HARPER: Trip ON WarRIOR AND TOMBIGBEE RIVERS 113 
glass, my notes made on the boat were chiefly confined to trees, 
shrubs, and vines. But our frequent stops at the bluffs, lasting 
from half an hour to several hours or all night, gave me oppor- 
tunity to note many herbs and check up my identifications of the 
woody plants, and to do a little collecting. A few short trips 
away from the river on foot were also made, the longest being from 
Beckley’s Landing in Marengo County via Myrtlewood to Naheola 
ferry, a walk of eight or ten miles (with one companion), which cut 
off about fifteen miles of river. The plants noted on such trips 
are not counted with those on the river-banks, however. 
One extreme method of treating the plants observed would be 
to combine them all in a single list, and the other would be to 
consider each geological formation separately and classify the 
plants growing on or near it according to habitat, as far as pos- 
sible. But the former method would obscure some interesting 
features of distribution, and my notes are not complete enough to 
make it worth while to attempt the latter. Another method would 
be to consider the Warrior and Tombigbee Rivers separately, thus 
dividing the journey into two equal parts. But this would be 
rather arbitrary, for the upper Tombigbee does not flow through 
any kind of country that the Warrior does not, so that there 
was no perceptible change of natural conditions as we passed 
Demopolis except an increased flow of water. (If we had started 
at the head of navigation on the Tombigbee instead of on the 
Warrior, then Demopolis might have been a logical dividing point, 
for the Warrior doubtless brings down from the mountains seeds 
of quite a number of plants which do not grow along the upper 
Tombigbee at all.) 
The dividing line between the Cretaceous and Eocene forma- 
tions is a line which several species growing along the river do 
not seem to cross, and by dividing the notes there some interesting 
features of distribution can be brought out. That method is here 
adopted. 
In the following table the plants of the Cretaceous and Eocene 
portions of the river banks are listed in parallel columns. They 
are divided first into trees, shrubs, and herbs, then arranged in 
order of apparent frequency, the number prefixed to each being 
the number of times it was noted in the region under which it is 
