Belg. J. Zool., 135 (supplement) : 175-177
December 2005
Density and cover preferences of Black-and-rufous
elephant-shrews (Rhynchocyon petersi) in Chome Forest
Reserve, Tanzania

Stephanie Coster and David O. Ribble
Department of Biology, Trinity University, One Trinity Place, San Antonio, TX 78212
Corresponding author : David O. Ribble, e-mail address :
ABSTRACT. The objective of this study was to determine the density and habitat preference of the Black-and-
rufous elephant-shrew (Rhynchocyon petersi) in Chome Forest Reserve, Tanzania. Chome Forest (143km2) is
located in the South Pare Mountains and provides critical habitat for endangered R. petersi. Twelve 300m transects
were cut through the centre of the forest in an east-west direction and the number of elephant-shrew nests within 2.5
meters on each side of the transects was recorded. The mean number of nests per 100m transect (0.39 0.47 [1SE])
translated to a density estimate of 19 elephant-shrews per km2 (SE=23). Nest sites tended to be found in areas with
greater than expected cover at the low (<5m) levels. These results indicate the population density of R. petersi is
lower in the Chome Forest Reserve than in most populations in the Eastern Arc Mountains. The reasons for this dif-
ference and the conservation implications are discussed.
KEY WORDS : elephant-shrew, sengi, conservation, density.
ries, their presence is an indication of a healthy forest eco-
Africa's tropical forests are home to large diversity of
Of the giant elephant-shrews, the most is known about
species, many of which are endemic to the African conti-
the Golden-rumped elephant-shrew (R. chrysopygus) and
nent. With increases in both human population and defor-
there are few records about R. petersi. The objective of
estation, more and more animals are becoming threatened
this study was to estimate the density of elephant-shrews
(MYERS, 1988). The elephant-shrews or sengis (order
based on nest counts and analyze habitat preferences in an
Macroscelidea) are one such group. There are 15 species
undisturbed forest reserve where R. petersi were known
in this strictly African mammal group, three of which are
to occur (STANLEY et al., 1996).
referred to as "giant" elephant-shrews and are of the
genus Rhynchocyon (Peters 1847). All three Rhynchocyon
species are considered threatened due to habitat destruc-
tion and fragmentation, including the species that is the
focus of this study, the Black-and-rufous elephant-shrew
Chome Forest Reserve is located within the Eastern
(R. petersi, Bocage 1880) (NICOLL & RATHBUN, 1990).
Arc Mountains (37o 58' 0.12" E, 4o 17' 60" S), a range on
the southeast coast of Kenya and the eastern coast of Tan-
The giant elephant-shrews share similar life histories in
zania. Chome Forest is made up of mostly wet montane
that they are diurnal insectivores that live in lowland and
forests (submontane, montane, and upper montane) with
montane forests and dense woodlands (RATHBUN, 1984).
elfin forest on high ridges and heathlands on rocky, acidic
They can be found in altitudes ranging from sea level-
soils. It covers approximately 142.8km2 and is situated on
2300m. While foraging they use their long proboscis to
the ridges and plateau of the South Pare Mountains in the
turn over leaf litter and dig up beetles, termites, other
district of Same in Kilimanjaro Region, Tanzania. The
insects and centipedes. Once the arthropods are exposed,
reserve was established in 1951 under the National Forest
the sengi's long tongue extends and scoops them up
Policy and Draft Act to ensure ecosystem stability
(KINGDON, 1997).
through conservation of forest biodiversity, water catch-
For shelter, the giant elephant-shrews build nests. The
ment and soil fertility. Because of the high annual rainfall
dimensions of their shelters are typically one meter wide
and pristine forest cover, the forest has a high water
with a body-sized bowl of 20cm long, 15cm wide and
catchment value and is an important resource for the 22
10cm deep (RATHBUN, 1979). Giant elephant-shrews live
surrounding villages in the catchment area. The altitude in
in monogamous pairs with defined territories and there-
the reserve ranges from 1250-2400m. The estimated
fore each animal can make and maintain up to ten nests in
annual rainfall is 1400mm. During the dry season fire is a
one territory with several nests in use at one time (FITZ-
problem because it replaces dry and lower forests with
GIBBON & RATHBUN, 1994). Their territories are typically
heath land. Historically fire was not a threat but fires have
about 1-1.7 hectare (RATHBUN, 1979). Because of their
increased with human activity near the forest. STANLEY et
dependence on undisturbed forest and their large territo-
al. (1996) noted the presence of R. petersi in the Chome

