Does the matrix matter? A forest primate in a complex agricultural landscape

B I O L O G I C A L C O N S E RVAT I O N

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/biocon

Does the matrix matter? A forest primate in a complex agricultural landscape Julie Andersona,b,*, J. Marcus Rowcliffea, Guy Cowlishawa a

Institute of Zoology, Zoological Society of London, Regents Park, London, NW1 4RY, UK Wakuluzu, Friends of the Colobus Trust, P.O. Box 5380, 80401, Diani Beach, Kenya

b

A R T I C L E I N F O

A B S T R A C T

Article history:

Many threatened primates now exist in fragmented forest habitats. The survival of these

Received 12 April 2006

populations may depend on their ability to utilise agricultural or other matrix habitats

Received in revised form

between forest fragments, but this is poorly known. Here, we systematically investigate

4 October 2006

an arboreal primate’s use of a heterogeneous matrix in a fragmented forest landscape:

Accepted 12 October 2006

the Angola black-and-white colobus (Colobus angolensis palliatus) in southern Kenya. We used a novel technique, based on semi-structured interviews with local informants, to address the difficulty of sampling relatively rare but important events, such as dispersal

Keywords:

between fragments. We found that colobus frequently travelled and foraged in indigenous

Colobus angolensis

matrix vegetation (such as mangrove, wooded shrubland and shrubland) up to 4 km from

Primate

the nearest forest fragments. Agricultural habitats, such as perennial plantation (coconut,

Matrix

mango and cashew nut) were also used by colobus for travelling and foraging (in remnant

Forest fragmentation

indigenous trees). The probability of sighting colobus in the matrix increased with the pro-

Agricultural landscape

portion of both tall (>6 m) vegetation cover and food tree cover, but declined with distance from forest habitat. Our findings suggest that certain matrix habitats are important for C. a. palliatus, and that future primate conservation initiatives might benefit from adopting a ‘landscape-level’ approach to habitat management, particularly in fragmented forest systems. Ó 2006 Elsevier Ltd. All rights reserved.

1.

Introduction

Primates are a highly threatened taxon, dependent upon tropical forest ecosystems. Almost 90% of all primate species are found within this biome (Mittermeier and Cheney, 1987), and more than half of the 250 existing species are considered to be of conservation concern by the Primate Specialist Group of the World Conservation Union (IUCN, 2005). Habitat loss and fragmentation is currently the greatest threat facing these taxa. One in four species are either Endangered or Critically Endangered as a consequence of habitat loss, and without better protection these species may be extinct in the next 20

years (Mittermeier, 1996; IUCN, 2005). Some of the most threatened primate communities now survive only in fragmented forest habitats (Cowlishaw and Dunbar, 2000; Marsh, 2003). It is important therefore, to understand the ecological flexibility or limits of such communities and species, in order to implement effective management strategies for their future conservation (Harcourt, 1998, 2002; Lindenmayer, 1999). Research into tropical deforestation and its effects on primate populations has largely focused on three major processes: the degradation of habitat, its reduction in area, and its insularisation (Andren, 1994; Marsh, 2003). This research has shown that the quality and spatial characteristics of

* Corresponding author: Tel.: +44 079 0575 9312; fax: +44 020 7483 2237. E-mail addresses: [email protected] (J. Anderson), [email protected] (J.M. Rowcliffe), [email protected] (G. Cowlishaw). 0006-3207/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2006.10.022

B I O L O G I C A L C O N S E RVAT I O N

forest fragments play an important role in understanding any remnant population’s behaviour, density, viability, and ultimately conservation and management (Chapman and Lambert, 2000; Fahrig, 2003; Marsh, 2003). However, to date, most if not all primate research has focused within the boundaries of forest fragments (Chapman and Peres, 2001). The complex mosaic of vegetation types that lies between forest patches has received little attention in these analyses. Often referred to as ‘‘matrix’’ habitat (Gascon et al., 1999; Ricketts, 2001), these non-forest vegetation types are generally believed to be of limited use to forest-dwelling primates. The wider importance of the matrix has recently been highlighted by the integration of conventional habitat-fragmentation studies with the complementary discipline of landscape ecology, i.e. the study of landscape structure and its effects on ecological processes (Turner, 1989). This approach has been highly instructive, demonstrating the importance of studying fragments within a larger, dynamic landscape mosaic (Fahrig and Merriam, 1994). For example, matrix composition and connectivity have been found to influence the dispersal, diversity, abundance and persistence of a variety of diverse taxa, including insects (e.g. Goodwin and Fahrig, 2002; Bonte et al., 2003), birds (e.g. Aberg et al., 1995; Wethered and Lawes, 2003) and mammals (e.g. Gascon et al., 1999; Laurance and Laurance, 1999; Pardini, 2004). However, this approach has yet to be considered for primates. Some primates exhibit great ecological and behavioural flexibility in response to their changing environment. For example, anthropogenic land transformation has allowed some species to exploit agricultural fields in the matrix on the boundaries of their natural habitats (Cowlishaw and Dunbar, 2000). Subsistence crops offer an alternative food source to most frugivorous cercopithecine and pongid species (Naughton-Treves, 1998; Reynolds et al., 2003), and occasionally colobines (Naughton-Treves, 1998). Indigenous vegetation in the matrix can also provide additional food resources, in the form of mangrove foliage (Siex, 2003) or secondary vegetation in fallow fields and coastal shrubland areas (MorenoBlack and Maples, 1977). Anecdotal reports also exist of arboreal primates dispersing through plantations (Laidlaw, 2000; Umapathy and Kumar, 2000; Olupot and Waser, 2001; Li, 2004) and wooded shrubland (Marsh, 1979; Wieczkowski, 2004). In most of these studies, the rarity of primate dispersal between groups, and the difficulties of surveying primates within the matrix, have prevented a more focused study of this phenomenon. If indeed ‘the matrix matters’ (Ricketts, 2001), further investigation into its potential effects may be vitally important for the future conservation management of fragmented primate populations. The Kwale District of southern Kenya provides an ideal landscape for the study of matrix quality and its significance to arboreal primates. Firstly, Angola black-and-white colobus (Colobus angolensis palliatus) populations inhabit 55 out of the 124 coastal forest fragments in this region (Anderson et al., in press (a)), and it is largely folivorous and forest-dependent (Moreno-Black and Maples, 1977; Davies, 1994). Indeed, canopy structure has a significant influence on the incidence of C. a. palliatus populations within patches (Anderson et al., in press (b)), highlighting its dependency on closed-canopy veg-

