distributions (IUCN SSC East African Plants Red List Authority 2013a). 

Of specimens only identified to genus level Turraea has 29 species whereof five have been 

evaluated by IUCN. Four species are listed as vulnerable, endangered or critically endangered. 

Due to distribution, endemism and ecology, the species found in KPFR is undeniably, not one of 

35



the threatened species. The species could likely be Turraea mombassana a very common shrub in 

costal forest shrub land forest, or Turraea floribunda a species which was found in 2016 (IUCN 

SSC East African Plants Red List Authority 2013b). 

In 2016, 19 individuals of Dovyalis macrocalyx were registered. I suspect that the unknown 

Dovyalis species found in 2004 is also D. macrocalyx a relatively common species in fringing 

forest but with no previous official records in Zanzibar (Hyde et al. 2016). 

Psychotria bibracteatum and Psychotria goetzei were both found in 2004 and 2016. A third 

unknown Psychotria was also found in 2004. Psychotria has a very long list of IUCN evaluated 

species. Several critically endangered. It is not possible to determine which species the sample 

from 2004 is, or if it is endangered or not. Other evaluated species that have not been identified to 

species level do not have any threatened or endangered species with a likely range on Zanzibar. 

The main concern should be focused towards conservation of E. hildebrandtii as it also is an 

important P. kirkii food species and is occasionally excavated for ornamental purposes in hotel 

gardens. More knowledge on species distribution on Zanzibar and a thorough investigation or 

publication of collected data is desirable to further investigate if the consequences of the ongoing 

wood cutting on Zanzibar for floral species composition. 

Human disturbance showed to have a negative interference on species composition. Plots were 

placed as precise as possible in the same locations to analyze species composition in a capture - 

recapture method. As several new species were found in 2016 which were not present in 2004, 

there is a high chance some of these new species have the same functional traits and a species 

turnover has occurred. Even taking this into consideration the species richness has still dropped 

considerably and probably most remarkable is, that richness has dropped in accordance to 

increasing human activity. 

36



CONSEQUENTIAL RESULTS 

Food species availability showed a positive correlation P. kirkii location, though not significant. 

Correlations with distance to rim and abundancy of larger trees outline that food is patchy 

distributed in core forest with bigger trees and in more shrubby areas, were smaller groups were 

encountered, believed to be foraging parties. This also explains why correlations and linear 

regressions with location of colobus are weak or not significant. Encounter rate and group size was 

higher in higher coral forest, but shrubs were occasionally visited, splitting correlations. 

Shrubby areas are often closer to higher levels of disturbance and I believe this reflects on the 

relation to colobuses habitat use. Seasonal changes in available foods is larger in coral rag forest 

than in cultivated shambas. This has elsewhere led to extremely inflated population densities of 

550 individuals/km 2 (Siex & Struhsaker 1999b). This has earlier been misinterpreted as a habitat 

preference where I support the original findings concluding that this is an exceptional case only 

possible because of dietary diversity requirement satisfied in the adjutants NP, as Piliocolobus 

have high dietary diversity requirement (Siex & Struhsaker 1999b; Siex & Struhsaker 1999a; 

Onderdonk & Chapman 2000). Colobuses were occasionally observed in smaller groups closer to 

forest edge right after sunrise. During return from transect walks at midday, groups had often 

retreated to the core forest for midday rest in the shade of the greater canopy cover. As these 

observations where outside of census they have not been included in the analysis but does 

nonetheless point out some population behavioral tendencies. 

During studies in 2004 two adjacent transects, K1 and K4 were also traversed. These transects 

were not used during this recent study as they were not revivable. They would have required a lot 

of work clearing the transects possibly having greater consequences for the forest and was also 

prohibitive due to time limitations. The area around transect K1, located north of the main road 

to Kinyasini, has been completely cleared of higher coral forest and is now a patchy shrub forest. 

Interviews with locals and investigation of the area showed no signs of colobus monkeys living in 

this area. The increase of both colobus and wood cutting at transect B in Mchekeni, located south 

of the main road may have happened as a reaction to the degradation of the northern transect K1 

area. 

37



With disturbance in general being higher towards forest rim it could indicate the forest reserve 

being degraded from forest borders and inwards, causing a higher deforestation from north and 

south edge of the forest reserve, where transect K1 and K4 were located. Areas at the north and 

south boarder have previous been proposed as “high protection zones” because of being crucial 

for continued survival of Zanzibar wildlife, and experiencing extensive threat, demanding 

excessive protection (Siex 2011). 

