3i 
illustrated by the South African Flora . 
graphical or evolutionary factor, whatever we care to call it, we are 
neglecting another very useful method of attack, and we fail to explain 
phenomena as fully as we might. The dominance of Podocarpus fatcata 
in certain Drakensberg forests cannot be explained by considering habitats 
alone, for it is absent from apparently identical habitats in other parts of 
the range. The reason becomes clear when we consider its origin from the 
widespread P. latifolia. The presence of a mesophytic species of Rhus 
in a patch of moist scrub is not due simply to the fact of the more meso- 
phytic conditions alone. We must connect its presence there with that of 
a widespread parent species over drier areas. In a word, it is not the 
present species alone that we must study, but wherever possible we must 
go back a step or two and consider allied or parent species and their 
behaviour and distribution. 
The whole of our argument is based on the hypothesis that a species 
in the course of its migrations, when it comes into contact with conditions 
different from those which produced it, is, in many cases at least, capable 
of differentiation or of giving rise to new species suited to the new conditions. 
In a larger sense, and over a longer period of time, the same applies to larger 
groups, the genera, tribes, and families. While, in this paper, effort has been 
directed chiefly to bringing forward evidence of as varied a nature as possible 
to show that this takes place, there are many other interesting points which 
have not been dealt with. What conditions, for instance, are most effective 
in producing new. species? Is the passage from tropical to temperate 
conditions more important than from dry to moist (or vice versa in 
each case) ? 
The examples of tropical species being replaced in South Africa by 
nearly related more temperate species are very numerous, but this is to be 
expected from the size of the areas involved. On the other hand, the 
examples of xerophytic shrubby species being closely allied to mesophytic 
forest species are also very numerous. The transition from siliceous to 
calcareous soil conditions has been closely investigated in Europe from this 
standpoint, but few observations have been made in South Africa on the 
effect of differences in soil conditions. 
While the ecological aspects of this study are what have been kept most 
in view the methods of studying present-day tendencies obviously have 
a bearing on the question of the origin of the South African flora as a whole 
and its connexions with outside areas. From what was said in the intro- 
duction, it is clear that there are only two general views to be considered. 
One lays emphasis on the South African connexions with the floras of 
Australia (particularly South-West Australia) and of South America and 
postulates land connexions to explain them, across what are now deep 
ocean basins. The difficulty with hypothetical changes of sea and land on 
a vast scale is always, as Wallace has expressed it again in his last work 