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

213

etation. Secondly, these 124 forest fragments are threatened by anthropogenic landscape change, and the entire area is listed as one of 11 ‘priority regions’ for international conservation investment due to the ongoing forest destruction (Brooks et al., 2002). The fragments are currently surrounded by a heterogeneous mix of agriculture, urban development and indigenous (non-forest) vegetation. Thirdly, the regional dynamics of C. a. palliatus patch populations may also function at a metapopulation level (Anderson, 2005). Since this has important implications for conservation management, any information regarding limited colobus dispersal between forest patches, that would help to elucidate this possibility, would be valuable (Hanski, 1994). In this context, this study has three main aims. To (1) provide a novel approach to data collection within the matrix (in recognition of the difficulties of collecting systematic data on what may be rare but important patterns of habitat use) utilising local informants as surveyors of their own landscape, (2) investigate the occurrence and behaviour of C. a. palliatus in non-forest matrix habitats using these methods, and (3) identify the key habitat attributes that determine C. a. palliatus usage of the matrix. This information will contribute to the development of future C. a. palliatus conservation management plans within Kenya.

2.

Methods

2.1.

Study site

The Kwale District, in the Coastal Province of Kenya, lies between Mombasa and the border of northeastern Tanzania (3°30 0 , 4°45’S; 38°31 0 and 39°31’E). The District (8322 km2) is largely an agro-ecological zone (Muchoki, 1990) with only 3% of land cover supporting closed-canopy indigenous coastal forest (i.e. 255 km2, within 124 forest fragments: Anderson, 2005). Human development and agricultural activities have given rise to a heterogeneous mix of ‘matrix’ land cover. Indigenous matrix comprises mangrove, coastal shrubland, shrub grassland and wooded grassland areas. Historical human land use has transformed coastal forest (and other indigenous vegetation) into pastoral grasslands, perennial plantations (coconut, cashew nut and mango), timber plantations, annual croplands (e.g. maize, rice, sugar cane and root vegetable crops) and areas of human development and settlement (Muchoki, 1990). Most agriculture is locally managed. Largescale commercial production of sugar cane, bixa and cashew nut collapsed before 1990 with the closure of the District’s main agro-factories (Anderson, 2005). For the purpose of this study, we classified the matrix into 17 categories (Table 1). These were based on gross structural characteristics, i.e. primary vegetation type (if any), canopy cover and height (Grunblatt et al., 1989), and were further grouped into ‘indigenous’ and ‘human land use’ types.

2.2.

Sampling colobus in the matrix

Since observations of C. a. palliatus in the matrix were expected to be relatively rare events, standard survey techniques (e.g. line transects) were considered inappropriate. We therefore devised an alternative approach, in which two

214

B I O L O G I C A L C O N S E RVAT I O N

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

Table 1 – Matrix types within the Kwale District, Kenya Matrix type

Matrix structurea (height)

Examples of matrix composition

80–100% closed-canopy tree coverage (>10 m), remnants of the coastal forestsb 80–100% closed-canopy mangrove (>10 m), encompassing eight tree speciesc 50–79% dense shrubs (1–6 m), 20– 49% indigenous coastal trees (>10 m) 50–79% dense grass (<1 m), 20– 49% indigenous coastal trees (>10 m) 80–100% closed shrub (1–6 m), 2– 19% indigenous coastal trees (>10 m) 50–79% dense grass (<1 m), 20– 49% shrubs

Adansonia digitata, Combretum schumannii, Lecaniodiscus fraxinifolius, Ficus spp. Rhizophora mucronata, Ceriops tagal, Bruguiera gymnorrhiza, Avicennia marina. Grewia spp., Brachystigia spiciformis, Cynometra webberi, Paramacrolobium spp. Acacia spp., Commiphora edulis, Hyphaene compressa, Terminalia spp. Manilkara sulcata, Diospyros conrnii, Croton spp., Dobera glabra, Adenia spp. Acacia spp., Thespia danis, Lantana spp., Annona senegalensis, Phoenix reticulate Sandy or rocky, bare ground only Deposits of coral and sands 630 m from the Phoenix reclinata, Hyphaene spp., Elaeis guineensis, Raphia spp. Indian Ocean, rivers, lakes

1

Indigenous Indigenous coastal forest vegetation

2

Mangrove

3

Wooded shrubland

4

Wooded grassland

5

Shrubland

6

Shrub grassland

7

Bare ground

8

Sand

9

Swamp

10 11

Water Human land use Perennial plantation

12

Timber plantation

13

Annual cropland <1 m

80–100% closed crops (<1 m)d

14

Annual cropland 1–3 m

80–100% closed crops (1–3 m)d

15

Grassland

80–100% closed grass (<1 m)

16

Human development

0% vegetation cover (0 m)

17

Quarry

0% vegetation cover (0 m)

a b c d

0% canopy coverage (0 m), no secondary vegetation (0 m) 0% canopy coverage (0 m), no secondary vegetation (0 m) 2–19% sparse trees (>10 m), 50– 79% water (0 m) 100% water = no vegetation (0 m) 50–79% dense trees (>10 m), 20– 49% grass, shrubs or annual crops (<3 m) 50–79% dense trees (>10 m), 20– 49% grass or shrubs (<3 m)

Anacardium occidentale, Cocus nucifera, Mangifera indica Casuarina equisetifolia, Eucalyptus spp., Pinus spp., Cupressus spp. Solanum tuberosum, Vigna ungiculata, Oryza sativa, Zea mays, Manihot esculenta Saccharum officinarum, Musa spp., Carica papaya, Zea mays, Manihot esculenta Hyparrhenia spp., Digitaria mombasana, Andropogon spp., Setaria spp. Buildings, tarmac roads and settlements Open quarries for coral, lime and minerals

Matrix structure details the % coverage and height (m) of dominant vegetation types (Grunblatt et al., 1989). For more detailed descriptions and lists of tree species see (Burgess et al., 2000; Robertson and Luke, 1993; White, 1983). For descriptions of species see (Richmond, 1997). For descriptions of commonly used cultivated food species see (Maundu et al., 1999).

researchers used semi-structured interviews to gather information about colobus sightings from local informants living in the matrix. Using this method, researchers were utilising the historical knowledge of the interviewees’ experience, treating the local community as the ‘surveyors’ of their own landscape (cf. Van der Hoeven et al., 2004). To achieve representative sampling within the matrix, a median of seven sites (mode = 9, range = 3–9) were selected for each of 10 selected matrix types, i.e. mangrove, shrubland,

wooded shrubland, grassland, shrub grassland, wooded grassland, annual cropland 1–3 m, perennial plantation, timber plantation and swamp. Matrix types were chosen that would provide a wide spectrum of matrix structural differences to compare and contrast. Sites were chosen that represented homogeneous expanses of each matrix type (>10 ha). The numbers of sites selected for each vegetation type were proportional to the relative land cover within the District as a whole (Anderson, 2005).

