( 249 ) 
Reproduction is normally sexual. In some of the dioecious spe- 
cies, antheridia are very rarely found and development must in many 
cases be parthenogenetic. Vegetative multiplication also takes place 
in various ways: by bulblets, formed upon either stem or rhizoids, 
by branches with naked base, or by branches similar in all respects 
except origin to the pro-embryos. 
The systematic eccentricities of the family are of this nature. 
When thoroughly understood a great majority of the plants collected 
will fall with perfect distinctness into a few well-marked groups, 
though more than one under the prevailing nomenclature is spoken 
of as aspecies. Yet within the limits of these groups the differ- 
ences may be very considerable and of the most varied nature. 
Even a so-called variety may come to comprise plants which show 
many distinguishing characters. Further, the grouping thus pro- 
duced sets at defiance all laws of geographical distribution. Chara 
fragilis Desv. is reported from all of the continents; Chara 
gymnopitys A. Br. was originally described from Tasmania, 
C. gymnopitys keukensts Allen from Lake Keuka in New York 
State. Here the species and so-called variety are very different, 
though certainly in the same group; in the other case plants from 
Asia and America greatly resemble one another. 
Again, the dioecious C. cazescens Loisel. is widely distributed in 
Europe, but male plants have been collected on extremely few occa- 
sions. It occurs in eastern America, although here male plants 
seem unknown. But in the west of America, from the Saskatche- 
wan to California, there occur two extremely similar species, C. Azr- 
suta Allen and C. evoluta Allen, unquestionably monoecious, and 
another species frequently identified with the last has been collected 
in central Asia. C. canescens, generally known by its later name 
C. crintta Wallr., in its European acceptation is an aggregate spe- 
cies, but were it not for this one seemingly constant character C. evo- 
Zuta would certainly not be the first species to be segregated from it. 
It is at present the custom to divide these blanket species into 
‘¢ forms,” and when this is done the latter are almost always found 
to have a well-marked distribution, limited only too often to a single 
pool. But so far as evidence is available, there is every reason to 
believe that each form retains its peculiarities constant for long 
periods of time. It might seem that the proper course of procedure 
is to raise these forms to specific rank, but to this there are objec- 
tions. The number of resulting species not only would be very 
