illustrated by the South African Flora . 29 
species. Otherwise none of the ordinary ecological methods could 
account for their presence in one place, their absence in another. 
Coni ferae. 
The genus Podocarpus is of great interest, not only because it has 
been found in the fossil Tertiary deposits of northern Europe and America, 
though it is now confined to Japan and the Southern Hemisphere and reaches 
such widely separated points as New Zealand, South Africa, and southern 
Chili, but also because species are so often dominant in the chief South 
African forests. 
The two species Podocarpus latifolia (P. 7 hunbergii) and P. elongata 
are quite distinct and have, more or less, the same range. They are both 
old species ; Podocarpus falcata is closely allied to P. latifolia , and it is 
probably identical with, or at any rate extremely close to, P . melanjianus . 
P. falcata is completely dominant in forests of a narrow belt along the 
southern end of the Drakensberg, but is apparently as completely absent 
from exactly similar habitats north of the Mont aux Sources. If it is the 
same species as P. melanjianus it begins again in Rhodesia and East 
Africa, and we have a striking example of discontinuous distribution in 
a dominant forest species. If it is not the same species as P. melanjianus , 
we have a case of two derivatives from the same species differing slightly 
in two distinct areas. 
Isolated Types. 
The above survey of families, genera, and species has now been made 
of sufficient length to show that the flora of South Africa affords abundant 
evidence of the origin of new species, according to the principles sum- 
marized to begin with. If I had cared to utilize all the notes that have 
been prepared the survey could have been extended to at least three times 
the length. While in the larger genera it is often very easy to see the 
nearest relationships and probable origin of rare endemics or the splitting 
up of variable species, &c., there are, on the other hand, many isolated' 
species or genera, or even families (e. g. Bruniaceae), to which at the present 
stage of our knowledge it is impossible to apply those methods. Many 
examples are to be found among our forest-trees and shrubs, e. g. Ocotea 
bidlata , Pygeum africanum , Cunonia capensis , Clausena inaequalis , Calo- 
dendron capense , Ilex mitis , Noltea africana , Hippobromus alata , Curtisea 
faginea , Heteromorpha arborescens , Leucosidea sericea , Cunonia capensis , 
Platylophus trifoliatus, Choristylis rhamnoides , Xymalos monospora , Gerrar- 
diana foliosa , &c., which have no obvious connexions in South Africa, 
though several of them do have such in the tropics. In the latter case, 
their distribution in South Africa must be explained by the ordinary methods 
of migration. They must be assumed to have migrated into South Africa 
by themselves. They are not the descendants of immigrants. In the case 
