.REPLACEMENT OF Tlii: ANCIENT E. AFRICAN FOREST 491^ 



A similarity in the species comprising- Chi])ete and the part 

 of Chirinda that was, apparently, last connected with it is sug- 

 .Sl'estive. but not conclusive, for bird-carried species are largely 

 concerned. 



To sum up: Chirinda itself, shaped like an hour-glass by 

 thf tires that formerly ran in between its twin heights, shows 

 us an early step in forest-splitting. Chipete, only a few hundred 

 yards away, and with liighly-suggestive evidence of a former 

 Cfintinuity with Chirinda, the next. The other patches of the 

 countrv show us everv gradation in the matter of distance 'from 

 one another, and I have described an instance above in which 

 a gap was greatly widened b\' the coniiplete desitruction of a 

 forest-patch. The evidence generally is suggestive of former 

 continuity in the section I am particularly concerned with, and 

 of the view that the patches actually represent our most primi- 

 tive type of forest. 



Should we accept this much it is hard not to extend our 

 conclusions. The actual distribution of a number of plants and 

 animals, and the facts touched on in the last paragraph 

 of Section 2, suggest strongly that the idea of a great East 

 African forest tha't once connected Knysna with Elgon and 

 bevond is not far-fetched. Our Gazaland dense forests are the 

 meeting-place of Nyasan and Sotithern species, Podocarpus 

 inilanjiana carries us on beyond to the German and British East 

 African mountains, and cases like Khaya nyasica (not much more 

 than a sub-species of K. sciiegalcnsis) take us round with a 

 swiuiJ-. 7'ia Nyasaland, into the forests of the west. Finally, 

 Pscudoi'oly.v, Pcvcilosfachys, and other apparent relics of the 

 connection with Madagascar, enable us to picture the forest as 

 existing already in the earlier Tertiaries, when, as the great 

 luimmulitic deposits between Chirinda and the sea alone sufifice 

 to show, the country's configuration must have been very dif- 

 ferent, and land must have been under water that to-day carries 

 isolated forest and was once perhaps clothed with it. 



5. Factors in Forest-distribution. 



(fl) High Primitive Forest. — I have already incidentally 

 indicated the view (i) that the general area in which Chirinda- 

 like forest can occur is strictly limited by conditions of rainfall; 

 and (2) that one important factor which has brought about its 

 present scattered distribution is the annual grass- fires. It has 

 been suggested to me that the rainfall generally is gradually 

 diminishing, and that this may be the factor that is causing the 

 forests " to die out." 



Directly, no — or so one would say: for the surviving 

 forests are, for the mo.st part, in far too flourishing a condition 

 to allow one to suppose that any such detrimental influence is at 

 work within them. At the same time, a realization of the exact 

 conditions under which some of them grow suggests that the 

 possibility should not be lightly dismissed. Thus Neave (Geo- 

 graphical Journal, February, 1910) writes of the " mitu," or 



