12 Bews . — Some General Principles of Plant Distribution as 
objection, it would, in this case, be brought forward only to justify the first, 
since some explanation on the theory of migration would naturally be 
expected of one species being widespread and another closely allied species 
having restricted or discontinuous distribution. The objection as a whole 
applies chiefly to polygenesis as an explanation of discontinuous distribu- 
tion. The case for polygenesis has already been sufficiently argued, and it is 
maintained that the onus of proof is on the other side. It is clearly as difficult 
to disprove polygenesis as to disprove migration, and the former, in many 
cases, is the simpler explanation. Some critics would admit that wide- 
spread species might produce closely allied but not identical species at 
two or more separate points, i.e. they readily admit No. i but deny No. 5 
of our categories as given above. In reply to this, it might be asked, how 
often does it not happen that the systematist allows himself to be influenced 
by geographical distribution in deciding that two closely allied species are 
distinct? It makes him magnify minor differences which are not constant 
and it makes him sometimes see differences which do not exist. The result 
is clearly reasoning in a circle. Species are said to be distinct simply 
because they are widely separated geographically and the chances of 
migration are remote, and then it is denied that identical species can arise 
in separate geographical areas. 
A third objection is that many of the examples quoted are not really 
distinct species, though they are so described in the ‘ Flora Capensis’. That 
probably in many cases is quite true, but it is not really an objection, for, 
though they may be considered merely varieties, the same evolutionary 
principles are involved. In fact, it has only been because it was thought 
unnecessary that the words £ or variety ’ were not repeated after £ species ’ in 
each case above. The £ Flora Capensis ’, too, has been the work of many 
different specialists, and it is unlikely that they all have erred in the making 
of too many species. The earlier work of Harvey more probably erred in 
the other direction, for he was not inclined to split species too much. In 
a large flora of about 13,600 species the mass of evidence to be sifted is so 
great that the presentation of it is a matter of considerable difficulty. 
Considerations of space prevent us from dealing with all the families. An 
effort has been made to make as varied a selection of examples as possible, 
and if any reader is still not convinced, he can read through the ‘ Flora 
Capensis ’ for himself and find many more. Our aim for the moment is to 
illustrate the method, and its bearing on other aspects of botanical study 
will be discussed later. Though the ‘ Flora Capensis * has been mentioned, 
it must be admitted that the earlier volumes are so incomplete as a record 
as to be of little use, though the later volumes are better. I have relied 
chiefly for details regarding distribution on the various local check-lists, 
supplemented in the case of Natal by my own various field investigations. 
The lists referred to are those by Bolus and Wolley-Dod for the Cape ( 14 ), 