B I O L O G I C A L C O N S E RVAT I O N

During the period of February until June 2003 a minimum of six interviews were obtained from each sample site. A day was spent at each site, scanning the area for interview candidates who were either actively working in the matrix or living in settlements surrounded by it. Participants were selected on the basis of their knowledge of C. a. palliatus and length of exposure to the matrix. Thus, key informants were locals who spent a significant proportion of their working day outside, e.g. farmers and herdsmen in agricultural and grassland matrix, and fishermen in mangroves. For swamp, timber plantations and mangrove vegetation, participants lived and worked on the edges of these matrix types. When participant sourcing methods proved difficult, snowball sampling was used to locate key informants (Bernard, 2002), i.e. after each interview, participants were asked to name and direct the researchers towards other appropriate informants. All interviews were geo-referenced using a Garmin 12XL global positioning system (GPS) and distances to nearest forest boundaries were calculated using ArcView geographic information system (GIS) (ArcView GIS 3.2, ESRI Inc. 1999; Anderson, 2005) and Nearest Features extension (Version 3.6e, 2001). The forest boundary GIS was derived from a July 2001 GPS mapping survey involving all Kwale District forests (Anderson, 2005). A standardised framework of questions was put to each participant in the semi-structured interview. However, interviews were flexible in terms of the scope, extent, order and emphasis with which different questions were explored. Participant credibility was assessed by quantifying their exposure to the matrix (i.e. total number of hours) in terms of working hours per day, days per week, number of seasons and years. Participants were also asked if they were born at the site, with historical exposure quantified wherever possible. Colobus knowledge was then tested using a threestage process; participants had to list all the monkey species within the District (and mention colobus), describe two things about the colobus (e.g. appearance, uniqueness of white infants, habitat preference, arboreal nature or movement, ecology or behaviour), and lastly, correctly identify the subspecies from five differing primate photo ID cards. Participants were also ranked in accordance to the level of prompting required, i.e. 0 (none) to 3 (excessive). The interview responses used in this analysis were restricted to those participants who expressed a good level of knowledge about the colobus (with little or no prompting) and spent P3000 h (i.e. the equivalent of 1 year, 5 days/week, 4 h/day) within the matrix. All other interviewees were excluded. Participants who were classed as ‘incidental’ in the matrix (i.e. those who lived within the matrix, but worked daily elsewhere), were also excluded. The interviews established if participants had observed colobus within their lifetime at the sample site, and recorded details of the frequency of sightings both within the past year (i.e. weekly, monthly, rarely, never) and historically. In all cases, the participant’s exact response was documented in detail and summarised by researchers post-hoc. Rare sightings were described by the number of sightings/ year. If colobus sightings were reported, the following details were discussed: (1) the number of animals observed, (2) first

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

215

behaviour observed, (3) pattern of locomotion (e.g. types of arboreal, terrestrial), (4) height above ground, support use, and speed of travel (feeding/resting, slow and fast), and (5) reasons for movement. Responses to the latter question were later summarised by researchers as either ‘travelling’, ‘feeding’ (including details of food types), or ‘chased’ (e.g. dogs, people and throwing stones). In some instances, participants could take researchers to the exact location of the colobus sighting, such as a feeding tree, and these points (n = 53) were also geo-referenced using a GPS. Using the above GIS, the minimum distance (km) of each geo-referenced colobus sighting to the nearest forest boundary was calculated. Only sightings within the past two years were used for this analysis (n = 48), to ensure comparability with the 2001 forest boundary maps. Interviews ended with a discussion on primates as local agricultural pests, to determine the severity of conflict between people and colobus. Sometimes participants had no direct experience of observing colobus in the sample matrix. However, they had observed colobus outside the forest, within matrix elsewhere in the Kwale District. In these instances, the same interview format was adhered to, gathering as much information as possible about these additional sightings. A total of 386 interviews were completed, although only 347 reports were used based on the ‘credibility’ filter regarding the participant’s knowledge and exposure. The majority of participants were male (76%) and between the ages of 40 and 60 years (53%). Occupations of participants were either farmers (66%), fishermen (12%), herdsmen (10%), or miscellaneous outdoor workers, e.g. kiosk owners, medicine men, loggers, carpenters and maintenance workers (12%), with a history of working an average of 85,000 h ± 4,000 SE (range = 3000–481,000 h) within, or on the edge of, their respective matrix sites.

2.3.

Matrix vegetation surveys

At each sample site, four intercept-vegetation transects (Grieg-Smith, 1983; Bullock, 1996), each measuring 150 m in length and 5 m in width, were randomly placed and walked (with positions logged by GPS). Using this method, the beginning and end of all noticeable gaps in the upper vegetation canopy (if any) were noted, and at every 5 m the height (m) of the uppermost canopy level and the presence of (1) colobus food trees (taken from Anderson, 2005), (2) bare ground, (3) dwarf <1 m (short grass, crops, shrubs), (4) low 1–3 m (overgrown grass, crops, shrubs) and (5) tall >6 m (shrubs, trees) vegetation were recorded. Crop and canopy tree species (i.e. indigenous or perennial trees >6 m) were also recorded if their canopy was directly above the 5 m sampling track. The coverage of individual vegetation layers (2–5) were recorded as it was hypothesised that these variables could influence both colobus movement and local (observer) visibility of this large primate. Human paths, roads and rivers that intersected transects were also recorded as possible barriers. At a small number of sites, transects were not conducted due to difficult terrain (e.g. surface water, thick vegetation barriers); these were swamps (no sites sampled), and mangrove and coastal shrubland (two sites sampled each). No primates were observed during vegetation sampling.

216

2.4.

B I O L O G I C A L C O N S E RVAT I O N

Statistics

Using the statistical software R, version 1.9.1 (www.r-project.org), a stepwise generalised linear model (GLM) analysis with binomial error structure (Crawley, 1993) identified the key attributes that influence the frequency of C. a. palliatus sightings within the matrix. Each matrix site was regarded as a unit, with the proportion of colobus sightings weighted as a ‘two-vector’ response variable (i.e. total number of ‘positive’ responses were bound together with the total number of ‘negative’ responses into a single object). Using data gathered from the intercept vegetation transects it was possible to calculate the average canopy height (m), number of canopy gaps, gap length (m), proportion of bare ground cover, dwarf (<1 m), low (1–3 m), and tall (>6 m) vegetation cover, and proportion of colobus food tree cover for 47 matrix sites, across nine matrix types. The distances (km) of these 47 matrix sites to nearby forest were also estimated from the average distances of site interviews (n = 47 sites, n = 165 interviews). ‘Human activity’ levels within each of the sites were calculated as the average number of hours/year participants worked within each site, and were included to avoid the possibility that variation in exposure might confound the effects of other variables. The habitat structure, distance and human activity variables for each site were subsequently entered into a full model, then sequentially removed, least significant first, until a minimal adequate model was reached. Statistical significance (p 6 0.05) was tested using deletion F-tests.