Population estimates predicted an advancement in total abundance of colobus within the sampled 

area, which is not consistent with the tendencies in group size, vegetation and disturbance, why I 

can only argue that the elevated population size, could be a result of population compression from 

north, south and rim inwards. Population compression is a phenomenon within Piliocolobus 

history and has earlier been predicted in P. kirkii in other areas of Zanzibar (Nowak 2007). 

Population compression occurs as a result of habitat loss or degradation, causing inflated 

population densities by immigration to more adequate habitats (Struhsaker 2010). Without 

vegetation analysis population densities, can be a very misleading indicator of habitat quality (Siex 

& Struhsaker 1999b). 

I believe the population compression is due to immigrations from these northern and southern 

areas of the reserve, to Mchekeni and core forest areas. 

Because of the simultaneous high human and colobus activity at Mchekeni, the general picture 

from transect K2 and K3 of colobus preferring less disturbed central/core areas of the reserve and 

humans dispersing with opposite tendencies, cannot be transferred to Mchekeni. Additionally, as 

transect B being so short and the ecological structure of the forest in this area being quite different 

from the overall compositions the tendencies in this area understandably very from the rest of the 

transects. The vegetative indicators of human disturbance observed at transect B, specifying a 

rising human-colobus conflict, between where monkeys are more abundant and the increased 

human activity. 

38



MANAGEMENT STATUS 

As mentioned, The 2013 list of Priority Primate Areas, listed Kiwengwa – Pongwe Forest Reserve 

as of special interest for the protection of the endangered Zanzibar red colobus (Davenport et al. 

2013). In general, there is an array of different boards, groups, organization and councils to manage 

and oversee the nature and conservation of Zanzibar, embracing the importance of integrating the 

local community and their needs in conservation management. This new generation approach of 

conservation management is well organized, but it appears that the lack of sufficient funds leaves 

most agreements and projects, started by the administrational parties, not able to be integrated at 

community level. During my time in Kiwengwa I did not once experience any signs of community 

and conservation cooperation, nor did the established collaborations in JCNP, show any functional 

commotion when visiting the park visitor center. This implying a lack in implementation of 

conservation strategies. The majority of communities surrounding the PA’s are highly dependent 

on forest related labor, having nonexistence secondary income opportunities. This gives the 

community based organizations (and the government) managing the areas, little ability of inducing 

sustainable change in the natural resource harvest from the forests, causing the continued 

degradation (Hassan & Said 2011). Tanzania generally experiences a lack in active and adequate 

management of government protected forests. High density human settlements adjacent to 

government protected area, consequently linger, a lower conservation success than in unmanaged 

forest in low density of human settlements areas (Davenport et al. 2013). 

39



CONCLUDING REMARKS 

In this study of the endangered P. kirkii, in Kiwengwa-Pongwe Forest Reserve. Habitat protection 

has not (or at least not yet), had the anticipated conservation effects. We did find a possibly higher 

total density of P. kirkii and a higher encounter rate, but these results were based on a dataset of 

fewer observations than what is generally required for population estimates. We also only sampled 

a more central section of the forest reserve, were wildlife is more secluded from disturbing human 

activity. Comparing the in general very low average group sizes, the significant reduction in group 

size between the two sampled years, the shift in location, the decreased forest density, and the 

avoidance of human activity, I can only conclude that the colobuses are still suffering habitat loss 

forcing to compress furthest from humans, in safety of the patches of higher coral rag that still 

remain. 

For conservation management to have an impact here, I believe that the root to the problem has 

to be addressed. In Kiwengwa – Pongwe the problem imposing the largest threat to the reserve 

wildlife, is the surrounding human settlements being so dependent on forest products, particularly 

fuelwood. This is a socioeconomic concern cause by lacking affordable alternatives, to illegal but 

cheap, wood collection in the reserve. It has little effect creating rules and regulations through 

management, if the people you are affecting by this, have so limited resources that they have no 

ability to follow conservation attempts, and a reassessment of the conservation management 

implementation is critical needed. 

Before the Zanzibar collective society can support a more prosperous community, commonly 

affording alternatives to fuelwood, I do not believe that despite community including management 

strategies, we will see a positive effect on the conservation of the Zanzibar endemic Piliocolobus 

kirkii. 

