Regional Distribution of the Cape Flora. 251 



The order Umbellifera? is perhaps one of those which, 

 as a rule, have the most marked general resemblance. Yet 

 in South Africa I found at Somerset East in the wooded 

 kloofs an umbelliferous tree, Hctcroniorpha arhorcscens, whose 

 leaves are usually teruate but often almost simple, while 

 Alepidca, amongst the long grass and bulbous plants on the 

 top of the mountain, has a rosette of entire ciliated leaves 

 and a striking resemblance to what one may call the 

 Hieracium type. On the other hand, near Seapoint, and in 

 certain other places, I foimd a rosette of very broad, flat 

 ciliated leaves which for a long time puzzled me ; eventu- 

 ally I found the flowers, and made it out to be a dicecious 

 nmheUiier, Arctopus echinatus ; while the nearest approach to 

 the ericoid form is shown by a Ehyticarpus {BiLpleurum 

 difforme) which has only the petiole left, and rigid, almost 

 woody, stems. In Orchids, Disa grandiflora, the Pride 

 of Table ^Mountain, has very large flowers, but few in 

 number, while a common species in the eastern districts, 

 Disa polygonoides, has a very large number of excessively 

 small flowers crowded together at the ends of the peduncle 

 with exactly the same general appearance as the flowers of 

 plants of the ericoid type. The Lilies, and in a less degTee 

 the Irids, also seem to me to show similar tendencies. 



This may seem very speculative, but I have so often noticed 

 the same thing, and in such difterent plants, that I think it 

 will supply at least a good working hypothesis in studying 

 the influence of physical conditions. 



