^if^ RRTL ■\CK:NrENT OF THE ANCIENT E. AFRICAN FOREST 



stron.s^ly-iiiarked dry season, il would a])|)ear thai this forest 

 g-ives place, after its destruction for cultivation and the subse- 

 quent aband(^nnient of the crround, to secondary forest more or 

 less dense, which becomes largely replaced in time by the 

 orig-inal forest. Below that figure, and with a sufficiently 

 strongly-marked dry season, annual fires take place outside the 

 forest, and the latter, as it is destroyed by them or by cultiva- 

 tion, is rejjlaced by grass land, of fire-bearinu' grasses, inter- 

 mingled throughout great areas with a secondary tyi)e 

 of wooding which is also ])yroi)hytic. This is vsry dif- 

 ferent, apparently, from the secondary type in the other class 

 of area, and to be regarded as intervening between it and the 

 primary type; for an equivalent to what is elsewhere the secon- 

 dary type occurs here as a third stage, should fires cease. The 

 latter is at first i)artly a matter (^f the denser massing of the 

 ]jyrophytic trees themselves, but it is chiefly composed, eventually, 

 of shade-beariny species that belong neither to the true forest 

 nor to the pyrojihytes. Where these thickets, or the secondary 

 pyro])hytic bush, are near enough to true forest to receive seeds 

 from it, they are eventtially swami^ed and replaced bv it — fires 

 still not taking place. 



One other point is ])erha]is of some importance. Our 

 retreating forests of to-day are surrounded by gradtiated zones, 

 varying in width with circumstances, that might be called 

 dcnudatioii-zones, should we acce])t wash as the chief factor in 

 the impoverishment that follows forest — ^though nati\e cultiva- 

 tion and burning contribute strongly — or, better, ijerha])s, liiiiuiis- 

 .'::oiies. When the forest was still unbroken, but was already 

 retreating before the fires, these zones would have existed as 

 continuous lines or concentric rings along or round the wh(~)le 

 forest area, and even the inner lines would still have been con- 

 tinuous for some little time after the forest became broken 

 through, while the outer zones may be regarded as still connected 

 to-day. As the fragments became more numerous and widely- 

 sj^aced, various com])licating factors will have arisen. Members 

 of a ]5articular zone wdll have persisted far longer in some cir- 

 ciunstances than in others ; the inner zones Avill not only ha\e 

 surrounded forest-patches, Init have survived them for a time 

 as islands, or, should the forest have advanced again, have dis- 

 appeared through being swallowed up in it, and changes, through 

 siltins', etc., in the distribution of the richer elements of the soil 

 will have led to much dis])lacement. Allowing for all this. 

 I believe that a fuller realization of* the former continuity of 

 forest and the consecpient extension of the zonal idea to eml)race 

 these sub-divisions of some of the altitudinal zones, dei)endent 

 not merely on altitude but on degrees of soil-impoverishment fol- 

 lowing the destruction of forest sub-divisions, Avill give us 

 the correct basis for a study of distribtition, as well as for our 

 more local necological studies, faunal and floral. Thus it is 

 highly suggestive that in the Chirinda region our connections 

 with Kilimanjaro, both in l)irds and plants, extend to several of 



