Mammals in the forests of eastern africa - Biogeography and Ecology of the Rain Forests of Eastern Africa - page 237

the more conservaive form of tree hyrax should
be resricted to the Tanzanian mountains and
islands, having presumably been displaced from
the surrounding lowland forests, which are now
inhabited by the very widespread D. arboreus,
which is intermediate in form between D. validus
and D. osalis (Kingdon, 1971). D. arboreus is in
tum a montane species in western Uganda, while
the lowland forests are in the process of being
colonised by D. orsalis (which is replacing
arboreus). t is as though three layers of advancing
arborealists are superimposed on each other.
However, the hyrax story may have sill further
subtleies, because the rocky alpine habitats on
Mount Rungwe are occupied by the semi-
arboreal Heterohyrax brucei laemanni although it is
enirely ringed by forest where Dendrohyrx
arboreus is dominant. This suggests that a Hetero-
hyrax which had already adapted to subalpine
Mount Rungwe might have pre-empted later col-
onisaion by Dendrohyrx.
The suni, Neotgus moschatus, has an extensive
range all down easten Africa. It represents a
stable and successful but broadly very conserva-
ive bovid type that is almost certainly of non-
forest origin. Its closest relaives, . batesi in the
Congolian block and N. pygamaeus in upper
Guinea, are wholly forest dwellers but appear to
be more derived species than soni (see Kingdon,
1982a).
The greater galagos have recently been sub-
divided into two disinct species (Olsen, 1979).
One of these is the easten endemic Ga/ago
ganetti, which is more terresrial and can tolerate
very much drier thickets and more open forests
than its ecological equivalent the potto, Perodiai-
cus potto, which feeds on similar foods and shares
several aspects of its ecological niche with the
greater galago. G. rssicauatus (unil recently the
species to which G. ganetti was assigned)
occupies drier counry including arid woodland,
and the absence of any large galagos anywhere in
the main forest block suggests that galagos,
especially the larger species, were never true for-
est species. If this also applies to the smaller spe-
cies this would put the easten endemic G.
zanzibaicus into a crucial intermediate posiion
because it resembles both G. emiovii in the main

forest block and the Senegal galago G. senegalensis
in the savannas. A thick-tailed dwarf galago that
builds leaf nests and may be semi-social has
recently been discovered to coeist in forest south
of the Rufiji River with the much commoner and
more widespread G. zanzibaius.
The dog mongooses also conorm to a patten
of disjuncion between easterly and westerly for-
est forms. Bdeogale crassicaua is a more conserva-
ive species that is widely disributed in easten
African woodlands and forests. The Zanzibar
Island race, B. c. tenuis, has been suggested to be
the sranded fragment of a sill earlier population.
Further up the coast another race, B. c. omnivora,
may represent an incipient species. This
mongoose is also sranded in the northernmost
coastal enclave of Sokoke Forest and is inter-
mediate in colouring and tooth structure between
B. rssicauda and B. nigipes, which is a specialist
forest mongoose that has a patchy disribuion
between Cameroun and Mount Kenya. Stranding
is likely to be an appropriate way of describing the
origins of several Zanzibar endemics but especi-
ally a giant rat of peculiar affiniies, Cricetomys
cosensi. This pouched rat is related not only to the
widespread . gambianus but also to the easten
endemic Beamys hinei which, perhaps because it
is extremely rap-shy, has been recorded in only a
few montane and coastal localiies between the
mountains around Lake Nyasa and the East Afri-
can coast. The interesing point about . cosensi is
that its closest affiniies are with . emini, a
western forest species that is unknown from
Tanzania. . ga mbianus occupies a wide variety of
relaively moist habitats on the mainland, where it
is sympatric with Beamys in the few places where
the latter has been found.
If . gambianus is the most recently evolved type
of pouched rat, which seems likely, the relict
status of Beamys, a smaller but demonstrably less
evolved member of the same lineage, is more
comprehensible. What is uncertain is the degree
to which Beamys is physiologically or adapively
restricted to montane forest, riverine and lowland
well-wooded habitats, or is limited by compeiion
with other rodents, notably Cricetomys. Its forest
or near-forest adaptaion could thereore be a by-
product of its evolutionary and compeiive