3.

Results

3.1.

Colobus in the matrix

A total of 98 out of 347 reports documented colobus within the matrix (Table 2). These sightings were divided into two types. Firstly, 54 reports (15% of all reports) were specific to 23 matrix

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

sites (i.e. 34% of all sites visited). Since it was difficult to rule out the possibility of pseudoreplication within sites (i.e. participants within the same site observing the same colobus individuals), our data are summarised by the number of ‘site’ reports as well as ‘interview’ reports. Secondly, 44 participants gave information regarding matrix colobus sightings in other regions of the Kwale District. All 98 reports were used to document the qualitative nature of colobus behaviour within the matrix. However, only the 54 colobus sightings connected to sample sites were used to estimate sighting frequencies because these reports could be related to the participants’ matrix exposure time at each site. The patterns of colobus incidence showed a strong pattern of variation across matrix types, which was consistent whether the data were summarised by site or by interview (Table 2). Both wooded shrubland and shrubland were sampled less intensively due to the difficulty of finding large homogeneous areas of these remnant vegetation types (Table 2). Nevertheless, an overall trend is apparent, whether one looks at the proportion of sites, interviewees (at those sites), or interviewees with additional reports. Namely, C. a. palliatus is most frequently observed within three main matrix types: wooded shrubland, perennial plantation (mixed coconut, cashew nut and mango), and mangrove. In contrast, colobus were never reported within grassland, shrub grassland or swamp vegetation. Despite the low sampling effort in shrubland areas, colobus sightings were reported within half of these sites. The significance of this fourth matrix type is highlighted more clearly when sightings frequencies are summarised for each matrix type (Fig. 1). Within all wooded shrubland, mangrove and shrubland sites where colobus were reported, all participants gave accounts of regular colobus sightings (i.e. weekly or monthly, within the past year). Interviewees in four out of seven perennial plantation sites also reported regular colobus sightings, although the remaining sites only reported rare encounters.

Table 2 – Colobus presence in the matrixa Matrix type

Sites ns

Individual interviews

nps (propn)

ni

npiSITE (propn)

npiEXTRA

Total

Wooded shrubland Perennial plantation Mangrove Shrubland Wooded grassland Timber plantation Annual cropland 1–3 m Grassland Shrub grassland Swamp

3 9 9 4 8 6 7 7 6 9

3 7 6 2 3 1 1 0 0 0

(1.00) (0.78) (0.67) (0.50) (0.37) (0.17) (0.14) (0) (0) (0)

11 50 46 17 46 32 37 33 35 40

8 (0.73) 20 (0.40) 16 (0.35) 2 (0.12) 6 (0.13) 1 (0.03) 1 (0.03) 0 (0) 0 (0) 0 (0)

10 24 5 2 3 0 0 0 0 0

18 44 21 4 9 1 1 0 0 0

Total

68

23 (0.34)

347

54 (0.15)

44

98

nps (propn), total number (and proportion) of positive sites where colobus were observed in the matrix. npiSITE (propn), total number (and proportion) of positive interviews reporting colobus in the matrix, where interview reports relate specifically to the sample site. npiEXTRA, additional reports of colobus in the matrix elsewhere in the Kwale District. Total, npiSITE + npiEXTRA. a Results of semi-structured interviews, summarising the total number of matrix sites visited (ns) and the total number of interviews completed (ni) for each matrix type.

B I O L O G I C A L C O N S E RVAT I O N

217

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

1.2

Rare sightings

(3)

Proportion of matrix sites reporting colobus

1.0

Regular sightings within past year (9)

0.8 (9) 0.6

(4) (8)

0.4 (6) 0.2

(7)

(7)

AC

Gr

(6)

(9)

0.0 WSh

Per

Mang

Shrub

WGr

TP

ShGr

Sw

Matrix type Fig. 1 – Proportion of matrix sites reporting colobus. Matrix types, WSh; wooded shrubland; Per, perennial plantation; Mang, mangrove; Shrub, shrubland; WGr, wooded grassland; TP, timber plantion; AC, annual crops 1–3 m; Gr, grassland; ShGr, shrub grassland and Sw, swamp. Regular sightings, the proportion of sites where the modal interview response to colobus sighting frequency (within the past year) was ‘weekly’ or ‘monthly’; Rare sightings = the proportion of sites where sightings occurred only once or twice within the experience of the participant. Details of all rare encounters of colobus were as follows: three perennial plantation sites (only once within 16, 23 and 38 years respectively), and one wooded grassland (once in 48 years), timber plantation (once in 8 years) and annual cropland site (twice in 54 years).

3.2.

Colobus activity in the matrix

Both solitaries (10%) and colobus groups (90%) were observed within the matrix (n = 98). Locomotion was primarily arboreal, namely climbing, quadrupedalism and leaping (90%), although quadrupedal ground movement was also observed (10%). During arboreal locomotion, the supporting branches of mangroves, coastal shrubs, timber tree spp., indigenous tree spp., cashew nut and mango trees were utilised, as well as the fronds of coconut palms. Colobus activity (classified from the first behaviour observed) within differing matrix types is summarised in Table 3. Feeding activity was the most frequently observed colobus behaviour, occurring largely within wooded shrubland, perennial plantations, shrubland and mangrove. Colobus were reported to eat the leaf buds and young leaves of Rhizophora mucronata, Heritiera littoralis and Ceriops tagal in mangrove. Within perennial plantations, colobus were reported to feed primarily on the leaves of indigenous trees, which were retained at these sites as sacred trees, shade trees, meeting places, or sources of fruit and/or medicinal products. Colobus were also reported to feed on the exotic timber trees Ceiba pentador and Azadirachta indica, and ornamental species Delonix regia and Bougainvillea spectabilis, within this matrix type. It was difficult to determine whether C. a. palliatus travelling activity reflected dispersal between habitat patches, because colobus may have entered the matrix solely for

foraging purposes. This explanation might certainly apply in those cases where colobus were initially seen travelling, but subsequently observed to arrive at an indigenous food tree and proceed to feed (60% of observations). Nevertheless, it is also true that animals travelling between forest patches would be likely to feed when passing food trees during their journey. Unfortunately, the available data are insufficient to distinguish between these two explanations. To error on the side of caution, we therefore view all instances of ‘travelling’ as supporting evidence for the potential of colobus dispersal through the matrix, rather than direct evidence of dispersal itself. All interviewees stated that C. a. palliatus was not a significant agricultural pest compared with the more frugivorous