40



FURTHER RESEARCH 

For a more thorough re-assessment of the conservative state of KPFR, reopening of transects K1 and 

K4 should be considered. The doubts of what consequences this might have, creating easy access 

routes through the forest also beneficial for wood cutters, might indeed not be that great if the areas 

of the forest already have undergone a great degradation. We did consider doing it for this study, but 

the time restrictions did not allow the time for it, nor would I have been able to conduct enough 

repetitions of each transect within the time limit. Conducting this amount of field research for one 

researcher within three months is not recommended, lenition of time restrictions allowing additional 

rest days and increasing repetitions would enable further representative results. 

Starting a long-term research project on all four transects, like they have in JCNP, with repeated walks 

several times yearly, would also create a great opportunity of following the changes in habitat quality 

and population densities, as long-term monitoring is the most reliable monitoring method (Hassel- 

Finnegan et al. 2008). Because of the easier access and a functional forest office, most ecology and 

conservation research is based in JCNP. Establishing long term monitoring program would not only 

benefit KPFR, but would also provide diversity to Zanzibar based biological exploration, this not 

only addressing P. kirkii, but the broad spectra of Zanzibar’s idiosyncratic flora and fauna. 

Investigating the surrounding settlements yearly requirement of forest related resources would also 

greatly apprise KPFR’s coming future. It could possibly become the support gaining the needed 

attention upon the assessment of the forest reserves conservation status. 

One of the things that in my opinion could be most interesting to research from here on, is the 

established wildlife corridors. In 2011 wildlife corridors were establish to strengthen the terms of 

representativeness and connectivity of PA in order to preserve ecology and evolutionary processes 

necessary for a continued survival of the unique Zanzibar flora and fauna (Siex 2011). So far there 

has not been any thorough research investigating if the corridors are being used or fulfilling the 

purpose of engagement. 

41



REFERENCES 

Burgess, N., Clarke, G.P. & Rodgers, W.A., 1998. Coastal forests of eastern Africa: status, 

endemism patterns and their potential causes. Biological Journal of the Linnean Society, 

64(3), pp.337–367. 

Bösenberg, J.D., 2010. Encephalartos hildebrandtii. The IUCN Red List of Threatened Species 

2010. Eastern Arc Mountains & Coastal Forests CEPF Plant Assessment Project. 

Available at: http://dx.doi.org/10.2305/IUCN.UK.2010-3.RLTS.T41931A10603165.en 

[Accessed September 8, 2016]. 

CITES, 2016. Convention on International Trade in Endangered Species of Wild Fauna and 

Flora, Geneva, Switzerland. 

Davenport, T.R.B., Nowak, K. & Perkin, A., 2013. Priority Primate Areas in Tanzania. Oryx, 

48(1), pp.39–51. 

Davies, G. & Oates, J.F., 1994. Colobine Monkeys: Their Ecology, Behaviour and Evolution, 

Cambridge, UK: Cambridge University Press. 

Gaba, E., 2010. Graphic Lab. Available at: https://commons.wikimedia.org [Accessed October 4, 

2016]. 

Galat-Luong, A. & Galat, G., 2005. Conservation and Survival Adaptations of Temminck’s Red 

Colobus (Procolobus badius temmincki), in Senegal. International Journal of Primatology, 

26(3), pp.585–603. 

Gereau, R.E. et al., 2016. Globally Threatened Biodiversity of The Eastern Arc Mountains and 

Coastal Forests of Kenya and Tanzania. Journal of East African Natural History, 105(1), 

pp.115–201. 

Google Inc., 2016. Google Earth. Available at: https://www.google.com/earth/. 

Groves, C.P., 2007. The taxonomic diversity of the Colobinae of Africa. JASs Invited Reviews 

Journal of Anthropological Sciences, 85, pp.7–34. 

Grubb, P. et al., 2003. Assessment of the Diversity of African Primates. International Journal of 

Primatology, 24(6). 

Hassan, I.H. & Said, T.A.., 2011. Assement / Evaluation of The Community Forest Activities to 

Date: Learning The Best Practices and Problem Areas in Zanzibar, Zanzibar, United 

Republic of Tanzania. 

Hassel-Finnegan, H.M. et al., 2008. How reliable are density estimates for diurnal primates? 

International Journal of Primatology, 29(5), pp.1175–1187. 

42



Hyde, M.A. et al., 2016. Flora of Zimbabwe: Species information: Records of: Dovyalis 

macrocalyx. Available at: http://www.zimbabweflora.co.zw/speciesdata/speciesdisplay.php?species_id=140830&sortvar=qdssort 

[Accessed October 20, 2016]. 