Stephanie Coster and David O. Ribble
Forest, but no study has documented the densities of these
layer (< 5m) elephant-shrews selected higher shade
elephant-shrews in this forest.
classes for nests than expected (2 = 8.14, d.f.= 2, P <
This study was conducted in the rainy season between
0.05; Fig. 1). In particular the shade class of 36-55% was
April 11th and 29th 2001. Chome Forest is accessible by
higher than expected (Fig. 1).
road, but the forest itself is navigated only by footpaths.
One such trail that is heavily trafficked bisects the middle
of the forest from a west to east direction and is used by
locals trekking from Mhero to Kanza or Mhero to Bombo.
The estimated population of R. petersi in Chome Forest
Because the path was established and because there had
was 19 (SE=23) elephant-shrews per km2. This estimate
been sengi sightings in the area, this trail was chosen to be
is most interesting when compared to density estimates of
a reference path for transects. In order to estimate ele-
R. petersi from other forest reserves in the Eastern Arc
phant-shrew density and habitat preference, transects
Mountains. HANNA & ANDERSON (1994) estimated popu-
were cut through the forest starting from the forest edge
lation density of R. petersi for seven study sites in the
on the western side (near Mhero boundary). Twelve
Eastern Arc Mountain range using similar techniques to
transects were cut perpendicular from the path, each
this study (Table 1). The population density of Chome
300m long and paced 500m apart. The first transect was
Forest Reserve is low when compared to these other sites
500m from the forest edge.
where R. petersi was found. According to HANNA &
Nest frequencies within 2.5 meters on each side of the
ANDERSON (1994), the available habitat for R. petersi
transect were tallied. Both newer (in use) and older nests
tends to be fragmented given its location at higher eleva-
were recorded. For more qualitative data, the number of
tions, which is typically on isolated mountains. Current
scraping/digging sites was also tallied. Using a density
logging and hunting pressures on these forests have fur-
conversion factor from FITZGIBBON & RATHBUN (1994)
ther exacerbated the lack of habitable areas for R. petersi.
study on R. chrysopygus, the population density per km2
Though Chome Forest is closed to timber harvesting and
was estimated.
hunting, pit saws and traps were sighted in the forest and
To examine if giant elephant-shrews were selecting
therefore human activity in the forest could be limiting
specific shade classes for their nests, percent canopy
the numbers of R. petersi. Also because of the proximity
cover was estimated every 20m along each transect at
of the village and the lack of a buffer zone, the forest is
each of three layers of canopy : < 5m, 5-15m, and > 15m.
isolated which could prevent immigration into the exist-
Percent canopy cover was divided into four shade classes,
ing population. FITZGIBBON (1994) suggested that for R.
of 0-15%, 16-35%, 36-55% and >56%. Canopy cover was
chrysopygus in Kenya, selective tree felling and pole cut-
also recorded at each observed nest site and compared to
ting in protected areas have little effect on elephant-shrew
available cover with 2 analysis.
densities but she warns that in unprotected areas, human
pressure may be more of a threat.
The average number of nests found per 100m of
Estimated population densities of Rhynchocyon petersi in forest
transect was 0.39 (SE= 0.47). Using the density conver-
sites throughout Eastern Tanzania (from HANNAH & ANDERSON,
sion factor from the FITZGIBBON & RATHBUN (1994)
study, the estimated density was 19 per km2 (SE=23). By
extrapolation a liberal estimate for the whole reserve
Population density in pristine
would be approximately 2700 R. petersi.
Forest site areas
areas of forest reserve (No./

