East African coastal forest botany - Biogeography and Ecology of the Rain Forests of Eastern Africa - page 79

relaives elsewhere. If two communiies are apart
they also tend to diverge, through differences in
pattens of local exincion and immigraion
(Gleason, 1926; Macarthur & Wilson, 1967).
Species are not always found in all apparently
suitable coastal forests, especially where such
suitable habitats are isolated from others of the
same type. Conversely, some species of limited
geographic range occur in a wide range ofhabitats
locally. The Northen species Aristogeitonia
monophyla and Mitiostigma greenwayi, or
instance, occur in a broad spread of coastal forest
and thicket habitats around Mombasa, but not in
the Southern forests. This is no doubt because
they flourish in certain (limestone) areas in the
North, and from such bridgeheads make incur-
sions into temporary habitats (e.g. forest gaps and
stream banks over sandstones) in the surrounding
landscape. Similarly, Baphia puguensis which
favours steep sided valleys in the Pugu Hills
occurs from Moist valley forest to dry ridge top
thicket there, yet in spite of this local versaility is
unknown elsewhere.

environmental heterogeneiy, and because of
regular human and other disturbances.

These three actors, and paricularly the first
two can account in principle for the compleity
observed in the coastal forests, at all levels. The
north-south division of geographical elements
expresses part ofthis diversiy only at the broadest
level, and even so only rather crudely. The
Gleason individualisic model of the community is
most apt and evident in the coastal forests. It is for
this reason that coastal forest as a whole has to be
deined not by florisic content but by geographi-
cal posiion.
There is little doubt that the Tanzanian moun-
tain arc has been a refuge for coastal forest spe-
cies during inclement imes. Species surviving
climatic change in shifing belts of vegetaion in
the mountains, including species that evolved
there in isolaion, will have repeatedly recolonised
the coast when favourable condiions reuned. It
has been seen, however, that plant species have
almost certainly evolved within the coastal mosaic
itself. Coastal forests are today the only refuge for
a wide variety of plants, especially Dry forest,
riverine or rock-loving ones. There must have
been imes, possibly during drier climates, when
coastal forests have enriched the mountain com-
muniies. The coastal vegetaion is a bufer zone
between rain orest and mariime communiies,
and it seems likely that at least certain Oceanic or
coastal lineages like Ggasiphon, By ttneia and
Mitiostima reached the mountain forests via a
coastal forest form. At least the first of these is
capable of long distance dispersal across the sea
(Verdcourt, 1981).
Some Northern species may have previously
occurred in the South, and vice versa. Further
exploraion will probably show that some species
are distributed more widely than thought now.
However, the complex pattens and the north-
south differences in the coastal forest flora cannot
be attributed enirely to inadequate sampling or
random exincions. Species, even genera, have
been weaving themselves into and out of this
tapesry of communiies for many millennia.
There is little chance that any further develop-
ment will be one of enrichment, although one

3. Disturbance and compeiion. The influences
discussed in previous paragraphs shape the dis-
incive species pool in each area: disturbance,
compeiion and, on the level of single species,
more specific interacions influence on how the
members of this species pool associate on a finer-
grained level. The edge of environmental
determinism is sharpened by compeiion, and
especially by diffuse compeiion with the whole
of the surrounding forest. Conversely, the rela-
ionship between landscape and species distribu-
ion is generally made less close by disturbances,
especially canopy gaps. Gaps often allow the inva-
sion of species less well suited to the site if such
species happen to arrive, for instance from a
neighbouring patch in the mosaic, before a better
adapted species. The invasion of coastal forest
gaps by savanna trees like Brachystegia sp icormis,
or of Pugu Hills valley bottom forest by Baphia
puguensis, are manifestaions of this first-come,
first-served influence. The relaive conribuion
of this influence, i.e. the influence of the local
mosaic on one patch within it, is paricularly
important in the coastal forests because of the