Its Origin , Migrations , and Evolutionary Tendencies. 219 
type of plant is an older form than the temperate herbaceous type. In 
addition to the evidence from other branches of botany, their argument 
from the geographical standpoint is summed up as follows : ‘ There is great 
preponderance of herbs in temperate regions and of woody plants in the 
tropics. The latter climate probably approaches more nearly to that under 
which Angiosperms first appeared. Herbs have a short life cycle and are 
therefore able to survive periods of cold underground or in the form of seeds. 
Their great development in temperate regions has probably been in response 
to the progressive refrigeration of the climate during the course of the 
Tertiary.’ 
Though there are a great many herbs (especially bulbous Monocotyle- 
dons) in the temperate African flora, there is also a high proportion of small 
woody shrubs, and the number of endemics is exceptionally large. Now 
Sinnott and Bailey argue that the endemic plants in a flora are to be regarded 
usually as its most ancient element, a conclusion exactly the opposite of that 
reached by Willis (8). Sinnott and Bailey would have it that the herbaceous 
element of the temperate African flora is recent and derivative, while 
the woody element is ancient, a conclusion which receives no support from 
its present-day distribution, nor from the phylogeny of the families to which 
most of the woody types belong, Compositae, Ericaceae, Proteaceae, &c. 
The same areas are occupied by both elements, herbs and woody shrubs. 
Before discussing these views further we may investigate what is to be 
learned by comparing the temperate and tropical floras. 
Comparison of Temperate and Tropical Floras. 
Certain families have already been compared as regards their tropical 
and subtropical representatives. The conclusions there reached, viz. that 
the tropical types were older than the subtropical, apply also in many cases 
when the former are compared with temperate types. This is the case in 
the Violaceae, Loranthaceae, Thymelaeaceae, Apocynaceae, and Asclepiada- 
ceae taken together, and similarly the Myrsinaceae and Primulaceae and 
Araliaceae and Umbelliferae, each pair taken together. In other families 
the same thing is seen, the more primitive section in floral characters being 
tropical, the more recent temperate, e. g. Boraginaceae, Santalaceae, Ulma- 
ceae, Liliaceae, Rubiaceae, Leguminosae, and others. It is noteworthy that 
the more primitive tropical types usually have fleshy indehiscent fruits, 
while the more recent temperate types have usually capsular fruits. 
There are, however, a considerable number of families in which the 
temperate representatives have apparently retained more of the primitive 
ancestral characters in their floral morphology. In the Verbenaceae the 
subfamily Stilboideae has usually nearly regular corollas and endospermic 
seeds, and is more temperate than the Verbenoideae, which has irregular 
corollas and exendospermic seeds, or than theother subsections, the Viticoideae 
