CAMPBELL: THE ORIGIN OF THE HAWAIIAN FLORA 91 
quite in line with certain facts of plant distribution which appear 
to have been overlooked. 
The occurrence in the Islands of many hygrophilous liverworts 
and the filmy ferns (Hymenophyllaceae) seems to make it practi- 
cally certain that, as in the case of the snails cited by Pilsbry, the 
presence of these in the Islands can be explained only by considering 
them as remnants of the flora of a formerly much more extensive 
area connecting the Islands with some ancient continent. These 
plants are peculiarly unfitted for transportation over long dis- 
tances and it is difficult to see how they could possibly have sur- 
vived the exposure to heat and dryness to which they must have 
been subjected, assuming that they have come directly from either 
the American tropics or the remote tropical regions to the south. 
These plants inhabit, for the most part, the cool dark rain-forests 
and are quickly destroyed by exposure to the heat and sunshine 
of the lower levels. 
During the past summer the writer made a brief visit to the 
Islands, with a special view to examining the hepatic flora; and 
although the collections made were not as comprehensive as it 
was hoped to make them, owing to the remoteness of the col- 
lecting grounds, the results tend to confirm Pilsbry's view of a 
connection with the Malaysian and Australasian region. 
The most conspicuous of the liverworts in the lower forests 
are two species of Dumortiera, a genus peculiarly adapted to wet 
dark conditions. According to Stephani's* list of the liverworts 
of Hawaii, these species are D. trichocephala (Hook.) N. ab E., a 
species widely distributed through the eastern tropics, and D. 
hirsuta (Sw.) R. Bl. & N., an even more widely spread species. 
Itis probable, however, that a critical examination of the Hawaiian 
plants will show that they are not identical with those species. 
The so-called D. trichocephala is certainly quite different from 
material of the same species collected in the Malayan region, and 
the form attributed to D. hirsuta resembles very closely the Java- 
nese D. velutina Schiffn. Stephani states also that the monotypic 
genus Wiesnerella, which is closely related to Dumortiera, occurs 
in Hawaii, and immature material collected by the writer perhaps 
belong# here. This species occurs also in Java, the Himalayas, 
and Japan. 
* Stephani, F. Hepaticae Sandvicenses. Bull. Herb. Boiss. 5: 840-849. 1897. 
