The Use by Rain Forest Mammals of Natural Forest Fragments in an Equatorial African Savanna. Utilizacion de Fragmentos de Bosque Natural por Mamiferos de Selva Lluviosa en una Sabana Ecuatorial Africana

The Use by Rain Forest Mammals of Natural Forest Fragments in an Equatorial African Savanna C. E. G. TUTIN,* L. J. T. WHITE,† AND A. MACKANGA-MISSANDZOU‡ *Centre International de Recherches Médicales de Franceville, B.P. 769, Franceville, Gabon; and Department of Biological and Molecular Biology, University of Stirling, Stirling, U.K. †Wildlife Conservation Society, New York; and Programme de Conservation et Utilisation Rationnelle des Ecosystèmes Forestiers d’Afrique Centrale (ECOFAC), Composante, Gabon (Projet FED, EU DG VIII) ‡Ministère des Eaux et Forêts, Libreville, Gabon; and Programme de Conservation et Utilisation Rationnelle des Ecosystèmes Forestiers d’Afrique Centrale (ECOFAC), Composante, Gabon (Projet FED, EU DG VIII)

Abstract: Two contrasting vegetation types occur in the Lopé Reserve in central Gabon: tropical rain forest dominates but areas of savanna containing natural forest fragments also exist. How forest mammals use the forest fragments has broad implications for conservation and management of tropical ecosystems because the natural landscape at Lopé mimics the fragmentation of forests produced increasingly by human action. The Lopé savannas result from climate induced vegetation changes over the past 20,000 years and are currently maintained by active management because, without regular burning, they are colonized by forest. Forty-five species of large mammal (body weight $2 kg) have been recorded at Lopé and, although none are savanna specialists, some use the savanna habitat. Sweep censuses were conducted monthly over 2 years in 13 forest fragments. The census sites were small (0.4–11 ha), completely or largely surrounded by savanna, and up to 450 m from continuous rain forest. Population density and biomass were calculated for the 26 species of mammals encountered (nocturnal species were rarely seen) and compared to data previously collected in adjacent continuous forests. Total biomass (6010 kg km22) was highest in the forest fragments. Compared to adjacent, continuous forest, elephants ( Loxodonta africana) were less common whereas buffalo (Syncerus caffer) and red river hog ( Potamochoerus porcus) were much more numerous. Of eight species of diurnal primates, four were more common, two occurred at similar densities, and two were much less common in the fragmented habitat. Most mammalian species moved between continuous forest and the fragments but two species of guenon and six species of duiker appeared to reside permanently in some fragments. The diversity and high biomass of large mammals found within the forest outliers at Lopé is surprising and suggests that fragmentation per se will not be catastrophic for most of these species. Utilización de Fragmentos de Bosque Natural por Mamíferos de Selva Lluviosa en una Sabana Ecuatorial Africana Resumen: En la Reserva Lopé en Gabón Central ocurren dos tipos de vegetación contrastantes: domina la selva tropical lluviosa pero también hay áreas de sabana que contienen fragmentos de bosque natural. La forma en que los mamíferos del bosque utilizan los fragmentos de bosque tiene implicaciones en la conservación y el manejo de ecosistemas tropicales porque el paisaje natural en Lopé es una replica de la fragmentación de bosques producida por acción humana. Las sabanas de Lopé son consecuencia de cambios en la vegetación inducidos por el clima en los últimos 20,000 años y en la actualidad son mantenidas con manejo activo ya que, sin quemas frecuentes, son colonizadas por bosques. En Lopé se han registrado 45 especies de mamíferos mayores (peso corporal $ kg) y, aunque ninguna es especialista de sabana, algunas utilizan el hábitat de la sabana. Durante dos años se condujeron censos de barrido mensuales en 13 fragmentos de bosque. Los sitios censados fueron pequeños (0.4–11 ha) y estaban completa o semicompletamente rodeados por sabana y a una distancia hasta de 450 m de bosque lluvioso continuo. Se calculó la densidad poblacional

Paper submitted June 20, 1996; revised manuscript accepted December 4, 1996.

1190 Conservation Biology, Pages 1190–1203 Volume 11, No. 5, October 1997

Tutin et al.

Mammalian Use of Natural Forest Fragments

1191

y la biomasa de las 26 especies de mamíferos que encontramos (raramente observamos especies nocturnas), y comparamos los datos con registros previos en bosques continuos o adyacentes. La biomasa total (6010 kg km22) fue mayor en los fragmentos de bosque. Al comparar con el bosque continuo adyacente, los elefantes ( Loxodonta africana) fueron menos comunes, mientras que los búfalos (Synceras caffer) fueron mucho más numerosos. De ocho especies de primates diurnos, cuatro fueron más comunes, dos ocurrieron en densidades similares y dos fueron mucho menos comunes en el hábitat fragmentado. La mayoría de las especies de mamíferos se movieron entre el bosque continuo y los fragmentos, pero dos especies de guenón y seis de duiquero aparentemente residían permanentemente en algunos fragmentos. La diversidad y la biomasa de mamíferos mayores encontrados en los afloramientos en Lopé son sorprendentes y sugieren que la fragmentación no será catastrófica para la mayoría de estas especies.

Introduction Forest habitats throughout the tropics are becomingly increasingly fragmented as a result of human action, and maintaining species diversity in these modified landscapes is a major preoccupation of conservation biology. In addition, in much of tropical Africa, natural savannas with gallery forests are widespread and the importance of small patches of forest, within savanna habitats, for the survival of fauna has been emphasized (e.g., Fay 1988; Mühlenberg et al. 1990). These small areas of forest are under threat in many places because they are cleared both for subsistence agriculture and for firewood. Few data exist on patterns of use by large mammals of natural forest fragments because studies of fauna able to persist in artificially fragmented forests have focused on small mammals and birds. We compare mammalian biomass in small natural forest patches embedded in a matrix of savanna with those in adjacent continuous canopy tropical rainforest in the Lopé Reserve in central Gabon. Information on the ability of “forest” mammals to move between forest outliers across open grassland and their capacity to survive without access to large areas of continuous forest is needed for conservation and management of such habitats, whether they are natural or human-made. Tropical rain forest is the predominant vegetation in present day equatorial west Africa, but extensive islands of grass savannas also occur. Some of these savanna areas, such as the Bateké Plateaux (Gabon & Congo) and the Odzala National Park (Congo), support (or supported until recently) species of large savanna mammals such as lion (Panthera leo), hyena (Crocuta crocuta), and antelope (e.g., Kobus defassa, Sylvicapra grimmia), whereas others, such as the Booué savannas, which straddle the equator between latitudes 11–128E in central Gabon, do not. However, the Booué savannas are not without animals because some species from the neighboring continuous forest use them. The Booué savannas cover about 2000 km2 and are surrounded by tropical rain forest (Aubreville 1967). The grass savannas are interspersed with narrow ribbons of gallery forest along watercourses and with scat-

tered “bosquets” (isolated forest blocks) ranging in size from 0.3 to 15 ha. The Lopé Reserve (5000 km2) includes about 800 km2 of the Booué savanna zone. The savanna landscape adds immeasurably to the beauty of the region, and the diversity of habitats was cited as a prime reason for gazetting Lopé as a protected area in 1946. Ecological and archaeological research has been in progress in the northern part of the Lopé Reserve since 1982 and while the former has concentrated on continuous forest habitats (e.g., Tutin et al. 1991, 1994; White 1994; Williamson et al. 1990), archaeological sites investigated to date are mostly in the savannas (Oslisly 1993). Recent analysis of ∂13C values and carbon dating has shown that some savannas have existed for at least 8000 years (Oslisly & White 1996). Savanna dominated vegetation was much more extensive in the region during periods of drier weather in the past (Maley 1989, 1993), but under present climatic conditions savannas are actively colonized by forest if protected from fire. There is a long history of human occupation in the area (Tutin & Oslisly 1995), and fire has played a central role in the recent history of vegetation changes at Lopé (White et al. 1996). Present management of the Reserve includes a program of savanna burning to maintain the full habitat and species diversity. However, no quantitative data on animal use of the savannas and natural forest fragments existed and, as well as being relevant to management at Lopé, such data can help conservation planning for other tropical areas where natural forests are being fragmented by human action. Information on the different propensities of forest mammals to use a naturally fragmented habitat will help predict the vulnerability of different species to habitat alteration and may allow general guidelines to be made with respect to ecological characters such as diet, body size, or activity pattern. With few exceptions the 45 species of large mammal that have been recorded at Lopé (Table 1) are species that occur in continuous tropical rain forest and no savanna specialists occur. The bushbuck (Tragelaphus scriptus) is an edge species occurring in mixed habitats, and the sitatunga (Tragelaphus spekei ) is associated with swampy vegetation (Haltenorth & Diller 1977).