IUCN, 2016. IUCN Redlist. Available at: http://www.iucnredlist.org/ [Accessed November 2, 

2016]. 

IUCN SSC East African Plants Red List Authority, 2013a. Erythrococca berberidea. The IUCN 

Red List of Threatened Species 2013. Available at: 

http://dx.doi.org/10.2305/IUCN.UK.2013-2.RLTS.T179317A1575069.en. [Accessed 

September 13, 2016]. 

IUCN SSC East African Plants Red List Authority, 2013b. Turraea mombassana. The IUCN Red 

List of Threatened Species 2013. Available at: http://www.iucnredlist.org/details/179336/0 

[Accessed September 13, 2016]. 

Kingdon, J. & Happold, D., 2013. Mammals of Africa., London, UK: Bloomsbury Publishing. 

Lovett, J. & Clarke, G.P., 1998. The IUCN Red List of Threatened Species 1998. Available at: 

http://www.iucnredlist.org/details/32890/0 [Accessed October 19, 2016]. 

Lwanga, J.S. et al., 2011. Primate population dynamics over 32.9 years at Ngogo, Kibale 

National Park, Uganda. American Journal of Primatology, 73(10), pp.997–1011. 

Marshall, A.R. et al., 2005. Monkey abundance and social structure in two high-elevation forest 

reserves in the Udzungwa Mountains of Tanzania. International Journal of Primatology, 

26(1), pp.127–145. 

Ministry of Natural Resources and Tourism, 2014. Extending the Coastal Forest Protected Area 

Subsystem in Tanzania, Zanzibar, United Republic of Tanzania. 

Minteer, B.A. & Miller, T.R., 2011. The New Conservation Debate: Ethical foundations, 

strategic trade-offs, and policy opportunities. Biological Conservation, 144(3), pp.945–947. 

Mittermeier, R.A., Anthony B. Rylands, Christoph Schwitzer, Lucy A. Taylor, Federica Chiozza 

and Elizabeth A. Williamson (eds.). 2012. Primates in Peril: The World’s 25 Most 

Endangered Primates 2010–2012. IUCN/SSC Primate Specialist Group (PSG), 

International Primatological Society (IPS), and Conservation International (CI), Arlington, 

USA. 40pp. 

Mittermeier, R.A. et al., 2007. Primates in Peril : The World ’ s 25 Most Endangered Primates , 

2006 – 2008, Conservation International (CI), Arlington, USA. 40pp. 

Mittermeier, R.A., Wallis, J., Rylands, A. B., Ganzhorn, J. U., Oates, J. F., Williamson, E. A., 

Palacios, E., Heymann, E. W., Kierulff, M. C. M., Long Yongcheng, Supriatna, J., Roos, C., 

43



Walker, S., Cortés-Ortiz, L. and Schwitzer, C. (eds.). 2009. Primates in Peril: The World’s 

25 Most Endangered Primates 2008–2010. IUCN/SSC Primate Specialist Group (PSG), 

International Primatological Society (IPS), and Conservation International (CI), Arlington, 

USA. 84pp. 

Myers, N. et al., 2000. Biodiversity hotspots for conservation priorities. Nature, 403(6772), 

pp.853–8. 

National Bureau of Statistics, 2014. Basic Demographic and Socio-Economic Profile, Zanzibar 

Tanzania, Dar es Saleem, United Republic of Tanzania. 

National Research Council, 1981. Techniques for the Study of Primate Population Ecology, 

Washington, D.C.: National Academies Press. Available at: 

http://www.nap.edu/catalog/18646 [Accessed July 21, 2016]. 

Nowak, K., 2007. Behavioural flexibility and demography of Procolobus kirkii across floristic 

and disturbance gradients. Ph.D Dissartation. University of Cambridge, Cambridge, UK. 

Nowak, K., 2008. Frequent water drinking by Zanzibar red colobus (Procolobus kirkii) in a 

mangrove forest refuge. American journal of primatology, 70(11), pp.1081–92. 

Nowak, K., 2013. Mangrove and peat swamp forests: Refuge habitats for primates and felids. 

Folia Primatologica, 83(3–6), pp.361–376. 

Nowak, K., Cardini, A. & Elton, S., 2008. Evolutionary Acceleration and Divergence in 

Procolobus kirkii. International Journal of Primatology, 29(5), pp.1313–1339. 