Table 3 – Colobus activity in the matrix Activitya

Matrix type Feeding Wooded shrubland Perennial plantation Mangrove Shrubland Wooded grassland Timber plantation Annual cropland 1–3 m Total sightings

15 (0.83) 27 (0.61) 10 (0.48) 3 (0.75) 3 (0.33) – – 58 (0.59)

Resting

Travel

– 2 (0.05) 3 (0.14) – – – – 5 (0.05)

3 (0.17) 15 (0.34) 8 (0.38) 1 (0.25) 4 (0.66) 1 (1.00) 1 (1.00) 35 (0.36)

a Data under each column represents the total number of local sightings within the matrix (n = 98) where C. a. palliatus feeding, resting or travelling activity was observed. Figures in parentheses indicate each colobus activity as a proportion of the total number of sightings in each matrix type.

218

B I O L O G I C A L C O N S E RVAT I O N

primates, i.e. the yellow baboon (Papio hamadryas cynocephalus), Syke’s monkey (Cercopethicus mitis albogularis), and vervet monkey (Chlorocebus aethiops). However, within the southernmost region of Kwale, colobus were reported to periodically eat the skins of unripe mangos (Mangifera indica) and oranges (Citrus sinensis), and the leaves of cassava (Manihot esculenta), sweet potato (Ipomoea batatas) and cow pea (Vigna unguiculata) crops. These reports were unique to this area. Analysis of GPS sighting locations data indicated that colobus were found to travel within the matrix up to 4.2 km from the nearest forest boundaries with a median travel distance of 0.6 km (range = 0.07–4.2 km, n = 48). There was no significant difference between travel distances of solitaries (median = 1.6 km, range 0.2–4.2 km, n = 6) and colobus groups (median = 0.6 km, range 0.07–4.1 km, n = 42) (Mann–Whitney test: U = 72, Z = 1.68, p = 0.096). Since travel distances were calculated as the shortest (Euclidean) distances from nearby forest, it is possible that this method is underestimating the potential range of true travelling distances for C. a. palliatus. Individuals are seldom likely to disperse in straight lines but rather follow routes that would be influenced by matrix preferences (Opdam, 1990; Bennett, 1998).

3.3.

Predicting colobus use of the matrix

GLM analyses indicate that C. a. palliatus became increasingly rare with distance from forest edge (Table 4). When controlling for distance, two matrix-quality variables also proved important: colobus were more common in areas of tall (>6 m) vegetation cover and food tree cover (Table 4 and Fig. 2). The combination of these three factors explained

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

Table 4 – Predictors of colobus use of the matrix Parametera co-efficient

Predictor variable Distance from forest Coverage of colobus foodplants Coverage of tall (>6 m) vegetation

0.67 5.37 3.11

SE

F

p

0.211 6.77 0.01 1.724 5.99 0.02 1.168 4.39 0.04

a GLM analysis of the variables influencing local sightings of C. a. palliatus within the matrix. GLM binomial proportional model (null deviance = 104.05, residual deviance = 68.19, df = 43.1).

34.5% of the variance in colobus sighting reports. Neither human activity, matrix canopy height (m), number of canopy gaps, gap length (m), proportion of bare ground cover, nor proportion of dwarf (<1 m) or low (1–3 m) vegetation cover explained any further significant variance.

4.

Discussion

Information concerning species dispersal can be difficult and time-consuming to collect. This is largely due to the rarity of dispersal events and the challenge of tracking animals through a complex landscape (Sutherland et al., 2000; Bowne and Bowers, 2004). However, we have shown that valuable qualitative and quantitative data can be gathered in a short time by drawing on local knowledge. In this case, for the arboreal primate C. a. palliatus, our findings suggest that the matrix does indeed matter. This colobus monkey uses the

1

Tall >6m vegetation Colobus food trees

Proportion cover

0.8

0.6

0.4

0.2

0 Mang

TP

Per

WSh

WGr

Shrub

AC

ShGr

Gr

Matrix type Fig. 2 – Differences in the proportion of tall (>6 m) vegetation cover and colobus food tree cover by matrix type. Ordered on x-axis in declining value of tall vegetation cover. Values on y-axis represent the mean proportional cover ±1 SE, derived from all intercept vegetation transects within each matrix type. Mang, mangrove; TP, timber plantation; Per, perennial plantation; WSh, wooded shrubland; WGr, wooded grassland; Shrub, shrubland; AC, annual cropland 1–3 m; ShGr, shrub grassland and Gr, grassland.

B I O L O G I C A L C O N S E RVAT I O N

matrix to forage and, probably, to travel between forest patches. Colobus use of each matrix type appears to depend on its structural similarity to forest (i.e. the coverage of tall >6 m vegetation), the availability of food trees, and the distance to nearest forest patch. The frequency of colobus sightings varied between matrix types and sites, which may indicate different uses of different matrix areas.

4.1.

The matrix as a facilitator of inter-patch travel

The potential for C. a. palliatus to move out of forest habitat, and hence travel to other patches, was higher than expected for an arboreal, forest-dependent colobine moving in non-forest matrix. Colobus were found up to 4.2 km from forest, and it is likely that higher travel distances occur given that only straight-line shortest distances could be explored here. Similar travel distances have also been recorded in two arboreal primate species of similar body mass (Procolobus rufomitratus, 2 km: Marsh, 1979, Alouatta palliata, 3 km: Glander, 1992). C. a. palliatus travelled through a diverse range of habitats, encompassing seven matrix types, and the coverage of tall (>6 m) vegetation was an important predictor of relative use across these habitats. Mangrove had the greatest coverage of tall vegetation (80–100%) and a comparably high incidence of colobus movement. Perennial and timber plantations also provided colobus with dense coverage (50–79%) of tall vegetation, and along with mangrove may offer the best ‘structural’ connectivity between habitat patches (Laidlaw, 2000; Tischendorf and Fahrig, 2000; Umapathy and Kumar, 2000). Nevertheless, C. a. palliatus movement is not strictly dependent upon closed canopy. Colobus were reported to move through areas of wooded shrubland, wooded grassland and shrubland. One report of colobus moving through cropland (‘annual cropland 1–3 m’), and additional observations of terrestrial locomotion, illustrate that C. a. palliatus can exhibit considerable adaptability. Similarly, howler monkeys (Alouatta spp.) have been found to travel across cornfields and grasslands in Mexico (Pozo-Montuy and Serio-Silva, 2003; Mandujano et al., 2004), whilst Tutin et al. (1997); Tutin (1999) documented black colobus (Colobus satanus) crossing savanna in Gabon. However, at all of these sites (including Kwale District) the long-distance terrestrial movement of arboreal primates is relatively uncommon. This likely reflects the possibility of higher energy expenditure and increased exposure to predation (Waser et al., 1994; Olupot and Waser, 2001), and also the scarcity of resources such as food, shelter and refuge from predators (Bennett, 1998; Baum et al., 2004). In the Kwale District, domestic dog predation and road traffic accidents are two of the documented risks to C. a. palliatus in the open matrix (Kanga, 2000; Cunneyworth and Rhys-Hurn, 2004).

