194 BOTANICAL GAZETTE [SEPTEMBER 
choked by the density of the Scevpus americanus Pers., among 
which they grew. Where the shores were free from Scirpus, as 
they were for stretches north of Bush River station, P. B. & W 
R. R., and at the mouth of Otter Creek, they were composed of 
pure gravel, subject to shifting with the action of the waves. 
y. I have indicated thus the location of a well-grown speci- 
men found floating in the bay at the mouth of Furnace Creek 
over 3*™ from the nearest known colony. 
It appears from this list of known stations that /soetes saccharata 
has a general distribution in the fresh water portions of Chesa- 
peake Bay and its tributaries, but is at present unknown from 
any other locality. If it has a wider distribution, our ignorance 
of that fact is not entirely due to lack of observation, for Dela- 
ware Bay, which ‘furnishes the most accessible suitable habitats, 
has been explored by students of the genus, with the result that 
several stations have been located for J/soetes riparia Engelm., 
but not a single specimen of J. saccharata Engelm. has been 
found. , 
If the last named species is indeed limited to Chesapeake 
Bay, it will be of interest to consider the causes to which the 
restriction may be due. Endemic species have a peculiar interest 
in their bearing upon problems of biogeography. A cosmopoli- 
tan distribution is evidence of a high degree of adaptability to 
variation of habitat, and is also evidence either of an old species 
or of efficient means of dispersal. On the other hand, a species 
_ which is limited to a single locality is either a remnant of a once 
more widely distributed form, or it is a relatively new species 
which has arisen in the more or less isolated region to which it 
is now limited. In the first instance it has been protected by 
the barriers which surround it, or by peculiarly favorable con- 
ditions for its growth from the extinction which has overtaken 
the species in other regions; in the second case it has been pre- 
vented by barriers from spreading beyond the region of its 
origin. Cosmopolitan species seem to deny the existence of 
barriers, while local species not only confirm their existence but 
give a clue to the nature and position of these barriers. As 
biogeography has to do preeminently with barriers, it is evident 