Nowak, K. & Lee, P.C., 2011. Demographic Structure of Zanzibar Red Colobus Populations in 

Unprotected Coral Rag and Mangrove Forests. International Journal of Primatology, 32(1), 

pp.24–45. 

Nowak, K. & Lee, P.C., 2013. Status of Zanzibar Red Colobus and Sykes’s Monkeys in Two 

Coastal Forests in 2005. Primat Conservation, 2013(27), pp.65–73. 

Nowak, K., Perkin, A. & Jones, T., 2009. Update on Habitat Loss and Conservation Status of the 

Endangered Zanzibar Red Colobus on Uzi and Vundwe Islands. Wildlife Conservation, 

(September Issue). 

Oates, J.F. & Ting, N., 2015. Conservation consequences of unstable taxonomies: The case of 

the red colobus monkeys A. M. Behie & M. F. Oxenham, eds., Canberra, Australia.: ANU 

Press, The Australian National University. 

Onderdonk, D.A. & Chapman, C.A., 2000. Coping with forest fragmentation: The primates of 

Kibale National Park, Uganda. International Journal of Primatology, 21(4), pp.587–611. 

Rabe, L. & Saunders, F., 2014. Community-based Natural Resource Management of the Jozani- 

44



Pete Mangrove Forest : Do They Have a Voice ? Western Indian Ocean Journal of Marine 

Science, 12(2), pp.133–150. 

Saunders, F., 2011. It’s Like Herding Monkeys into a Conservation Enclosure : The Formation 

and Establishment of the Jozani-Chwaka Bay. Conservation Society, 9(4), pp.261–273. 

Schwitzer, C., Mittermeier, R.A., Rylands, A.B., Chiozza, F., Williamson, E.A., Wallis, J. and 

Cotton, A. (eds.). 2015. Primates in Peril: The World’s 25 Most Endangered Primates 

2014–2016. IUCN SSC Primate Specialist Group (PSG), International Primatological 

Society (IPS), Conservation International (CI), and Bristol Zoological Society, Arlington, 

USA. iv+93pp. 

Schwitzer, C. Mittermeier, R. A., Rylands, A. B., Taylor, L. A., Chiozza, F., Williamson, E. A., 

Wallis, J. and Clark, F. E. (eds.). 2014. Primates in Peril: The World’s 25 Most Endangered 

Primates 2012–2014. IUCN SSC Primate Specialist Group (PSG), International 

Primatological Society (IPS), Conservation International (CI), and Bristol Zoological 

Society, Arlington, USA. iv+87pp. 

Siex, K.S., 2011. Protected Area Spatial Planning for Unguja and Pemba Islands, Zanzibar 

World Wide Fund for Nature (WWF) From Wildlife Conservation Society (WCS), Bronx, 

New York. Available at: World Wildlife Fund for Nature (WWF) and Wildlife 

Conservatory Society (WCS) [Accessed October 4, 2016]. 

Siex, K.S. & Struhsaker, T.T., 1999a. Colobus monkeys and coconuts: A study of perceived 

human-wildlife conflicts. Journal of Applied Ecology, 36(6), pp.1009–1020. 

Siex, K.S. & Struhsaker, T.T., 1999b. Ecology of the Zanzibar red colobus monkey: 

demographic variability and habitat stability. International Journal of Primatology, 20(2), 

pp.163–192. 

Siex, K.S. & Struhsaker, T.T., 2016. Piliocolobus kirkii. The IUCN Red List of Threatened 

Species 2016. Available at: http://dx.doi.org/10.2305/IUCN.UK.2016- 

1.RLTS.T39992A92630131.en. [Accessed October 4, 2016]. 

Struhsaker, T.T., 2005. Conservation of Red Colobus and Their Habitats. International Journal 

of Primatology, 26(3), pp.525–538. 

Struhsaker, T.T. et al., 2004. Demographic Variation Among Udzungwa Red Colobus in 

Relation to Gross Ecological and Sociological Parameters. International Journal of 

Primatology, 25(3), pp.615–657. 

Struhsaker, T.T., 2000. The effects of predation and habitat quality on the socioecology of 

African monkeys: lessons from the islands of Bioko and Zanzibar. In P. F. Whitehead & J. 

45



J. Clifford, eds. Old world monkeys. Canbridge, UK: Cambridge University Press, pp. 393– 

430. 

Struhsaker, T.T., 1975. The Red Colobus Monkey G. B. Schaller, ed., Chicago, USA: The 

University of Chicago Press. 