4.2.

The matrix as an additional foraging habitat

Our findings indicate that C. a. palliatus exploits certain matrix types for food resources. Mangrove may be particularly important in this respect, since it contains the greatest proportion of food trees. The Zanzibar red colobus (Procolobus kirkii), and Temminck’s red colobus (Procolobus badius temminicki) also use mangrove as both foraging sites and refuges from human disturbance (Siex, 2003; Galat-Luong and Galat, 2005).

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

219

Wooded shrubland, wooded grassland and shrubland were additional indigenous matrix types where feeding, or travelling towards food trees, was observed. All of these matrix types exhibit a 20–49% coverage of colobus food trees. Our vegetation transects also uncovered differences between forest clearance methods for cultivation. Annual crop cultivation (e.g. maize, cassava and sugarcane) tends to rely on ‘clear-cutting’ (slash and burn), resulting in few or no indigenous coastal forest remnants. In contrast, in mixed perennial plantations (i.e. coconut, mango, cashew nut), remnant forest trees remain because of their shade, aesthetic and medicinal values. If forest canopy is replaced by a perennial canopy, C. a. palliatus can still access the indigenous food trees even when they are sparsely distributed. These trees may also provide food when resources are scarce in small or poor-quality forest fragments (Cowlishaw and Dunbar, 2000). Similarly, P. kirkii exploits perennial plantations for indigenous foods, and also regularly eat mango leaves and immature coconuts (Siex and Struhsaker, 1999).

4.3.

Conservation implications

Our findings suggest that the matrix is important to C. a. palliatus for both foraging and travel between forest patches. Future research that incorporates genetic methods with demographic approaches should help to further elucidate these patterns. In the meantime, our results suggest that future conservation planning in the Kwale District should emphasise the preservation or improvement of existing high-quality matrix structure between coastal forest fragments, either as viable ‘corridors’ (Beier and Noss, 1998; Bennett, 1998) or ‘stepping stones’ (Baum et al., 2004) that could structurally link isolated habitat patches. Our results indicate that it should be useful for wildlife managers to gain a greater understanding of matrix habitat in the Kwale District – to promote C. a. palliatus movement, population continuity and the management of specific habitats within the landscape – with the ultimate goal of ensuring the future persistence of this taxon in Kenya. These findings, however, should not detract from the overall importance of coastal forest preservation. C. a. palliatus uses the matrix for travelling and foraging, but these activities would not be possible without adequate forest refuges. We would therefore advocate a more integrated land-management approach (Lindenmayer and Franklin, 2002) involving, in order of priority (with land cover data taken from Anderson, 2005): 1. Preservation and enrichment of remaining coastal forest, with an additional conservation focus on the preservation of remnant mangrove areas as valuable habitat for C. a. palliatus (land cover of these habitat types in Kwale district currently 2% and 3%, respectively). 2. Preservation of indigenous matrix vegetation such as shrubland and wooded shrubland (land cover in Kwale District currently 16% and 5%, respectively). These habitats can offer a barrier to edge effects within coastal forests, increasing the effective size of such fragments (Didham and Lawton, 1999; Mesquita et al., 1999) whilst providing additional foraging habitat for C. a. palliatus.

220

B I O L O G I C A L C O N S E RVAT I O N

3. Enrichment of existing perennial and timber plantations (land cover in Kwale District currently 13% and 2%, respectively), both structurally to maintain or improve connectivity between forest patches (Taylor et al., 1993) and functionally to provide additional colobus food trees (Medellin and Equihua, 1998). The latter might include the planting of indigenous or non-indigenous species that benefit both colobus and the local community, e.g. Delonix regia and Azadrachta indica (Bicca-Marques and Calegaro-Marques, 1994; Grimes and Paterson, 2000; Ratsimbazafy, 2002). Such trees could provide foliage for C. a. palliatus without eliciting conflict over food with the local community. There are substantial conservation benefits to adopting a landscape approach to managing the effects of tropical deforestation on forest species. Our results add to a growing body of research demonstrating that forest species are capable of using matrix habitats, and that the quality of the matrix acts as a selective filter for wildlife movement between tropical forest fragments (e.g. Laurance and Yensen, 1991; Laurance, 1994; Laurance and Laurance, 1999; Gascon et al., 1999; Viveiros de Castro and Fernandez, 2004; Antongiovanni and Metzger, 2005). By integrating the matrix into forest fragmentation studies we can not only enhance our understanding of wildlife population dynamics, but also assess the vulnerability of species based on their ability to use or tolerate the matrix, and ultimately plan more effective conservation action accordingly.

Acknowledgements We thank the Government of Kenya for permitting this research (permit MOEST 13/001/31C 58); the Natural Environment Research Council, for financial support; and Wakuluzu, Friends of the Colobus Trust, Diani Beach Kenya, for logistics and research support. Professor Volker Sommer for research support, and Colobus Trust field researcher Hamisi Pakia, for assistance during the field season. R E F E R E N C E S

Aberg, J., Jansson, G., Swenson, J.E., Angelstam, P., 1995. The effect of matrix on the occurence of hazel grouse (Bonasa bonasia) in isolated habitat fragments. Oecologia 103, 265–269. Anderson, J., 2005. Habitat fragmentation and metapopulation dynamics of the Angola black-and-white colobus (Colobus angolensis palliatus) in coastal Kenya. PhD Thesis. University College London. U.K. Anderson, J., Rowcliffe, J.M., Cowlishaw, G., in press (a). The Angola black-and-white colobus (Colobus angolensis palliatus) in Kenya: historic range contraction and current conservation status. American Journal of Primatology in press. Anderson, J., Cowlishaw, G., Rowcliffe, J.M., in press (b). Effects of forest fragmentation and habitat quality on the abundance of Colobus angolensis palliatus in Kenya’s coastal forests. International Journal of Primatology in press. Andren, H., 1994. Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat-a review. Oikos 71, 355–366. Antongiovanni, M., Metzger, J.P., 2005. Influence of matrix habitats on the occurence of insectivorous bird species in Amazonian forest fragments. Biological Conservation 122, 411–451.

