The montane butterflies of the eastern Afrotropics - Biogeography and Ecology of the Rain Forests of Eastern Africa - page 160

montane forest butterflies of Cameroun seems
most probable, but we should await phylogeneic
studies of the species concened before drawing
conclusions.

of the mountain areas. Admittedly the species of
the Papilio nireus group in the montane forests
together cover all eight combined areas listed in
Table 8.5, but apart from the fact that there is
much distribuional overlap, all species together
or the monophyleic subgroup mackinnoni-
charopus-honimani which occurs in seven of the
eight areas, do not form a monophyleic group
with an alloparic lowland species urther west
(see Figure 8.3). Similarly the Charxes species
listed together cover most of the study area, but
they belong to a number of widely different spe-
cies groups (Henning, 1988). The montane
Uranothauma species combined also occur in
seven of the eight combined areas, but their inter-
relaionships are unknown and there is no good
lowland candidate in the genus to form a
monophyleic group with the montane species.
The Chondrolpis species are closely related,
orming a monophyleic group, but they are not
alloparic and the sister group has a wide distribu-
ion in EAT. A twin relaionship between mon-
tane forest species and lowland forest species to
the west can be found in a few instances only.
Ceratrichia bonga is the only species of the genus in
Tanzania, other species living further west, so
even though the phylogeny of the genus is not
known the sister species of bonga must be sought
for further west. Biylus knia is found in cenral
Kenya/north Tanzania, its sister species or sister
group (mananes and/or auriudus) lives from
West Africa to west Kenya. Biylus jfeyi occurs
from east Zaire to Kenya/north Tanzania, its sister
taxon (orothea and/or moyses) is ound in the
West and Cenral African forest blocks. For the
Biyclus species, see Figure 8.4.
There are no widespread montane forest spe-
cies that at the same ime occur in the Central
African forest block as well. Of the only five spe-
cies occurring n seven or all eight combined areas
(Table 8.5), two are resricted to the study area,
two occur in the highlands of Cameroun as well
and one is also found in the montane forests of
South Africa. Further, it is clear from Tables 8.4
and 8.5 and the remarks above (Disincion and
coherence of areas of endemism, p. 155) that the
numbers of species and of endemics are very
unevenly disributed.

ragmentation of a Tetiay orest belt
In this scenario the Easten Arc mountains should
have the highest concenraion of montane orest
butterflies. This is actually found, but the
decrease in numbers with increasing distance
from the Eastern Arc mountains, which is a fur-
her predicion, is not found at all (see Table 8.5).
The other predicion of this scenario is a twin
relaionship between taxa restricted to the Easten
Arc mountains and taxa in the Cenral African
lowland forest block. In the cases in which phylo-
geneic data are available (Metisella, Ch ondrolpis,
Papilio, Biyclus, Amauis) this patten is not
found. Even though exact knowledge of the
phylogeny of the other butterflies is very limited
or absent, a remark can be made since we do
know the relaives (i.e. congeners) and their dis-
tribuions. In those genera in which all species are
resricted to lowland forest except for one or more
species in the montane forests of the Easten Arc
mountains, we may safely assume that this patten
occurs. However, the only such genus is
Ceratichia. Since there are 28 genera with
endemic species in the Easten Arc mountains,
nine of which are resricted to forests, it cannot be
said that there is convincing evidence for the
eistence of a Teriary forest belt across the Afri-
can coninent. This agrees with what was found
about the predicion of decreasing numbers just
menioned. It does not mean that a coninuous
orest belt never eisted in the Teriary, only that
so far there is no evidence for it. If the belt ever
eisted the races have probably been obscured or
wiped out by later developments. This is not sur-
prising for events so remote in ime, certainly in
the light of the dramaic climaic changes in the
Pleistocene that must have upset all ecosystems in
he area.

Framentation of a Pleistocene orest belt
None of the pattens predicted by this scenario
are actually found. There are no groups of closely
related allopatric species, together covering most