Struhsaker, T.T., 2010. The Red Colobus Monkey - Variation in Demography, Behavior and 

Ecology of Endangered Species 1st ed., New York, USA: Oxford University Press. 

Struhsaker, T.T., Cooney, D.O. & Siex, K.S., 1997. Charcoal consumption by Zanzibar red 

colobus monkeys: its function and its ecological and demographic consequences. 

International Journal of Primatology, 18(1), pp.61–72. 

Struhsaker, T.T. & Siex, K.S., 2016. Piliocolobus kirkii. The IUCN Red List of Threatened 

Species 2016. Available at: http://dx.doi.org/10.2305/IUCN.UK.2016- 

1.RLTS.T39992A92630131.en [Accessed November 25, 2015]. 

Thomas, L. et al., 2010. Distance software: Design and analysis of distance sampling surveys for 

estimating population size. Journal of Applied Ecology, 47(1), pp.5–14. 

Whitesides, G. et al., 1988. Estimating Primate Densities from Transects in a West African Rain 

Forest: A Comparison of Techniques G. H. Whitesides; J. F. Oates; S. M. Green; R. P. 

Kluberdanz. Journal of Animal Ecology, 57(2), pp.345–367. 

46



APPENDIX 

Appendix cover additional material outside of the original thesis, which I find relevant for the 

greater understanding of this study. 

47



Appendix 1 – 

Study site map 

Figure 17 Map of area 

surrounding The Kiwengwa – 

Pongwe Forest Reserve. The 

reserve border is outlined in 

white. A total of 10 local 

villages border the reserve. 

Transects start and end points 

are marked with flags and red 

line. Yellow line marks main 

road. 

48



Appendix 2 - Colobus and food species maps 

Figure 18 Satellite photo of study site with descriptive layer added. The layer colors are assigned by a 

mathematical model giving similar areas the same color based on aerial footage. Data on colobus group 

size encountered on each transect, as blue circles. Data layer is from 2014. 

49



¯ 

Legend 

Sum of Fields 

1.400 

Food_Count 

NoFood_Count 

0 0,25 0,5 1 1,5 2 

Kilometers 

Figure 19 Satellite photo of study site with descriptive layer added. The layer colors are assigned by a 

mathematical model giving similar areas the same color based on aerial footage. Data on proportion of 

colobus food species within each plot. Diagram size reflecting plot total stem number. 

50



Appendix 3 – Detection functions g(x) 

Figure 20 2004 Observation distances and detection function g(x) Distance sampling method 

Figure 21 2016 Oobservation distances and detection function g(x) Distance sampling method 

51



Figure 22 2004 Observation distances and detection function g(x) Whiteside method. 

Figure 23 2016 Observation distances and detection function g(x) Whiteside method 

52



Appendix 4 – Cluster size distribution 

Figure 24 Frequencies and cumulative frequencies distribution and sized of clusters observed in 2004 and 

2016 overall and by transect. 

53



Appendix 5 - Statistics 

Table 4 Pared students t-test for paired samples on vegetative differences between 2004 and 2016. 

Wilcoxon matched-pairs signed rank test for proportion of cut stems DBH out of plot total. 

Mean 2004 Mean 2016 T(DF=33) P-value Significant 

Stems in plot 261,21 ± 87,84 155,62 ± 55,66 6,83



Table 6 Kruskal – Wallis and Dunn’s multiple comparison test summarized. Only including significant 

result from Kruskal-Wallis for further analysis with Dunn’s multiple comparison test. 

Multiple comparison test results summary 

Stems in plot 

No. cut stems in plot 

Kruskal-Wallis test P-value Signific Sum Kruskal-Wallis test P-value Signific Sum 

ant? mary 

ant? mary 

2004 0,0102 Yes * 2004 0,019 Yes * 

Dunn's multiple Mean Signific Sum Dunn's multiple Mean Signific Sum 

comparisons test rank diff, ant? mary comparisons test rank diff, ant? mary 

B 2004 vs. K2 2004 -13,77 Yes ** B 2004 vs. K2 2004 -11,05 No ns 

B 2004 vs. K3 2004 -11,32 Yes * B 2004 vs. K3 2004 -12,46 Yes * 

K2 2004 vs. K3 2004 2,453 No ns K2 2004 vs. K3 2004 -1,415 No ns 

Cut stems DBH 

Proportion cut stems DBH of total 

Kruskal-Wallis test P-value Signific Sum Kruskal-Wallis test P-value Signific Sum 