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

Baum, K.A., Haynes, K.J., Dillemuth, F.P., Cronin, J.T., 2004. The matrix enhances the effectiveness of corridors and stepping stones. Ecology 85, 2671–2676. Beier, P., Noss, R.F., 1998. Do habitat corridors provide connectivity? Conservation Biology 12, 1241–1252. Bennett, A.F., 1998. Movements of animals through linkages. In: Bennett, A.F. (Ed.), Linkages in the Landscape: The Role of Corridors and Connectivity in Wildlife Conservation. IUCN, Gland, Switzerland, pp. 67–91. Bernard, H.R., 2002. Research Methods in Anthropology: Qualitative and quantitative approaches. Altemira Press, Walnut Creek, CA. Bicca-Marques, J.C., Calegaro-Marques, C., 1994. Exotic plant species can serve as staple food resources for wild howler populations. Folia Primatologica 63, 209–211. Bonte, D., Lens, L., Maelfait, J.P., Hoffmann, M., Kuijken, E., 2003. Patch quality and connectivity influence spatial dynamics in a dune wolfspider. Oecologia 135, 227–233. Bowne, D.R., Bowers, M.A., 2004. Interpatch movements in spatially structured populations: a literature review. Landscape Ecology 19, 1–20. Brooks, T.M., Mittermeier, R.S., Mittermeier, C.G., da Fonseca, G.A.B., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G., Hilton-Taylor, C., 2002. Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology 16, 909–923. Bullock, J., 1996. Plants. In: Sutherland, G.D. (Ed.), Ecological Census Techniques. A handbook. Cambridge University Press, Cambridge, pp. 111–137. Burgess, N.D., Clarke, G.P., Madgwick, J., Robertson, S.A., 2000. Distribution and status. In: Burgess, N.D., Clarke, G.P. (Eds.), Coastal Forest of Eastern Africa. IUCN, Gland, Switzerland and Cambridge, UK. Chapman, C.A., Lambert, J.E., 2000. Habitat alteration and the conservation of African primates: Case study of Kibale National Park, Uganda. American Journal of Primatology 50, 169–185. Chapman, C.A., Peres, C.A., 2001. Primate conservation in the new millennium: the role of scientists. Evolutionary Anthropology 10, 16–33. Cowlishaw, G., Dunbar, R., 2000. Primate Conservation Biology. The University of Chicago Press, Chicago. Crawley, M.J., 1993. GLIM for Ecologists. Blackwell Scientific Publications, Oxford. Davies, A.G., 1994. Colobine populations. In: Davies, A.G., Oates, J.F. (Eds.), Colobine Monkeys: Their Ecology, Behaviour and Evolution. Cambridge University Press, Cambridge, pp. 285–310. Didham, R.K., Lawton, J.H., 1999. Edge structure determines the magnitude of changes in microclimate and vegetation structure in tropical forest fragments. Biotropica 31, 17–30. Fahrig, L., 2003. Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics 34, 487–515. Fahrig, L., Merriam, G., 1994. Conservation of fragmented populations. Conservation Biology 8, 50–59. 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, 585–603. Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L., Laurance, W.F., Zimmerman, B., Tocher, M., Borges, S., 1999. Matrix habitat and species richness in tropical forest remnants. Biological Conservation 91, 223–229. Glander, K.E., 1992. Dispersal patterns in Costa Rican mantled howling monkeys. International Journal of Primatology 13, 415–436.

B I O L O G I C A L C O N S E RVAT I O N

Goodwin, B.J., Fahrig, L., 2002. Effect of landscape structure on the movement behaviour of a specialized goldenrod beetle, Trirhabda borealis. Canadian Journal of Zoology-Revue Canadienne De Zoologie 80, 24–35. Grieg-Smith, P., 1983. Quantitative Plant Ecology. Blackwell Scientific Publications. Grimes, K., Paterson, J.D., 2000. Colobus guereza and exotic plant species in the Entebbe Botanical Gardens. American Journal of Primatology 51, 59–60. Grunblatt, J., Ottichilo, W.K., Sinange, R.K., 1989. A heirarchical approach to vegetation classification in Kenya. African Journal of Ecology 27, 45–51. Hanski, I., 1994. A practical model of metapopulation dynamics. Journal of Animal Ecology 63, 151–162. Harcourt, A.H., 1998. Ecological Indicators of Risk for Primates, as Judged By Species’ Susceptibility to Logging. In: Caro, T. (Ed.), Behavioural Ecology and Conservation Biology. Oxford University Press, Oxford, pp. 56–79. Harcourt, A.H., 2002. Empirical estimates of minimum viable population sizes for primates: tens to tens of thousands? Animal Conservation 5, 237–244. Kanga, E.M., 2000. Survey of the Angolan black-and-white colobus monkey (colobus angolensis palliatus) in the Diani forests, Kenya. African Primates 4, 50–54. Laidlaw, R.K., 2000. Effects of habitat disturbance and protected areas on mammals of peninsular Malaysia. Conservation Biology 14, 1639–1648. Laurance, S.G., 1994. Rainforest fragmentation and the structure of small mammal communities in tropical Queensland. Biological Conservation 69, 23–32. Laurance, S.G., Laurance, W.F., 1999. Tropical wildlife corridors: use of linear rainforest remnants by arboreal mammals. Biological Conservation 91, 231–239. Laurance, W.F., Yensen, E., 1991. Predicting the impacts of edge effects in fragmented habitats. Biological Conservation 55, 77–92. Li, Y.M., 2004. The effect of forest clear-cutting on habitat use in Sichuan snub-nosed monkey (Rhinopithecus roxellana) in Shennongjia Nature Reserve, China. Primates 45, 69–72. Lindenmayer, D.B., 1999. Future directions for biodiversity conservation in managed forests: indicator species, impact studies and monitoring programs. Forest Ecology and Management 115, 277–287. Lindenmayer, D.B., Franklin, J.F., 2002. Conserving Forest Biodiversity: A Comprehensive Multiscaled Approach. Island Press, Washington, DC. Mandujano, S., Escobedo-Morales, L.A., Palacios-Silva, R., 2004. Brief report on Alouatta palliata movements among fragments in Los Tuxtlas, Mexico. Neotropical Primates 12, 126–131. Marsh, C.W., 1979. Female transference and mate choice amongh Tana River red colobus. Nature 281, 568–569. Marsh, L.K., 2003. Primates in FragmentsEcology and Conservation. Kluwer, Academic/Plenum Publishers, New York. Maundu, P.M., Ngugi, G.W., Kabuye, C.H.S., 1999. Traditional food plants of Kenya. Kenya Resource Centre for Indigenous Knowledge (KENRIK): National Museums of Kenya, Nairobi. Medellin, R.A., Equihua, M., 1998. Mammal species richness and habitat use in rainforest and abandoned agricultural fields in Chiapas, Mexico. Journal of Applied Ecology 35, 13–23. Mesquita, R.C.G., Delamonica, P., Laurance, W.F., 1999. Effect of surrounding vegetation on edge-related tree mortality in Amazonian forest fragments. Biological Conservation 91, 129–134. Mittermeier, R.A., 1996. Introduction. In: Rowe, N. (Ed.), The Pictorial Guide to the Living Primates, vol. 1. Pogonias Press, East Hampton, New York. Mittermeier, R.A., Cheney, D.L., 1987. Conservation of primates and their habitats. In: Smuts, B.B., Cheney, D.L., Seyfarth, R.M.,

