856 A muds <f the South African Museum. 



the spines being the indurated persistent peduncles ; Bimorpho- 

 theca, 8 species,. 1 new; Tripteris, 11 species, 3 new; and Ursinia, 

 9 species. 



" The arrangement of the genera follows that of Benthani and 



Hooker's ' Genera Plantarum, 1 the species being more or Less 



after Harvey's account in the ' Flora Capensis,' vol. iii. The 



Pteronias are grouped after a revision by Dr. Phillips and the 



writer.* The descriptions of the new Pteronias in the following 



list are included in this revision." f 



The Compositae now described, together with the other families 



which have been or are yet to be published in this Journal, were 



obtained by expeditions which passed through several of the recognised 



South African floral regions. Many of the earlier records are of little 



value for phytogeographical purposes on account of their lack of 



precision. Since it is certain that the boundaries of some of these 



regions will require revision, it is desirable that the localities recorded 



below should be as clearly indicated as possible. It may, therefore, 



be useful to give a summary of the more important places visited 



by the four expeditions concerned, and an indication of the floral 



regions in which they are situated. These indications are liable 



to revision, since, in most cases, the floral region has been identified 



in the field, not by a study of the statistics of the vegetation — the 



datit for which are generally incomplete. Nevertheless, the collections 



among which these Compositae are included will undoubtedly throw 



some light upon the phytogeographical problems of South-West 



Africa, and certain conclusions, briefly referred to in the following 



notes, are already foreshadowed. 



In the summary of localities given below, as in Mr. Hutchinson's 

 list, the "Cape" region is synonymous with that which Bolus called 

 "The South-Western Coast region." X The change is made merely 

 for the sake of convenience and, in particular, to obviate any possible 

 confusion with the " Western Coast region " to which a large pro- 

 portion of these plants belong. The latter region, as defined by 

 Bolus, § includes that part of the Khamiesberg range, of which little 

 was known of the flora when Bolus wrote. The following account of 

 the distribution of the Khamiesberg Compositae confirms the im- 

 pression derived from a study of the rest of the known vegetation of 

 this range, viz., that botanically its outstanding relationships are 



* Ann. S. A. Mus., vol. ix, 1916, pp. 277-329. 



t Loc. cit., p. 



X Bolus, loc. cit., 207. 



§ Loc. cit., 204. 