ant? mary 

ant? mary 

2004 0,0188 Yes * 2004 0,0058 Yes ** 

2016 0,0232 Yes * Dunn's multiple Mean Signific Sum 

comparisons test rank diff, ant? mary 

Dunn's multiple Mean Signific Sum B 2004 vs. K2 2004 -14,78 Yes ** 

comparisons test rank diff, ant? mary 

B 2004 vs. K2 2004 -12,87 Yes * B 2004 vs. K3 2004 -11,43 Yes * 

B 2004 vs. K3 2004 -10,43 No ns K2 2004 vs. K3 2004 3,352 No ns 

K2 2004 vs. K3 2004 2,44 No ns 

B 2016 vs. K2 2016 8,615 No ns 

B 2016 vs. K3 2016 12,64 Yes * 

K2 2016 vs. K3 2016 4,027 No ns 

55



Appendix 6 – Plot specific species decline



Appendix 7 - Species list 

Colobus food species 

highlighted 

Adenia gummifera 

Albizia glaberrima 

Allophylus aldabricus 

Allophylus parvillei 

Allophylus rubifolius 

Allophylus sp. 

Ancylobotrys petersiana 

Annona senegalensis 

Apodytes dimitiata 

Aporrhiza paniculata 

Berchermia discolor 

Bersama abyssinica 

Blighia unijugata 

Bridelia cathartica 

Bridelia micrantha 

Camptolepsis ramiflora 

Carpodiptera africana 

carpolobia goetzei 

Cassytha filiformis 

Cathium mombassica 

Citrus sinensis 

Clausena anisata 

Clerodendrum glabrum 

Clerodendrum 

myricoides 

Cocos nucifera 

Cremaspora triflora 

Croton 

pseudopulchellus 

Cussonia zimmermannii 

Dalbergia vaccinifolia 

Deinbollia borbonica 

Dichrostachys cinerea 

Dioscorea sansibarensis 

Diospyros abyssinica 

Diospyros consolatae 

Diospyros ferrea 

Diospyros natalensis 

Dodonaea viscosa 

Dovyalis macrocalyx 

Dovyalis spp. 

Drypetes natalensis 

Ehretia amoena 

Encephalartos 

hildebrandtii 

Erythrococca berberidea 

Euclea natalensis 

Euclea racemosa 

Euclea schimperi 

Eugenia capensis 

Euphorbia nyikae 

Ficus exasperata 

Ficus ingens 

ficus natalensis 

Ficus scasselatii 

Ficus sur 

Flacourtia indica 

Flacourtia spp. 

Flueggea virosa 

Grewia bicolor 

Grewia mollis 

Harrisonia abyssinica 

Hensia zanzibarica 

Hoslundia opposita 

Ixora narcissodora 

Jasminum fluminense 

Jusminum mauritianum 

Lannea schweinfurthii 

Lantana camara 

Lawsonia inermis 

Lecaniodiscus 

fraxinifolius 

Lepisanthes senegalensis 

Leptactina platyphylla 

Ludia mauritania 

Macphersonia gracilis 

Mallotus oppositifolius 

Mangifera indica 

Manilkara sulcata 

Margaritaria discoidea 

Maytenus andata 

Maytenus heterophylla 

Maytenus 

mossambicensis 

Maytenus spp. 

Mdalasini mwitu 

Mfuka duri 

Mimusops fruticosa 

Mkekundu 

Mkomba 

Mkuni 

57



Monanthotaxis 

fornicata 

Monodora grandidieri 

Monothotaxis 

trichocarpa 

mpanduka female 

Mrabundi 

Musa sp. 

Mystroxylon 

aethiopicum 

Mystroxylon 

mombaciana 

Olea woodiana 

Ozoroa obovata 

Pavetta gerstneri 

Phyllanthus reticulatus 

Pittosporum viridiflorum 

Polyspheria multiflora 

Polyspheria parvifolia 

Psiadia arabica 

Psychotria 

bibracteatum 

Psychotria goetzei 

Psychotria sp. 

Rapanea melanophloeus 

Rausonia lucida 

Rhoicissus revoilii 

Rhus longipes 

Rhus natalensis 

Salacia elegans 

Senna petersiana 

Senna sp. 

Sideroxylon inerme 

Sorindeia 

madagascariensis 

Stadmania oppositifolia 

Sterculia rhynchocarpa 

Strychnos angolensis 

Strychnos spinosa 

Strychnos sp. 

