Australian Flowering Plants. 211 



have been considered by some botanists as belonging to Hakea, 

 Grevillea, Persoonia, Banksia, Dryandra, and allied types. 

 . Although agreeing with Berry in his contention that the Pro- 

 teaceaB were at one time luxuriant types in the northern hemi- 

 sphere which were driven south at a later period, it is impossible 

 to review the evidence past and present without coming to the 

 conclusion that Hakea, Persoonia, Grevillea, Banksia, Dry- 

 andra, Protea, Leucondendron, and other allied types, never did 

 exist in the northern hemisphere, but that they arose in the 

 south as xerophytic developments of older widespread Creta- 

 ceous types. In Australia and Africa such genera as are enum- 

 erated here are all vigorous, hardy, specialized, aggressive, and 

 xerophytic. They avoid the Australian and African jungles; 

 they avoid all mild, moist, and sheltered positions ; and they 

 occur as highly specialized forms on the sandy barren waste of 

 the two countries under consideration. Xerophytic Myrtacese, 

 Composite, Leguminosse, Ericaceae, Epacridacea?, and Lilacese, 

 are their special associates, nevertheless like Eucalyptus, Mela- 

 leuca, Erica, Leucopogon, Goodenia, Candollea, Hibbertia, and 

 Pimelea, they appear to strain at a leash as though eager to 

 people new lands if only their large island home should be 

 connected by large barren sandy wastes with other countries. 



It seems permissible to infer that the development of the 

 Proteacese appears to have begun in the great northern lands, — 

 but it is difficult to decide upon which suborder was the earlier. 

 The Xucamentacese have the more regular perianth, but the 

 Folliculares also have many genera with regular perianths. 

 The high specialization, however, of certain genera such as 

 Hakea, Grevillea, Banksia, and Dryandra, in Folliculares, 

 especially with respect to their anthers, which are sessile upon 

 the perianth segments above the tube, together with the 

 irregular perianths of Grevillea and Hakea suggests the rela- 

 tive recency of such highly specialized genera and, moreover, 

 it would appear that they are vigorous and aggressive genera 

 because of such specialization. 



. The primitive type appears to have been a large tree of 

 luxuriant habit, with regular perianth, free stamens unattached 

 to the perianth tube, and with large pinnate, compound, or 

 lobed leaves. 



With the zoning of the climate, and the development of 

 aggressive plant types in the northern hemisphere, these Pro- 

 teace?e were gradually driven south, while the tropical types 

 became more and more superseded by the more complex Sym- 

 petala?, Dialypetala? and the Monocotyledons. The sandy wastes 