Conservation Biology Volume 11, No. 5, October 1997

1192

Mammalian Use of Natural Forest Fragments

Tutin et al.

Table 1. List of large and medium-sized mammals recorded in the Lopé Reserve, Gabon. Species

Size a

Activity pattern b

Dietary category c

Primates Gorilla g. gorilla (Savage & Wyman) Pan t. troglodytes (Blumenbach) Mandrillus sphinx L Colobus satanas Waterhouse Cercocebus albigena (Gray) Cercopithecus nictitans L Cercopithecus pogonias (Bennett) Cercopithecus cephus L Perodicticus potto (P. L. S. Müller) Arctocebus calabarensis ( J. A. Smith) Galago demidoff Fischer Galago thomasi Elliot Galago alleni Waterhouse Euoticus elegantulus (Le Conte)

L M M M M S S S S S S S S S

D D D D D D D D N N N N N N

frug/fol frug frug/fol frug/fol frug frug frug frug omn insect insect insect insect sap/insect

Proboscidea Loxodonta africana cyclotis Matschie

L

D&N

fol/frug

Carnivora Panthera pardus L Profelis aurata Temminck Mellivora capensis (Schreber) Viverra civetta Schreber Nandinia binotata Gray Genetta tigrina (Schreber) Genetta servalina Pucheran Bdeogale nigripes Pucheran Atilax paludinosus (G. Cuvier) Herpestes sanguineus Rüppell Aonyx congica Lönnberg Aonyx capensis (Schinzi)

L M M M S S S S S S S S

N N N N N N N N N N N N

carn carn omn omn omn omn omn omn omn omn carn carn

Artiodactyla Hippotamus amphibius L Syncerus caffer nanus (Boddaert) Tragelaphus scriptus (Pallas) Tragelaphus spekei Sclater Potamochoerus porcus (L) Cephalophus silvicultor (Afzelius) Cephalophus dorsalis Gray Cephalophus callipygus Peters Cephalophus leucogaster Gray Cephalophus ogilbyi (Waterhouse) Cephalophus monticola (Thunberg) Neotragus batesi De Winton Hyemoschus aquaticus (Ogilby)

L L L L L L M M M M S S M

Cr D&N D&N D&N D&N D N D D D D D N

herb herb herb herb omn frug/fol frug/fol frug/fol frug/fol frug/fol frug/fol frug/fol omn

Insectivora Potamogale velox Du Chaillu

S

N

carn

Pholidota Manis gigantea Illiger Manis tricuspis Rafinesque Manis tetradactyla L

L M M

N N N

insect insect insect

Hyracoidea Dendrohyrax dorsalis (Fraser)

S

N

omn

Size (adult weight): L, .50 kg; M, 5–49 kg; S, ,5 kg. Activity pattern: D, diurnal; N, nocturnal; Cr, crepuscular; D & N, day and night. Dietary class: frug, frugivore; fol, folivore; omn, omnivorous; carn, carnivorous; insect, insectivorous; herb, herbivorous; sap, sap eater.

a b c

Conservation Biology Volume 11, No. 5, October 1997

Tutin et al.

Other species have wide geographical ranges in tropical Africa including evergreen and deciduous forest, woodland, and even habitats dominated by savanna, notably chimpanzees (Pan troglodytes), leopard (Panthera pardus), buffalo (Syncerus caffer), and elephants (Loxodonta africana), although for the latter two, forest dwelling sub-species are recognized and it is these that occur at Lopé. The present day landscape in the northern part of the Lopé Reserve resembles areas of tropical South and Central America where native rain forest has been largely cleared to create pasture for cattle, leaving isolated, usually small, fragments of forest. Similar habitats exist in tropical Australia and Africa and the impact of humanmediated forest fragmentation on resident fauna has been assessed in several studies (Estrada et al. 1993; Laurence 1991; Lovejoy et al. 1986; Newmark 1991; Saunders et al. 1991). The situation at Lopé differs in some fundamental respects from these. At Lopé fragments are relicts of ancient climate change and now active management is necessary to protect the savannas from forest encroachment, whereas in most areas referred to above, management policies focus on preserving remnants of previously continuous forest. Another difference is the large assemblage of medium and large mammals (.2 kg) present at Lopé compared to other study sites where larger mammals were rare or absent and the focus was on birds and small mammals. Systematic censuses of mammals have been undertaken in forested parts of the Lopé Reserve adjacent to, and more distant from, savannas, providing information on population density and biomass in different forest types (White 1994). Over 2 years we collected data on the abundance of these mammals in gallery forests and bosquets.

Mammalian Use of Natural Forest Fragments

1193

present day savannas, and closed canopy forest (CCF), which occurs in the south of the Lopé Reserve and in mountainous areas to the west of the savanna zone. The distributions of these major vegetation types are shown in Fig. 1. Research on vegetation dynamics and archaeology has shown that MF is a result of extensive re-colonization of savannas at Lopé that occurred between 1400 and 700 years B.P. when humans were absent from the area and thus fires were uncommon (Oslisly & White 1996; White in press; White et al. 1996). The CCF is older forest, believed to have persisted during the last arid climatic phase 2000-2500 years ago (Maley 1992). Table 2 lists the 10 species of tree that rank highest in terms of basal area in CCF, MF, gallery forests, and bosquets (GB) and in open savannas. In the savannas, grasses dominate and five species of fire-resistant shrub occur sparsely, only two of which reach diameters at breast height (dbh) $10 cm. The three forest types are botanically distinct and, as shown in Table 2, 6–8 of the top 10 tree species are exclusive and only one (Aucoumea klaineana) features in three lists. The location of the 50 km2 main study area of the Station d’Etudes des Gorilles et Chimpanzés (SEGC) is shown in Fig. 1. For the present study, 13 small forested sites in the savanna zone were chosen: 3 gallery forests (narrow strips of forest along watercourses continuing at one

Study Area The Lopé Reserve covers 5000 km2 and extends south in a tapering triangle from the river Ogooué at 08039S between latitudes 118179–118509E (Fig. 1). The climate is characterized by a long dry season of about 3 months between June and September. Temperatures vary little over the year but are lowest in the dry season: mean monthly maxima vary from 26.8–30.88C and minima from 20.5–22.38C. Mean annual rainfall is 1548 mm (1984–1995). The 45 species of large and medium-sized mammals (generally $2 kgs but including all primate species) recorded at Lopé are listed, with gross categories of activity patterns and diets, in Table 1. The Reserve includes several vegetation types described in detail by White (1992) and Tutin et al. (1994). The two major forest types are Marantaceae Forest (MF), characterized by a dense herbaceous understory, which occurs close to the

Figure 1. The Lopé Reserve showing major vegetation types and the study area of the Station d’Etudes des Gorilles et Chimpanzés (SEGC). The 13 forest fragments censused for mammals are located in the savanna to the west and north of the research station.

Conservation Biology Volume 11, No. 5, October 1997

1194

Mammalian Use of Natural Forest Fragments

Tutin et al.