Suregada zanzibarensis 

Synaptolepsis kirkii 

Tarenna pavettoides 

Teclea nobilis 

Terminalia boivinii 

Thylachium densiflora 

Toddalia asiatica 

Toddalia sp. 

Trema orientalis 

Tricalysia microphylla 

Tricalysia ovalifolia 

Turraea floribunda 

Turraea sp. 

Vernonia zanzibarensis 

Ziziphus robertsiana 

Unknown 

Unknown climber species 

Unknown stump 

Unidentifiable 

58



Appendix 8 - Correlations 

Table 7 Correlation analysis statistical parameters 

VARIABLE OBSERVATIONS OBS. 

WITH 

MISSING 

DATA 

DBH OF LIVE STEMS 

2004 

DBH OF LIVE STEMS 

2016 

N TREES DBH > 10 CM 

2004 


2016 


OBS. 

WITHOUT 

MISSING 

DATA 

MINIMUM MAXIMUM MEAN ±SD 

34 0 34 431,1 1356,2 1009,4 217,8 

34 0 34 307,9 1097,0 671,7 203,7 

34 0 34 11,0 44,0 26,4 8,2 

34 0 34 0,0 27,0 11,5 7,7 

34 0 34 0,0 13,0 3,3 3,6 

2016 

N STEMS 2004 34 0 34 116,0 507,0 261,2 89,2 

N STEMS 2016 34 0 34 71,0 319,0 155,6 56,5 

N CUT STEMS 2016 34 0 34 0,0 39,0 19,5 10,0 

DBH OF CUT STEMS 

34 0 34 0,0 410,2 122,0 83,7 

2016 

DISTURBANCE 34 11 23 1,0 7,0 2,8 1,6 

FEWER STEMS 34 0 34 -281,0 137,0 -94,1 95,3 

SPECIES LOST 34 0 34 1,0 11,0 5,0 2,3 

METERS FROM RIM 34 0 34 0,0 1400,0 591,2 419,3 

N FOOD SPECIES 2004 34 0 34 63,0 298,0 139,5 57,5 

N FOOD SPECIES 2016 34 0 34 14,0 163,0 46,7 26,9 

FOOD SPECIES PROP. 

34 0 34 0,3 0,7 0,5 0,1 

2004 

FOOD SPECIES PROP. 

34 0 34 0,1 0,6 0,3 0,1 

2016 

N COLOBUS 2016 34 13 21 1,0 38,0 13,7 10,0 

N COLOBUS 2004 34 23 11 4,0 44,0 22,9 12,0 

59



Table 8 List of all positive correlations 

Positive Correlations 

N colobus 2004 

N food species 

2016 

N stems 2016 

DBH of live stems 

2016 


2016 

DBH live stems 

2016 



2016 


N trees DBH > 

10 cm 2016 


2004 

N trees DBH > 10 

cm 2016 


2016 

Food species prop. 

2004 

Food species 

prop. 2016 

Food species Prop. 

2016 


2016 


2004 


2016 

N stems 2016 


2016 

N stems 2016 


2004 

N stems 2004 

N stems 2004 


2004 


2004 

Food species 

prop. 2004 


2004 


cm 2016 

N colobus 2004 N stems 2016 


2016 



2016 


cm 2016 


2016 


cm 2016 


2016 


cm 2004 


2016 


cm 2004 


2016 

Fewer stems 


2004 


cm 2016 

Food species 

Prop. 2016 


2016 


cm 2016 

N colobus 2016 N stems 2016 DBH of live stems 

2016 


cm 2004 

Meters from rim N colobus 2004 N cut stems 

2016 

DBH of cut stems 

2016 

N stems 2016 Species lost DBH cut stems 

2016 

DBH of cut 

stems 2016 

N cut stems 2016 


2004 

Fewer stems N trees DBH > 25 

cm 2016



Table 9 List of all negative correlations 

Negative correlations 

N colobus 2016 N stems 2016 Species lost Meters from rim Fewer stems N food species 

2004 

Meters from rim Species lost N stems 2004 DBH cut stems 

2016 

Food species 

Prop. 2016 

DBH of cut stems 

2016 

Food species 

prop. 2016 



cm 2016 

Fewer stems 


2004 


2004 

N stems 2006 

Fewer stems 


2004 

61



62

Endangered Zanzibar Red Colobus Piliocolobus Kirkii

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?