Habitat Available
Nest Sites

e o
> 0 *
> 0 *
Chome This study

* Not enough animals were captured for a density estimate, although
a few animals were observed.
Rhynchocyon petersi chose nesting sites in areas of
Shade Class (%)
greater canopy cover than expected (Fig. 1), probably to
Fig. 1. Shade classes for nest sites of Rhynchocyon petersi
avoid predators and to find sufficient leaf litter to con-
compared to the available habitat data collected every 20m.
struct nests. Nests were observed frequently at the base of
trees, and typically wild coffee was the predominant
shrub of the understory. Scraping and digging sites were
There was no evidence that elephant-shrews were
often found near coarse woody debris perhaps due to the
selecting specific shade classes in the middle or upper
higher proportion of prey found living in this substrate.
canopy layers (P > 0.10). However, in the lowest canopy
Finally, the forest edge seemed to have more nests (aver-

Status of Black-and-rufous elephant-shrew
age of 2.7 nests/100m) than the other areas of the forest
further help from Jo Anderson. We would like to thank the
(0.3 nests/ 100m), indicating the Black-and-rufous ele-
authorities at Chome Forest Reserve for granting access. The
phant-shrew may perhaps forage in both the forest as well
Department of Biology at Trinity supported our travel to the 9th
as in the surrounding heathland.
International African Small Mammal Symposium, Sokaine Uni-
versity of Agriculture, Tanzania, where we first presented these
The population density conversion factor we used was
results and we are grateful for that support. We would lastly like
determined from data collected on R. chrysopygus (FITZ-
to thank Galen Rathbun for his continued help and encourage-
ment in our studies of sengis.
argued that R. petersi exist at lower densities than R.
chrysopygus. If this is the case, then the conversion esti-
mator results in a liberal density estimate for R. petersi.
When extrapolating the population estimate for the entire
forest, we assumed that the whole forest area provided
FITZGIBBON, C.D. (1994). The distribution and abundance of the
golden-rumped elephant- shrew Rhynchocyon chrysopygus
adequate habitat for R. petersi, but there is evidence of
in Kenyan coastal forests. Biological Conservation, 67 :
disturbed areas of the forest which include burnings and
heathland habitat which has not been known to support
FITZGIBBON, C.D. & G.B. RATHBUN (1994). Surveying Rhyn-
elephant-shrew nests. All of these surveys of R. petersi
chocyon elephant-shrews in tropical forest. African Journal
were conducted over short periods of time and ideally
of Ecology, 32 : 50-57.
longer studies should be conducted. However, in general
KINGDON, J. (1997). The Kingdon Field Guide to African Mam-
Rhynchocyon populations do not vary substantially over
mals. Academic Press. San Diego, CA : 142-152.
time (RATHBUN, 1979) so we are confident in our relative
HANNA, N. & J. ANDERSON (1994). Njule 92, Final Report,
comparisons between forests.
assessing the status and distribution of the black-and-rufous
The results of this study indicate that the population of
elephant-shrew. Unpublished report. Oxford University
expedition to Tanzania. Oxford, England : 67pp.
R. petersi in Chome Forest Reserve is low and isolated
MYERS, N. (1988). Tropical Forests and their Species : Going,
when compared to other populations in the Eastern Arc
going, ...? In : WILSON (ed), Biodiversity. National Academy
Mountains and thus long term conservation plans must
Press. Washington DC : 28-35.
safeguard the future of the forest. This study also showed
NICOLL, M. & G. RATHBUN (1990). African Insectivora and Ele-
that forest cover is essential to the elephant-shrew, pre-
phant-shrews : an action plan for their conservation. IUCN,
sumably to avoid predation, while leaf litter is crucial for
nesting materials. With the proposed community conser-
RATHBUN, G.B. (1979). The social structure and ecology of ele-
vation agreement and the re-opening of the forest to
phant-shrews. Advances in Ethology, Supplement to Journal
selective timber harvesting, the elephant-shrew popula-
of Comparative Ethology, 20 : 1-77.
tion should be closely monitored.
RATHBUN, G. (1984). Elephant-shrews, Order Macroscelidea.
In : MACDONALD (ed), The Encyclopedia of Mammals. Facts
on File Publications, New York : 730-735.
StANLEY, W., S.M. GOODMAN & R. HUTTERER (1996). Notes on
the insectivores and elephant shrews of the Chome Forest,
The senior author of this study completed the project under
South Pare Mountains, Tanzania. Zoologische Abhandlun-
the guidance of the School for International Training (SIT) with
gen, 49 8 : 132-147.