Foi 
~ 
36 ca AMERICAN FERN JOURNAL 
but vestigia terrent. These groups ate not based on a 
single character, but the species belonging to the same 
subgenus agree in several respects. The characters 
used are often apparently very minute and have previ- 
ously been seldom considered as of any value, but they 
are, nevertheless, very constant. The best and most 
constant t character i is the pubescence, i. e., the type of 
and seales. I have never found that a spevies 
varies as to the kind of pubescence; further, species that 
agree in other characters correspond exactly in pubes- 
cence, not in the more or less dense covering by | 
~~ scales but i in the shape of the hairs and scales. For 
ve always tog multicellular, articulated. hairs : 
ery characteristic toothed seales; those ae aS 
have always ste 
