146 ORIGIN, ETC., OF THE FLORA. 
Outside of America, the exploration of Central Africa has yielded the most startling 
results. Contrary to expectation, the relationships between the African and American 
floras are almost as strong as those between the African and Asiatic floras, including 
representatives of such otherwise exclusively, or almost exclusively, American families 
as the Canellacee, Caricaceez, Humiriacee, Hydnoracee, Loasacee, Mayacacee, 
Napoleonacez, Rapateacee, Strelitziacee, Velloziaceee, and Vochysiacee. There are 
equally striking generic and specific connections. Altogether there are records of the 
same, or specially representative, genera and. species belonging to upwards of sixty 
families or distinct groups. These facts, coupled with the zoo-geographical data 
and the paleontology of the region, have given rise to the theory of a Jand-connection 
between Africa and South America in Eocene times, a theory accepted by both 
botanists and zoologists of high repute. Von Ihering’s projected map of the assumed 
‘“Verteilung von Land und Meer zur Eocién-Zeit ” offers a plausible solution of some 
of the problems of the present longitudinal distribution of organisms, alike in southern 
latitudes and in remote islands. As long ago as 1885 (Introduction to the Botany of 
the ‘Challenger’ Expedition), I rejected the theory of the vegetation of remote volcanic 
islands and groups of islands, such as the Galapagos and Hawaii, being necessarily 
of derived origin rather than remnants of a former wider flora. This view does not 
exclude dispersal by various agencies in distant parts of the world and natural 
migrations across the widest continents, as exemplified by some certainly introduced 
plants. . 
The explorations of the last twenty years in Western and Central China furnish 
further evidence of the existence of close relationships between the floras of eastern 
temperate Asia and eastern North America. Types illustrating this feature are. 
usually representative species of near affinity, as, for example, of the genera 
Liriodendron, LIiquidambar, and Sassafras. This American-Asiatic element extends 
southward to Mexico, and is more prominent in the mountains of eastern South 
Mexico than it isin the north. The following genera of forest-trees represented in 
the Atlantic States, absent from the Pacific States, reach Mexico, namely :—Magnolia, 
Asimina, Tilia, Robinia, Liquidambar, Ilex, Diospyros, Bumelia, Ulinus, Celtis, Morus, 
Ostrya, Carpinus, and Carya. In a general sense, the western coast-forests are 
coniferous and the eastern deciduous. Certain genera are apparently now confined, 
or nearly so, to America and the Mediterranean region (including the Azores and 
Canaries)—such are Lelianthemum, Lupinus, Heberdenia, Platanus, and Corema. 
Apart from the subantarctic flora, there are genera that extend from Chili to Mexico 
and from Australasia to Borneo. The magnoliaceous genus Drimys belongs to this 
category. Judging from analogies, this is an example of a northern extension of 
a southern type. The Proteaces, now almost exclusively confined to the Southern 
Hemisphere, furnish a similar instance in the closely allied genera Helicia and Roupala. 
The former ranges from Australasia to China, Japan, and India, the latter from 
Brazil to Mexico, and neither has further extensions. Of course, the absolutists 