Figure 2. The savanna ecotone at Lopé with recently burned savanna in the foreground, a gallery forest in the center, a bosquet behind, and continuous forest in the distance on the right. end to join the main forest); 1 “corridor” (a narrow gallery linked at both ends to forest); and 9 bosquets (small forest blocks completely surrounded by savanna). Figure 2 illustrates the forest fragments within the Lopé savanTable 2. First 10 species $10 cm dbh ranked by basal area in closed canopy forest (CCF), Marantaceae forest (MF), galleries and bosquets (GB), and savanna (SAV) habitats at Lopé. Rank Species Dacryodes buettneri Aucoumea klaineana Scyphocephalium ocochoa Coula edulis Augouardia letestui Sindoropsis le-testui Strombosiopsis tetrandra Strombosia zenkeri Santiria trimera Cylicodiscus gabonensis Cola lizae Lophira alata Pentaclethra macrophylla Pentaclethra eetveldeana Diospyros polystemon Hylodendron gabunense Xylopia quintasii Pycnanthus angolensis Diospyros dendo Eriocoelum macrocarpum Millettia sanagana Dialium lopense Uapaca guineensis Heisteria parvifolia Crossopteryx febrifuga Nauclea latifolia

Conservation Biology Volume 11, No. 5, October 1997

Family

CCF MF GB SAV

Burseraceae 1 5 Burseraceae 2 1 1 Myristicaceae 3 Olacaceae 4 Caesalpiniaceae 5 Caesalpiniaceae 6 Olacaceae 7 Olacaceae 8 Burseraceae 9 Mimosaceae 10 Sterculiaceae 2 Ochnaceae 3 Mimosaceae 4 3 Mimosaceae 6 Ebenaceae 7 5 Caesalpiniaceae 8 9 Annonaceae 9 Myristicaceae 10 Ebenaceae 2 Sapindaceae 4 Papilionaceae 6 Caesalpiniaceae 7 Euphorbiaceae 8 Olacaceae 10 Rubiaceae Rubiaceae

nas. Altitudes of the study fragments range from 220– 300 m. The area of each site was determined by tracing the outline of large-scale photographs (1/10,000) from radar images taken in 1992 onto millimeter graph paper. The total area of the 13 sites was 58.42 ha. Details of each site’s characteristics are listed in Table 3. Management of the Lopé Reserve (by the Ministère des Eaux et Forêts) includes a program of savanna burning. Fires are lit in selected savanna areas once a year, usually between 15 July and 15 August, 4–8 weeks into the dry season. Fires die out during the night, probably because of high humidity, and attempts are made to minimize damage to edge forest and reduce disturbance to animals (Tutin et al. 1996).

Methods

1 2

Censusing mammals in rain forest is not easy because visibility is often restricted either at ground level or by overlapping canopies. Counts from line transects are commonly used to sample a representative area and methods of analyzing such data have been developed to minimize error (Buckland et al. 1993; Whitesides et al. 1988). However, in the case of small areas of forest surrounded by savanna, an alternative census method exists that eliminates many of the methodological problems of line transects. Marsh and Wilson (1981) described the “sweep” census method where several observers advance in a coordinated way to detect animals by sight or by sound when they flee. This method was particularly appropriate in the present case because almost all of the animals encountered at Lopé are reluctant to flee into grassland.

Tutin et al.

Mammalian Use of Natural Forest Fragments

Table 3. Description of the 13 sweep census sites.

Name

Area (ha)

Type*

DGG Tortue W Tortue E Kourou Centroplacus Septembre Mad Buffalo Chateau d’Eau F de A Raphia Upper Lac Lower Lac Klainedoxa

11.06 1.03 0.44 6.77 3.28 1.54 4.24 1.16 5.03 7.41 4.93 2.61 8.92

gallery bosquet bosquet gallery bosquet bosquet bosquet bosquet corridor gallery bosquet bosquet bosquet

Total

58.42

Shortest distance across savanna Flowing to continuous water forest (m) Yes Yes No Yes No Yes No Yes Yes Yes No Yes Yes

0 75 35 0 230 190 450 40 0 0 60 115 90

*Gallery, narrow strip of forest along watercourse continuing at one end to main forest block; corridor, a gallery linked to continuous forest at both ends; bosquet, small forest block completely surrounded by savanna.

Monthly sweep censuses were conducted in the 13 sites from February 1993–January 1995. Censuses were conducted on three, usually consecutive, mornings between 0800 and 1330 hours. The aim was to record all large and medium sized mammals present at each site. Three observers walked slowly from one end to the other, coordinating movements by sight or by communication with walkie-talkie radios. Routes were determined by the configuration of each site. For example, in the three gallery forests the census always started at the end connected to forest to minimize the possibility of animals fleeing undetected. The largest bosquet was 250 m broad in places, so the number of observers was increased to five in order to obtain complete coverage. During sweeps inter-observer distances varied from 10–50 m depending on the configuration of the sites. For each animal seen, or heard, the following data were collected: time; species; distance from observer; height (if above the ground); number of individuals seen; and estimate of total number if all individuals were not counted. It is important to avoid recording the same individual (or group) twice and communication between observers by radio was invaluable, both to avoid duplicating sightings and to identify animals that were heard moving toward another observer. The nest building behavior of gorillas and chimpanzees provides another way to census ape populations because each weaned individual builds a sleeping nest each night (Tutin & Fernandez 1984; Tutin et al. 1995). For nests we noted species; height of nest and height and species of tree (if applicable); and age of nest. Nest groups were defined as nests of the same age within 50 m of each other. Each nest was counted only the first time it was seen, eliminating the need to use average life span of nests to calculate population densities from nest-counts.

1195

Additional data are needed to calculate population densities (average group size) and biomass (body weights) from sighting frequencies. Average group size was calculated from observations during sweep censuses if sufficient counts were made. If not, average group sizes from the adjacent forested area were used (White 1994). For all primates except chimpanzees, a distinction was made between solitary individuals and groups because only sub-adult or adult males spend time alone or occasionally in pairs. For chimpanzees, elephants, and buffalo, temporary groups of 1–201 are formed by members of communities or local populations and lone individuals can be of either gender and of all but the youngest age-classes. Thus for these species, solitary individuals were included in calculations of mean group size. To derive biomass estimates from population densities, data on body weights are required. Calculation of average body weight is problematical even if good data exist on normal weights of different age and gender classes. Here we adopt the same values as White (1994) who used the protocol developed by Oates et al. (1990) to derive the weight of an “average” individual. For solitary primates (except chimpanzees), adult male body weight was used to calculate biomass. During the sweep censuses we recorded detection distances to test whether the assumption of 100% detection of mammals was true. Ease of detection differs between species depending on size, color, and behavior. There was little risk of not detecting large species (elephant, buffalo) or ones that associate in large groups (red river hogs, mandrills), but small arboreal monkeys and duikers can be very inconspicuous, particularly when solitary. Detection distances for Cercopithecus cephus and C. nictitans (the two smallest monkeys encountered frequently); the smallest duiker (Cephalophus monticola) and medium sized, red duikers (C. callipygus, C. dorsalis, C. leucogaster, and C. ogilbyi were not always possible to distinguish in the field and were therefore grouped in the analysis). were plotted and gave effective sighting distances, calculated following Whitesides et al. (1988), of 70, 100, 15, and 30 m, respectively. As observers were usually within 50 m of each other, we can conclude that Cercopithecus monkeys and red duikers were detected consistently, but it is likely that some Cephalophus monticola were missed. Any missed animals will result in under-estimations of population density and biomass.

Results Species Recorded Twenty-six species of mammals were recorded during the 24 sweep censuses of the 13 sites. The number of times each species was encountered is shown in Table 4. In ad-

Conservation Biology Volume 11, No. 5, October 1997

1196

Mammalian Use of Natural Forest Fragments

Tutin et al.

dition to mammals, crocodiles (Osteolaemus tetrapsis) were encountered five times. All species of large or medium diurnal mammals known to occur in adjacent forests were recorded during sweep censuses, except the rare Neotragus batesi which has been seen only four times at Lopé (SEGC, unpublished data). In contrast, nocturnal and crepuscular species were seldom seen. The four medium-sized, red duikers often fled rapidly and, as shown in Table 4, less than 50% could be identified to species. Of the 506 total records, most were of six species of duiker (50%) and eight species of diurnal primates (36%). The total number of encounters per species ranged from 1–79. The five species encountered only once all are nocturnal or crepuscular. Similarly, Cephalophus dorsalis is active only at night (Feer 1988), and the five sightings were of individuals disturbed from their sleeping sites at very close range.