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

221

Wrangham, R.W., Struhsaker, T.T. (Eds.), Primate Societies. University of Chicago Press, Chicago, pp. 477–490. Moreno-Black, G.S., Maples, W.R., 1977. Different habitat utilization of four Cercopithecidae in a Kenyan forest. Folia Primatologica 27, 85–107. Muchoki, C.H.K., 1990. Land use in the Kwale District: Technical Report No. 136. Andere, D.K. (Ed). Department of Resource Surveys and Remote Sensing, Ministry of Planning and National Development, Nairobi, Kenya, pp. 1–20. Naughton-Treves, L., 1998. Predicting patterns of crop damage by wildlife around Kibale National Park. Conservation Biology 12, 156–168. Olupot, W., Waser, P.M., 2001. Activity patterns, habitat use and mortality risks of mangabey males living outside social groups. Animal Behaviour 61, 1227–1235. Opdam, P., 1990. Dispersal in fragmented populations: the key to survival. In: Bunce, R.G.H., Howard, D.C. (Eds.), Species Dispersal in Agricultural Habitat. Belhaven Press, London, New York, pp. 3–18. Pardini, R., 2004. Effects of forest fragmentation on small mammals in an Atlantic Forest landscape. Biodiversity and Conservation 13, 2567–2586. Pozo-Montuy, G., Serio-Silva, J.C., 2003. Locomotion and feeding on the ground by black howler monkeys (Alouatta pigra) in a very fragmented habitat of Rancheria Leona Vicario, Balancan Tabasco, Mexico. American Journal of Primatology 60, 65. Ratsimbazafy, J.H., 2002. The role of exotic plant species in the diet of ruffed lemurs (Varecia variegata) at Manombo Forest in Madagascar. American Journal of Primatology 57, 30. Reynolds, V., Wallis, J., Kyamanywa, R., 2003. Fragments, sugar, and chimpanzees in Masindi District, Western Uganda. In: Marsh, L.K. (Ed.), Primates in Fragments: Ecology and Conservation. Kluwer Academic, New York, pp. 309–319. Richmond, M.D., 1997. A Guide to the Seashores of Eastern Africa and Western Indian Ocean Islands. Sida/Department for Research Cooperation, SAREC. Ricketts, T.H., 2001. The matrix matters: Effective isolation in fragmented landscapes. American Naturalist 158, 87–99. Robertson, S.A., Luke, W.R.Q., 1993. Kenya Coastal Forests: The report of the NMK/WWF Coast Forest Survey, National Museums of Kenya, Nairobi. Siex, K.S., 2003. Effects of Population Compression on the Demography, Ecology, and Behaviour of the Zanzibar Red Colobus Monkey (Procolobus kirkii). Duke University, Durham, North Carolina. Siex, K.S., Struhsaker, T.T., 1999. Colobus monkeys and coconuts: a study of perceived human- wildlife conflicts. Journal of Applied Ecology 36, 1009–1020. Sutherland, G.D., Harestad, A.S., Price, K., Lertzman, K.P., 2000. Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology 4, 16 [online] URL: http// www.consecol.org/vol14/iss1/art16. Taylor, P.D., Fahrig, L., Henein, K., Merriam, G., 1993. Connectivity Is a vital element of landscape structure. Oikos 68, 571–573. Tischendorf, L., Fahrig, L., 2000. On the usage and measurement of landscape connectivity. Oikos 90, 7–19. Turner, M.G., 1989. Landscape ecology: the effect of pattern on process. Annual Review of Ecology and Evolutionary Systematics 20, 171–197. Tutin, C.E.G., 1999. Fragmented living: Behavioural ecology of primates in a forest fragment in the Lope Reserve, Gabon. Primates 40, 249–265. Tutin, C.E.G., White, L.J.T., Mackanga Missandzou, A., 1997. The use by rain forest mammals of natural forest fragments in an equatorial African savanna. Conservation Biology 11, 1190–1203. Umapathy, G., Kumar, A., 2000. The occurrence of arboreal mammals in the rain forest fragments in the Anamalai Hills, south India. Biological Conservation 92, 311–319.

222

B I O L O G I C A L C O N S E RVAT I O N

Van der Hoeven, C.A., de Boer, W.F., Prins, H.H.T., 2004. Pooling local expert opinions for estimating mammal densities in tropical rainforests. Journal of Nature Conservation 12, 193–204. Viveiros de Castro, E.B., Fernandez, F.A.S., 2004. Determinants of differential extinction vulnerabilities of small mammals in Atlantic forest fragments in Brazil. Biological Conservation 119, 73–80. Waser, P.M., Creel, S.R., Lucas, J.R., 1994. Death and disappearance-estimating mortality risks associated with philopatry and dispersal. Behavioural Ecology 5, 135–141.

1 3 5 ( 2 0 0 7 ) 2 1 2 –2 2 2

Wethered, R., Lawes, M.J., 2003. Matrix effects on bird assemblages in fragmented Afromontane forests in South Africa. Biological Conservation 114, 327–340. White, F., 1983. Vegetation of Africa: A Descriptive Memoir to Accompany the Unesco/AETFAT/UNSO Vegetation Map of Africa. Unesco, Paris. Wieczkowski, J., 2004. Habitat structure of a non-forest corridor used by a group of Tana mangabeys (Cercopebus galeritus). American Journal of Physical Anthropology 123, 208.