The montane butterflies of the eastern Afrotropics - Biogeography and Ecology of the Rain Forests of Eastern Africa - page 161

Thus, the pattens ound do not support the
idea that the butterflies of the montane forests are
simply lowland forest species trapped in the
mountains when the forests became isolated
there. In this connecion it is signiicant to note
that species-rich forest genera like Telipna (27
species), Micropntila (3 1), Pentila (36), Liptena
(70), Ep itola (82), Euriphene (64) and Euphedra
(137) have no species at all in the montane forests
of EAT. It seems unlikely that a coninuous forest
belt would have left no traces of the occurrence of
these genera in EAT.

Figure 8.2), Papilio aethipsis, P. ch rapkowskii (see
Figure 8.3), Etesiolaus sp., and Gnphoes grogani.
Adaptaion from another habitat, most probably
woodland, must be assumed for Abantis meu,
Ampittia pava, Colotis elgonnsis and Yp thima
albia. In the other 37 genera such habitat
switches may also have occurred but the phylo-
geny is unknown and there are possibiliies for
twin relaionships that do not agree with the
present patten. The patten could as well be
common, but we must await phylogen e ic studies.
For the same reason we cannot say much about
predicion (2). Predicion (3), however, the high
concenraion of endemics in the Easten Arc
mountains, absolutely agrees with the patten
found. It is remarkable that there is one other
combined area, the Westen Border Range, with a
high concentraion of endemics. Since this moun-
tain range is much younger than the Easten Arc
mountains, and not older than other mountain
ranges in EAT, its high endemicity (although
lower than in the Easten Arc mountains) must
have a cause other than aniquity of the habitat.
The Westen Border Range differs from other
mountain ranges in EAT in the close proimiy of
a large reservoir of lowland forest species (and
thus, candidate highland forest species), i.e. the
Central African forest block. It could be an
explanaion for the high endemicity in the
Westen Border Range relaive to the other young
mountains in EAT. This idea would be supported
if we were to find that the endemics of the
Westen Border Range each had sister species in
the lowland forest nearby. Another possible
eplanaion will be given in the following
paragraph.

Ecological dferentiation
Staring rom the viewpoint that preference or
the montane habitat is a derived condiion, the
montane forest butterflies either arose from low-
land forest butterflies or from butterflies living in
non-forest habitats. The possibility of the mon-
tane forest butterflies originally being lowland
butterflies from the west trapped in mountain
refugia has just been dealt with. These are frag-
mentaion scenarios. Here we are concened with
a possible adaptaion rom a non-forest habitat or
a lowland forest habitat to montane forest without
geographical isolaion. If this really occurred the
following pattens are expected (see descripion of
the scenario, above):

1. Twin relaionships between montane
forest and neighbouring lowland forest or
open habitat/woodland within EAT.
2. No recurrent geographic patten of
phylogeneic relaionships.
3. Highest concentraion of endemics in
Easten Arc mountains, unless the species
originated here could later spread through
the other mountain ranges (see Intenal
development, p. 148).
Direct evidence of pattern (1) is rare since we
know so little about the phylogeny of the species
concened, but there is also indirect evidence (e.g.
because the species is the only one of the genus
living in montane forest, and all candidate sister
taxa occur nearby at lower levels and/or in other
habitats). Thus, adaptaion from lower to higher
alitude forest must be assumed for Goryra
bibulus, the Chondrolepis species combined (see

Intenal velopment
Here we must test two possible scenarios that may
both have occurred but not simultaneously: (i) the
montane forests once were coninuous irrespec-
ive of possible contacts with oress to the west,
and (ii) there has been large-scale local and dif-
fe renial exincion caused by desiccaion. The
relevant pattens are: (i) widespread species or
widespread monophyleic groups of alloparic
species; (ii) highest concenraion of endemics
and of species in general in wetter areas, i.e. he