The total number of encounters per site varied from 2–94. The number of sightings per month (for the 13 sites combined) ranged from 13–29 with a mean of 21.3 per month. Population Density and Biomass Population densities for each species were calculated from the mean monthly number of sightings in all 13 sites (Table 4). Total biomass found in the galleries and bosquets was just over 6000 kg km22. Of this, ruminants made up 36% (dominated by buffalo which accounted for 29%), elephants 35%, red river hogs 19%, and the eight species of diurnal primates 10%. During the 24 sweep censuses we recorded 184 chimpanzee nest groups containing 424 nests and 15 gorilla nest groups containing 55 nests. Four of the gorilla nest

Table 4. Number of sightings, population densities, and biomass of large and medium-sized mammals in gallery forests and bosquets.

Species Cercopithecus cephus C. cephus, solitariesa Cercopithecus nictitans C. nictitans, solitariesa Cercopithecus pogonias C. pogonias, solitariesa Cercopithecus albigena C. albigena, solitariesa Colobus satanas C. satanas, solitariesa Mandrillus sphinx M. sphinx, solitariesa Pan t. troglodytes Gorilla g. gorilla G. gorilla, solitariesa Galago demidoff Cephalophus silvicultor Cephalophus monticola “Red” duikers, combined C. dorsalis C. callipygus C. ogilbyi C. leucogaster Unidentified Loxodonta africana cyclotis Hyemoschus aquaticus Syncerus caffer nanus Tragelaphus scriptus Tragelaphus spekei Potamochoerus porcus Nandinia binotata Panthera pardus Genetta sp? Manis sp? Dendrohyrax dorsalis All species a b

No. sightings in 24 sweeps 69 10 56 9 4 1 6 1 13 1 2 2 6 1 1 1 29 78 151 5 17 24 20 86 8 2 18 18 4 9 2 1 1 1 1 506

Lone individuals (see text for details). From adjacent continuous forest (White 1994).

Conservation Biology Volume 11, No. 5, October 1997

Density (individuals km22)

Average weight (kg)

Biomass (kg km22)

2.0 1.0 1.1 1.1

29.5 0.7 41.1 0.6 3.6 0.1 8.1 0.1 7.4 0.1 21.4 0.1 1.0 0.4 0.1 0.1 2.1 6.1 11.4

2.0 4.1 3.2 6.3 2.2 4.4 4.1 10.0 8.4 11.0 10.2 32.5 38.7 78.1 110.0 0.5 56.7 3.9 15.5

59.0 2.9 131.5 3.8 8.1 0.3 33.2 0.7 63.5 1.5 218.3 2.7 38.1 29.7 7.7 0.1 117.4 23.9 177.0

2.1 1.0 5.8 1.1 1.0 28.6 1.0 1.0 1.0 1.0 1.0

1.2 0.1 7.4 1.4 0.3 18.4 0.1 0.1 0.1 0.1 0.1

1742.0 10.4 237.5 38.4 62.8 61.9 2.1 40.3 1.6 1.9 2.2

2125.2 1.5 1752.8 52.2 18.2 1136.5 0.3 2.8 0.1 0.1 0.2

Mean group size 6.0 10.3 12.6b 18.9b 8.0 150.0b 2.3 5.3

Total biomass (kg km22) 61.9 135.3 8.4 33.9 65.0 221.0 38.1 37.4 0.1 117.4 23.9 177.0

2125.2 1.5 1752.8 52.2 18.2 1136.5 0.3 2.8 0.1 0.1 0.2 6009.3

Tutin et al.

groups contained a single nest and could thus be attributed to solitary adult males. Chimpanzee nests were found in 12 of the 13 sites and in 23 of 24 months. Gorilla nests were found in 8 of the 13 sites and in 8 of 24 months. Density calculated from nests was 1.0 km22 for chimpanzees and 0.13 km22 for gorillas. The figure for chimpanzee is in accord with that derived from sightings. However, the population density of gorillas derived from nests is lower than the figure from sighting frequency, 0.5 km22. Seasonality of Use of the Savanna Zone The sweep censuses covered two annual cycles allowing an examination of seasonal differences in the use of this habitat by the mammalian community. There was no detectable difference in the total monthly sightings between the 2 years. Annual totals were 260 versus 251, Wilcoxon matched-pairs signed rank test on sightings per month, p . 0.5. In contrast, there was a difference in the number of chimpanzee nests (and nest groups) recorded in the 2 years: 299 (120 groups) were recorded in year 1 and only 125 (68 groups) in year 2. Thus, to examine variation between calendar months, data on sightings of Cercopithecus spp. and Cephalophus spp. from the 2 years were pooled but not those on chimpanzee nests. Data on sightings of Cercopithecus nictitans and C. cephus were combined because individual sighting frequencies were too low to allow statistical testing. Similarly, data on red duikers were pooled. For neither of these two groups were there statistical differences between monthly sightings: C. nictitans and C. cephus, x2 5 9.65, df 5 11, p 5 0.5; for red duikers, x2 5 8.1, df 5 11, p 5 0.7. The number of chimpanzee nests (and nest groups) found each month varied significantly in both years of data collection: year 1 nests, x2 5 142.8, p , 0.001; year 1 nest groups, x2 = 54.1, p , 0.001; year 2 nests, x2 5 130.5, p , 0.001; year 2 nest groups, x2 5 61.9, p , 0.001 (df 5 11 in all cases). Sighting frequencies for other species were too low to be examined statistically for seasonal patterns. Inter-Site Differences There was a significant positive correlation between site area and the total number of sightings for the 13 sites (rs 5 0.934, p , 0.001) and a weaker positive correlation between site area and the total number of chimpanzee nests found (rs 5 0.545, p < 0.05). Differences might be expected between bosquets and gallery forests because the former are completely surrounded by savanna whereas the latter provide at least one forested passage to continuous forest. The four galleries had a total area of 30.3 ha compared to 28.2 for the nine bosquets (Table 3), but the total number of sightings

Mammalian Use of Natural Forest Fragments

1197

(245 and 266, respectively) did not differ significantly (x2 5 3.05, df 5 1, ns). Comparisons of the sighting frequencies of Cercopithecus cephus, C. nictitans, and Cephalophus monticola were possible, but none differed significantly in gallery forests compared to bosquets. Similarly, no differences emerged between the total sightings either of all primates or of all ruminants between the two types of site. A comparison of the number of chimpanzee nests showed that significantly more were built in bosquets than in galleries (x2 5 55.3, df 5 1, p , 0.001). Special mention should be made of the one corridor (a narrow band of forest linked at both ends to continuous forest) site. Sighting frequencies of the three species and two species-groups tested in the corridor, did not differ between the corridor, gallery forests, and bosquets but it was striking that four of the five sightings of Cercopithecus pogonias were in the corridor. Removing the corridor data from calculations of population density and biomass reduced C. pogonias from 3.6 to 0.9 individuals km22 and 8.4 to 2.2 kg km22, respectively. These comparisons suggest that gallery forests and bosquets are used in similar ways by most of the mammals but that chimpanzees nest more often in bosquets than galleries. One species of monkey, C. pogonias, was rarely encountered in gallery forest “cul de sacs” and was never seen in bosquets. The isolation of sites might influence the number of sightings of animals, so we compared the number of sightings in the nine sites within 100 m of continuous forest with the four more distant sites (115–450 m), but there was no significant difference (x2 5 0.99, df 5 1, ns). Most of the sites had streams that flowed year round, but in four of the bosquets there was no water (Table 3). Again, neither total sighting frequencies nor those of species and species-groups seen sufficiently often differed significantly between these two types of site. The proportion of chimpanzee nests found did not differ significantly between sites with water and those without (x2 5 1.4, df 5 1, ns). The five sightings of crocodiles and the six of the two mammalian species associated with water (Hyemoschus aquaticus and Tragelaphus spekei) were all in sites with permanent running water. For the other species, there was no indication that the presence or absence of permanent water influenced their use of bosquets. The areas of the 13 sites (0.4–11 ha) are all small compared to typical home ranges of the recorded mammals. Only the small territorial blue duiker (Cephalophus monticola) occupies home ranges of a size (2–5 ha) (Feer 1988) similar to all but the four smallest sites. The larger duikers and the primates have home ranges of 40– 4000 ha (Feer 1988; Gautier & Gautier-Hion 1974; Ham 1994; Tutin 1996). Individual recognition of animals, or groups, was rarely possible, but the data suggest that some duikers and groups of Cercopithecus monkeys

Conservation Biology Volume 11, No. 5, October 1997

1198

Mammalian Use of Natural Forest Fragments

Tutin et al.

Table 5. Comparison of mammalian biomass in three forested habitats in the Lopé Reserve.

Species Cercopithecus cephus Cercopithecus nictitans Cercopithecus pogonias a Cercopithecus albigena Colobus satanas Mandrillus sphinx a Pan t. troglodytes Gorilla g. gorillaa Galago demidoff a,c Primate total Red duikers, combined Cephalophus silvicultor Cephalophus monticola Hyemoschus aquaticus a Syncerus caffer nanus Tragelaphus scriptusd Tragelaphus spekei a,d Potamochoerus porcus Loxodonta africana cyclotis Ungulate total

Closed canopy forest (CCF) (site 4; White, 1994)

Marantaceae forest (MF) (sites 1 & 5; White, 1994)

19.2 86.7 9.2 40.7 92.4 66.6 19.4 23.4

10.6 71.2 12.4 43.1 107.9 11.4 21.3 66.4

357.6

344.3

85.3 34.0 4.8 1.5 213.8

51.2 54.7 1.3 0.4 95.1

204.3 1741.7 2071.6

89.8 4093.0 4385.5

Galleries & bosquets (GB) (this study) 61.9 135.3 8.4 (2.2b) 33.9 65.0 221.0 38.1 37.4 0.1 601.1 177.0 117.4 23.9 1.5 1752.8 52.2 18.2 1136.5 2125.2 5404.7

Other species

4.2

4.3

3.5

Total biomass

2433.4

4734.1

6009.3

a

Species encountered five times, or less, during sweep censuses. Corrected figure, excluding the corridor sightings (see text). c Present but not seen on censuses in CCF and MF. d Not recorded, probably not present in CCF and MF. b

were permanent, or quasi-permanent, residents of particular sites. In the largest bosquet (Klainedoxa), a group of C. nictitans were seen during 19 of 24 sweep censuses and C. cephus in 18. Red duikers were seen in 23 months in Raphia gallery and in 18 in DGG gallery. The maximum number of months at any one site with sightings of Cephalophus monticola was 13 (in Lower Lac and Mad Buffalo) but this may reflect the difficulty of detecting this small cryptic species. However, the data for the larger species (.4 kg for primates; .20 kg for ungulates) indicate that most, if not all, were transient visitors. Comparison of Mammalian Biomass in Different Habitats at Lopé Biomass estimates for Marantaceae forest (MF) and closed canopy forest (CCF) at Lopé (White 1994) were compared to the results of this study (Table 5). Total biomass in gallery forests and bosquets (GB) of the savanna zone was 1.3 times greater than that found in adjacent MF and more than twice as great as that in CCF. Of the 12 ungulate species present at Lopé, two were recorded only in GB (the bushbuck and the sitatunga). The other 10 ungulates were found in all forest types and had the same rank order of biomass in each (the four species of red duikers were considered as a group), with elephant dominating, followed by buffalo

Conservation Biology Volume 11, No. 5, October 1997

and red river hog. Proportional to total biomass, elephant represented 86.5% in MF, 71.6% in CCF but only 35.3% in GB. The eight diurnal primate species occurred in each habitat-type but the rank order of biomass differed between the three. Differences between MF and CCF were less striking than those between both of the continuous forest types and GB. Four species, Cercopithecus cephus, Mandrillus sphinx, Cercopithecus nictitans, and Pan troglodytes, were more abundant in GB than in MF and CCF, particularly the two former species that were 3-20 times more numerous. On the contrary, Colobus satanas was less common than in MF and CCF, as was Cercopithecus pogonias (especially if corridor sightings were excluded). Gorilla gorilla was less common in GB than in MF but densities in CCF and GB were similar. Cercocebus albigena densities were similar in all forest types.

Discussion Comparisons of Biomass The diverse assemblage of diurnal mammals recorded and the high figure for biomass in the GB within the savanna zone of the Lopé Reserve is surprising because the species involved are, with one exception, associated with forest habitats. The species found at much greater

Tutin et al.

densities ($double) in this habitat than in adjacent continuous forest were Cercopithecus cephus, Mandrillus sphinx, Cephalophus silvicultor, C. monticola, Syncerus caffer nanus, and Potamochoerus porcus (plus Tragelaphus scriptus and T. spekei which were not recorded in continuous forest). All but Mandrillus sphinx and Tragelaphus spekei were recorded more than five times during sweep censuses. A high risk of error is introduced in estimating population density and biomass for species that were encountered rarely during sweep censuses, particularly for those with either large group size or large body weight. Thus, density and biomass for species encountered five times or less should be treated with caution. The high figure obtained for Mandrillus sphinx (Table 4) is derived from only four encounters, two with groups (one of 15–20 individuals, the other of at least 457) and two with solitary males. Mandrills have large home ranges of unknown size (.50 km2), which they use nomadically, and visits to the savanna zone at Lopé are concentrated in the long dry season (Rogers et al. 1996). Mandrill groups were censused in August 1993 and July 1994 and the average group size of 150 used in calculations of population density, came from 20 complete counts made over 11 years in the study area. Thus, the estimate of mandrill biomass at 220 kg km22 results from rare visits by large groups, and it may be inappropriate to conclude that mandrills are 4–20 times more common in GB than in the adjacent forest. Habitat Preferences and Dependency The GBs are embedded in a larger area of savanna. The two major ecotones at Lopé, savanna (including galleries and bosquets) and forest (i.e., continuous forest), differ in botanical and structural ways. These differences have two major implications for animals: availability of food and vulnerability to predators. For terrestrial grazing herbivores, food availability is higher in savanna than in forest whereas the reverse is true for arboreal and terrestrial frugivores. Vulnerability to predation depends on size and behavior, but assessing vulnerability is difficult as few data exist. It is clear, however, that arboreal species, and those that depend on trees for predator evasion, will be more vulnerable in savanna than in forest. We attempt to distinguish between species able to live entirely within the savanna ecotone and those that depend on continuous forest but make excursions into the savannas and to assess the ecological factors responsible for these differences. Buffalo, bushbuck, and sitatunga at Lopé eat Graminae (and Cyperaceae) as well as the young leaves of shrubs and small trees. Buffalo are observed rarely in continuous forest but indirect sign shows that they penetrate deep into this habitat. Bushbuck and sitatunga have not been seen in continuous forest at Lopé but the latter

Mammalian Use of Natural Forest Fragments

1199

species does occur in swamp vegetation within continuous forest in northern Gabon. Bushbuck are the only large mammal at Lopé not able to survive in continuous forest. Buffalo and sitatunga do occur in continuous forest but the former certainly occurs at higher densities because of access to savannas, whereas density of the latter probably depends on the amount of marsh habitat available. All the large ungulates at Lopé are active day and night, but during the day those in the savanna ecotone may concentrate in GB to seek shade. For buffalo, which obtain significant amounts of food from the grass savannas, it is misleading to present population densities (and biomass values) calculated from sightings in the minority, forested, part of this ecotone. Red river hogs eat the roots of herbaceous plants including Graminae but most of their diet at Lopé comes from the fruit, seeds, and roots of trees (SEGC unpublished data). Lopé elephants also have a diverse diet including leaves, bark, and fruit of trees and, in the savannas, they feed on roots, stems, and leaves of herbs and the seasonally available fruit of shrubs (White et al. 1993). Elephants and pigs are highly mobile species although exact day range and home range sizes are not known for forest populations. Both exploit herbaceous foods found only in the savannas. Thus it is surprising that pigs were 5–12 times more abundant in the galleries and bosquets, whereas elephant density in this habitat was similar to that in CCF but almost 50% less than that of the immediately adjacent MF. Censusing species that are active both day and night only during the day introduces a bias that may work in opposite ways for these two species. Red river hog are a preferred prey of leopards at Lopé (SEGC unpublished data) and may concentrate visits to the savanna ecotone during daylight hours to avoid exposure to the largely nocturnal leopards, whereas elephants, unconcerned by predation threats, may concentrate visits to the savanna ecotone at night to avoid physiological stress. Both elephants and pigs (if red river hog and bushpig are considered members of the same species) occur in a range of habitats and have wide geographical distributions, which suggest they may not require access to continuous forest, but elephants might be reduced to lower densities if denied access to the diverse fruit resources of continuous forest (Dudley et al. 1992; White et al. 1993). All of the duiker species are frugivorous (Feer 1988), but they also browse on young leaves of shrubs. We have never observed duikers feeding on grasses or sedges although they do eat fruit of savanna shrubs. Densities of all duikers were higher in the forest fragments than in either MF or CCF. Methodological problems of censusing probably exaggerated differences because it is likely that a line of observers flush duikers more reliably during sweep censusing than a single observer during line transects. The forest fragments are critical to duikers because little or no food is obtained from grassland.

Conservation Biology Volume 11, No. 5, October 1997

1200

Mammalian Use of Natural Forest Fragments

Predation pressure may be high when these small ungulates cross open savanna, but food availability may also be higher in fragments as edge vegetation provides an abundance of accessible browse. All eight primate species at Lopé have diets dominated by fruit or seeds and obtain the majority of their foods from trees (Tutin & Fernandez 1993; Tutin et al. 1997). The three largest species travel on the ground, whereas the five smaller species are arboreal and descend to the ground rarely. All primates show extreme caution when crossing savannas. Behavior observed includes vigilance by adult males before and during crossing, and running, jumping, and bipedal locomotion. This suggests that the risk of predation while crossing open grassland is perceived as great. Potential predators on primates are leopards, which are capable of attacking apes (Boesch 1991; Hiraiwa-Hasegawa et al. 1986; Tutin & Bernischke 1991; Tutin et al. 1981), Crowned Hawk Eagles (Stephanoaetus coronatus), which prey on monkeys (Gautier-Hion & Tutin 1989; Skorupa 1988), and chimpanzees which are known to prey on three of the monkey species at Lopé (SEGC unpublished data). Golden cat (Profelis aurata) and python (Python sebae) are also potential predators of the smaller primates. All primates are more vulnerable to predators when trees are sparse or absent. Most primate species make sporadic visits to forest fragments from the surrounding continuous forests. Chimpanzees are frequent visitors and density estimates in fragments from both sightings and nests were similar to those in continuous forest, but their use of the savanna ecotone varies significantly between seasons and years. Gorillas use the savanna zone less, particularly for nesting. Use of the galleries and bosquets by both species of apes is related to fruit availability and the scarcity of gorilla nests is explained by low densities of preferred herbaceous nesting materials in this habitat (Tutin et al. 1995). The groups of Cercopithecus nictitans and C. cephus that are resident in some larger bosquets occupy home ranges 8-25 times smaller than in continuous forest. C. cephus forage low in the forest canopy (Gautier-Hion 1980) and may be gap and edge specialists. This could explain their high density in galleries and bosquets where edge to area ratio is high. C. nictitans has a wide geographical distribution (Oates 1988) and extends into savanna dominated habitats in the north of Central African Republic (Fay 1988), suggesting it is an adaptable species. The contrast between Cercopithecus pogonias, which visits forest fragments (excluding the corridor site) infrequently, and the other two species is striking because body weights and diets are similar for these three guenons (Gautier-Hion 1980; Tutin et al. 1997). The behavior of Cercopithecus pogonias, which Fay (1988) also found to occur at very low densities in forest fragments, is a puzzle but may be linked to vulnerability to predators or competition with the other guenons for shared foods. Crossing open grass savanna to reach

Conservation Biology Volume 11, No. 5, October 1997

Tutin et al.

bosquets involves costs, in terms of locomotion, increased vulnerability to predators, and probably physiological stress. The benefits are difficult to identify but could include relaxation of either inter- or intra-specific competition for food. Behavioral factors may also be important: C. pogonias is a cryptic and vigilant monkey that spends the majority of time in polyspecific association with either Cercocebus albigena or Colobus satanas (Ham 1994; SEGC unpublished data). Although there is no obvious reason why C. pogonias would be more vulnerable to predation than the slightly smaller C. cephus or the slightly larger C. nictitans, their behavior suggests that they feel more vulnerable. Mobility, or dietary flexibility, is essential for primates resident in forest fragments because food resources can be dramatically reduced by sporadic visits of larger species. Mandrill groups, which can number up to 700 individuals, move large distances but can occupy a bosquet for periods of 1–3 days during the dry season at Lopé (Rogers et al. 1996). In 1994 a group of 457 mandrills (weight of average individual, 11 kg) spent 3 days in an 8-ha bosquet, which, in terms of relative body weight, represents 1256 days of occupancy by the resident group of six Cercopithecus cephus (average weight, 2 kg). Conservation Implications Throughout the tropics, naturally continuous forest vegetation is being increasingly fragmented by human action, and considerable research has been conducted on the impact of this change on fauna (e.g., Estrada et al. 1993; Lovejoy et al. 1986; Malcolm 1994) and on the implications for management (e.g., Saunders et al. 1991). The data from the naturally fragmented ecotone at Lopé show that most large and medium sized forest mammals use the savanna dominated habitat but few live permanently within a single forest fragment. Many depend on trees for their food and would need a large area of fragmented habitat to provide their dietary needs if continuous forest was not available. Body size and dietary correlates with propensity to use the savanna ecotone emerged in expected ways with small species (with small home ranges) being the only supposed permanent residents and species that feed on herbaceous vegetation of savannas being generally common. However, the behavior of one primate species, Cercopithecus pogonias, which used the savanna ecotone so rarely in contrast to two closely related species of similar size, was unexpected and remains unexplained. This warns against generalization as do the results of a survey of forest patches close to the Kibale Forest in Uganda. Onderdonk (1996) found some similarities between Kibale and Lopé; Cercopithecus ascanius (which is equivalent to C. cephus) and chimpanzees used fragments regularly. However, some differences emerged: Colobus guereza were frequent residents in small patches but Cercocebus

Tutin et al.

albigena and Cercopithecus mitis (equivalent to C. nictitans) were never found in fragments. The presence of savannas adds to the diversity of both plant and animal species protected within the Lopé Reserve. A proportion of plant (Tutin et al. 1994; White & Abernethy 1996) and bird (Christy & Clarke 1994) species occur only in the savanna ecotone and one species of large mammal (Tragelaphus scriptus) appears to depend on savannas. Under present climatic conditions, forest is actively colonizing the savannas unless this succession is arrested by regular burning. Thus, active management is necessary to preserve the totality of present biodiversity. On the other hand, protecting certain savannas from fire allows study of natural succession and has revealed, for example, the role of frugivorous mammals as seed dispersers in this process (L. J. T. White, unpublished data). The savanna ecotone is crucial for the development and long-term viability of Lopé as a protected area because it adds immeasurably to the site’s attraction to tourists. Data on seasonal patterns of use by large mammals are now used by tourist guides to maximize visitors’ chances of seeing particular species in the ideal viewing conditions of the savanna ecotone. All these factors must be weighed in developing management plans for a habitat that is far from static and has emerged from a long history of both climate induced changes in vegetation and interaction with humans. Management recommendations for fragmented habitats often stress the importance of maintaining, or creating, corridors of woody vegetation between blocks of forest (see references in Saunders & Hobbs 1991). Our data suggest this would be important for Cercopithecus pogonias but that the other forest mammals at Lopé are willing to travel several hundred meters across open grassland to visit forest fragments. However, our study was limited to relatively large species and the degree of isolation of the studied fragments was comparatively small. Further research is needed on small mammals, birds, and insects—groups for which open grassland is more likely to constitute a barrier to movement. The Lopé Reserve is well protected from poaching so predation pressure comes from nonhuman species and is thus limited. This is perhaps the starkest contrast between natural and human-mediated forest fragmentation. In most of the latter cases, large and medium sized mammals have been exterminated by hunting prior to forest fragmentation (e.g., Redford & Robinson 1991; Wilkie & Finn 1990; Wilkie et al., in press). The absence of mammalian seed dispersers poses a threat to the longer-term maintenance of plant species diversity in both fragmented and nonfragmented forest habitats (Alexandre 1980; Terborgh 1992). Some primate species have proven remarkably resistant to habitat fragmentation with groups surviving in forest patches in West Africa that have been isolated by human activities. At Bossou, in Guinea, chimpanzees

Mammalian Use of Natural Forest Fragments

1201

persist in a sacred forest grove surrounded by fields (Sugiyama 1989), and in Nigeria several populations of the severely endangered Sclater’s guenon (Cercopithecus sclateri) remain in small sacred forests (Oates et al. 1992). The recent evolutionary history of these primates has seen forest cover in west and central Africa much reduced on several occasions due to climate change in the past 20,000 years (Maley 1993), so preadapatations to fragmented habitats are likely for some species. The speed at which fragmentation occurs has been underlined as critical in terms of short-term threats to species diversity (Kellman 1996). Past reductions in forest cover caused by climate change happened slowly and the few studies of naturally fragmented tropical habitats have found surprisingly high levels of both plant (Kellman & Tackaberry 1992) and large mammal (this study) diversity. Human mediated fragmentation of tropical forests occurs rapidly and large and medium sized mammals are often exterminated by hunting. Some species may persist, often because of cultural taboos on hunting, but the long-term survival of small, isolated populations will depend on active management such as establishing linking corridors and protection from hunting. If conflicts between humans and large mammals (hunting and crop raiding) can be resolved, our data show that reintroduction of native primates and ungulates could also be envisaged as a management option. The presence of these animals will bring benefits both through dispersal of seeds and by providing possibilities for tourism and recreation that could benefit local communities. However, the choice of species would require careful study as illustrated by the contrasting patterns of use of forest fragments by primate species both within and between sites.

Acknowledgments We gratefully acknowledge core funding from the Centre International de Recherches Médicales de Franceville, Gabon and from Wildlife Conservation Society, New York. Part of this work was supported by the ECOFAC program funded by the European Union (DG VIII) and managed by AGRECO-GEIE. We extend special thanks to M. Fernandez for his invaluable assistance. We thank the following colleagues for contributing to data collection: R. J. Parnell, J. Dibakou, A. Sinaboure, D. Mala, K. A. Abernethy, P. Peignot, B. C. Voysey, K. E. McDonald, P. Christy, K. A. Abernethy, J. Maley, R. Oslisly, M. Fernandez, and L. Molloy. Three anonymous reviewers made constructive comments on the manuscript. Literature Cited Alexandre, D. Y. 1980. Charactère saisonnier de la fructification dans une forêt hygrophile de Cote d’Ivoire. Révue Ecologie 34:335–350.

Conservation Biology Volume 11, No. 5, October 1997

1202

Mammalian Use of Natural Forest Fragments

Aubreville, A. 1967. Les étranges mosaiques forêt-savane du sommet de la boucle de l’Ogooué au Gabon. Adansonia Série 2 7:13–22. Boesch, C. 1991. The effects of leopard predation on grouping patterns in forest chimpanzees. Behaviour 117:220–242. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling. Chapman & Hall, London. Christy, P., and W. Clarke. 1994. Guide des oiseaux de la Réserve de la Lopé. ECOFAC, Libreville, Gabon. Dudley, J. P., A. Y. Mensah-Ntiamoah, and D. G. Kpelle. 1992. Forest elephants in a rainforest fragment: preliminary findings from a wildlife conservation project in southern Ghana. African Journal of Ecology 30:116–126. Estrada, A., R. Coates-Estrada, J. Meritt D., S. Montiel, and D. Curiel. 1993. Patterns of frugivore species richness and abundance in forest islands and in agricultural habitats at Los Tuxtlas, Mexico. Pages 245–257 in T. H. Fleming and A. Estrada, editors. Frugivory and seed dispersal: ecological and evolutionary aspects. Kluwer Academic Publishers, Dordrecht. Fay, J. M. 1988. Forest monkey populations in the Central African Republic: the northern limits. A census in Manovo-Gounda-St. Floris National Park. Mammalia 52:57–74. Feer, F. 1988. Stratégies écologiques de deux espèces de Bovides sympatriques de la forêt sempervirente africaine (Cephalphus callipygus et C. dorsalis): influence du rhythme d’activité. Thèse de Doctorat d’Etat ès Sciences. Université de Pierre et Marie Curie, Paris. Gautier-Hion, A. 1980. Seasonal variation of diet related to species and sex in a community of Cercopithecus monkeys. Journal of Animal Ecology 49:237–269. Gautier-Hion, A., and J. P. Gautier. 1974. Les associations polyspécifiques chez les Cercopithèques du plateau de M’Passa, Gabon. Folia primatologia 22:134–177. Gautier-Hion, A., and C. E. G. Tutin. 1989. Mutual attack by a polyspecific association of monkeys against a crowned hawk eagle. Folia primatologia 51:149–151. Haltenorth, T., and H. Diller. 1977. A field guide to the mammals of Africa including Madagascar. Collins, London. Ham, R. M. 1994. Behavioural ecology of grey-cheeked mangabeys (Cercocebus albigena) in the Lopé Reserve, Gabon. Ph.D. thesis. University of Stirling, Stirling, United Kingdom. Hiraiwa-Hasegawa, M., R. W. Byrne, H. Takasaki, and J. M. E. Byrne. 1986. Aggression towards large carnivores by wild chimpanzees of Mahale Mountains National Park, Tanzania. Folia primatologia 47:8–13. Kellman, M. 1996. Redefining roles: plant community reorganisation and species preservation in fragmented systems. Global Ecology and Biogeography Letters 5:111–116. Kellman, M., and R. Tackaberry. 1992. Disturbance and tree species coexistence in tropical riparian forest fragments. Global Ecology and Biogeography Letters 3:1–9. Laurence, W. F. 1991. Ecological correlates of extinction proneness in Australian tropical rain forest mammals. Conservation Biology 5: 79–89. Lovejoy, T. E., et al. 1986. Edge and other effects of isolation on Amazon forest fragments. Pages 257–285 in M. E. Soulé, editor. Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, Massachusetts. Malcolm, J. R. 1994. Edge effects in central Amazonian forest fragments. Ecology 75:2438–2445. Maley, J. 1989. Late Quaternary climatic changes in African rain forest: Forest refugia and the major role of sea surface temperature variations. Pages 585–616 in M. Leinen and M. Sarnthein, editors. Paleoclimatology and paleometeorology: modern and past patterns of global atmospheric transport. Kluwer Academic Publishers, Dordrecht. Maley, J. 1992. Mise en évidence d’une péjoration climatique entre ca. 2500 et 2000 ans BP en Afrique tropicale humide. Bulletin de la Societé Géologique de France 163:363–365. Maley, J. 1993. The climatic and vegetational history of equatorial re-

Conservation Biology Volume 11, No. 5, October 1997

Tutin et al.

gions of Africa during the Upper Quaternary. Pages 43–52 in T. Shaw, P. Sinclair, B. Andah, and A. Okpoko, editors. The archaeology of Africa: foods, metals and towns. Routledge, London. Marsh, C. W., and W. L. Wilson. 1981. A survey of primates in peninsular Malaysian forests. University Kebangsaan, Malaysia. Mühlenberg, M., A. Galat-Luong, P. Poilecot, B. Steinhauer-Burkart, and I. Kühn. 1990. L’importance des ilôts forestiers de savane humide pour la conservation de la faune de fôret dense en Côte d’Ivoire. Revue Ecologique 45:197–214. Newmark, W. D. 1991. Tropical forest fragmentation and the local extinction of understorey birds in the eastern Usambara Mountains, Tanzania. Conservation Biology 5:67–78. Oates, J. F. 1988. The distribution of Cercopithecus monkeys in West African forests. Pages 79-103 in A. Gautier-Hion, F. Bourlière, J. P. Gautier, and J. Kingdon, editors. A primate radiation. Evolutionary biology of the African guenons. Cambridge University Press, Cambridge. Oates, J. F., P. A. Anadu, E. L. Gadsby, and J. L. Werre. 1992. Sclater’s guenon. National Geographic Research and Exploration 8:476–491. Oates, J. F., G. H. Whitesides, A. G. Davies, P. G. Waterman, S. M. Green, G. L. Dasilva, and S. Mole. 1990. Determinants of variation in tropical forest primate biomass: new evidence from West Africa. Ecology 71:328–343. Onderdonk, D. A. 1996. Effects of forest fragmentation on the primate community of Kibale National Park, Uganda. Paper presented at the 16th IPS/19th ASP Congress, Madison, Wisconsin. Oslisly, R. 1993. Préhistoire de la moyenne vallée de l’Ogooué (Gabon). 96. Editions de l’ORSTOM, Paris. Oslisly, R., and L. J. T. White. 1996. La rélation homme/milieu dans la Réserve de la Lopé (Gabon) au cours de l’Holocene: les implications sur l’environnement. Pages in SYMPOSIUM ECOFIT. Dynamique à long terme des ecosystèmes forestiers intertropicaux. ORSTOM, Paris. Redford, K. H., and J. G. Robinson. 1991. Subsistence and commercial uses of wildlife in Latin America. Pages 6–23 in J. G. Robinson and K. H. Redford, editors. Neotropical wildlife use and conservation. Chicago University Press, Chicago. Rogers, M. E., K. A. Abernethy, B. Fontaine, E. J. Wickings, L. J. T. White, and C. E. G. Tutin. 1996. Ten days in the life of a mandrill horde in the Lopé Reserve, Gabon. American Journal of Primatology 40:297–313. Saunders, D. A., and R. J. Hobbs. 1991. Nature conservation: the role of corridors. Surrey Beatty and Sons, Chipping Norton, Australia. Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18–32. Skorupa, J. P. 1988. Crowned eagles Stephanoaetus coronatus in rainforest: observations on breeding chronology and diet at a nest in Uganda. Ibis 131:294–298. Sugiyama, Y. 1989. Population dynamics of chimpanzees at Bossou, Guinea. Pages 134–145 in P. G. Heltne and L. Marquardt, editors. Understanding chimpanzees. Harvard University Press, Cambridge, Massachusetts. Terborgh, J. 1992. Maintenance of diversity in tropical forests. Biotropica 24:283–292. Tutin, C. E. G. 1996. Ranging and social structure of lowland gorillas in the Lopé Reserve, Gabon. Pages 58-70 in W. C. McGrew, L. F. Marchant, and T. Nishida, editors. Great ape societies. Cambridge University Press, Cambridge, United Kingdom. Tutin, C. E. G., and K. Bernischke. 1991. Possible osteomyelitis of skull causes death of a wild lowland gorilla in the Lopé Reserve, Gabon. Journal of Medical Primatology 20:357–360. Tutin, C. E. G., and M. Fernandez. 1993. Composition of the diet of chimpanzees and comparisons with that of sympatric lowland gorillas in the Lopé Reserve, Gabon. American Journal of Primatology 30:195–211. Tutin, C. E. G., M. Fernandez, M. E. Rogers, E. A. Williamson, and W. C.

Tutin et al.

McGrew. 1991. Foraging profiles of sympatric lowland gorillas and chimpanzees in the Lopé Reserve, Gabon. Philosophical Transactions of the Royal Society of London Series B 334:179–186. Tutin, C. E. G., R. M. Ham, L. J. T. White, and M. J. S. Harrison. 1997. The primate community of the Lopé Reserve, Gabon: diets, responses to fruit scarcity and effects on biomass. American Journal of Primatology 42:1–24. Tutin, C. E. G., W. C. McGrew, and P. J. Baldwin. 1981. Responses of wild chimpanzees to potential predators. Pages 136–141 in A. B. Chiarelli and R. S. Corruccini, editors. Primate behavior and sociobiology. Springer Verlag, Berlin. Tutin, C. E. G., and R. Oslisly. 1995. Homo, Pan and Gorilla: co-existence over 60,000 years at Lopé in central Gabon. Journal of Human Evolution 28:597–602. Tutin, C. E. G., L. J. T. White, and A. Mackanga-Missandzou. 1996. Lightning strike burns large forest tree in the Lopé Reserve, Gabon. Global Ecology and Biogeography Letters 5:36–41. Tutin, C. E. G., L. J. T. White, E. A. Williamson, M. Fernandez, and G. McPherson. 1994. List of plant species identified in the northern part of the Lopé Reserve, Gabon. Tropics 3:249–276. White, L. J. T. 1992. Vegetation history and logging disturbance: effects on rain forest mammals in the Lopé Reserve, Gabon. Ph.D. thesis. University of Edinburgh, Edinburgh. White, L. J. T. 1994. Biomass of rain forest mammals in the Lopé Reserve, Gabon. Journal of Animal Ecology 63:499–512. White, L. J. T. In press. Forest-savanna dynamics and the origins of ‘Marantaceae Forest’ in the Lopé Reserve, Gabon. In B. Weber, L. J. T.

Mammalian Use of Natural Forest Fragments

1203

White, A. Vedder, and H. Simons-Morland, editors. African rain forest ecology and conservation. Yale University Press, New Haven, Connecticut. White, L. J. T., and K. A. Abernethy. 1996. Guide de la végétation de la Réserve de la Lopé, Gabon. ECOFAC, Libreville, Gabon. White, L. J. T., K. Abernethy, R. Oslisly, and J. Maley. 1996. L’Okoumé (Aucoumea klaineana): expansion et déclin d’un arbre pionnier en Afrique centrale atlantique au cours de l’Holocene. Pages 195– 198 in SYMPOSIUM ECOFIT. Dynamique à long terme des ecosysèmes forestiers intertropicaux. ORSTOM, Paris. White, L. J. T., C. E. G. Tutin, and M. Fernandez. 1993. Group composition and diet of forest elephants, Loxodonta africana cyclotis, Matschie 1900, in the Lopé Reserve, Gabon. African Journal of Ecology 31:181–199. Whitesides, G. H., J. F. Oates, S. M. Green, and R. P. Kluberdanz. 1988. Estimating primate densities from transects in a West African rain forest: a comparison of techniques. Journal of Animal Ecology 57: 345–367. Wilkie, D. S., B. Curran, R. Tshombe, and G. A. Morelli. In press. Sustainability of subsistence farming and hunting in the Ituri forest of northeastern Zaïre. Conservation Biology. Wilkie, D. S., and J. T. Finn. 1990. Slash-burn cultivation and mammal abundance in the Ituri forest, Zaïre. Biotropica 22:90–99. Williamson, E. A., C. E. G. Tutin, M. E. Rogers, and M. Fernandez. 1990. Composition of the diet of lowland gorillas at Lopé in Gabon. American Journal of Primatology 21:265–277.

Conservation Biology Volume 11, No. 5, October 1997