MAXON—STUDIES OF TROPICAL AMERICAN FERNS. 27 
general resemblance among the species of Cnemidaria, which stamps 
them as obviously related and as forming a natural group; but that 
there is any logical ground for setting these apart as a genus distinct 
from Hemitelia is not apparent. 
As with most members of the Cyatheaceae there are serious difh- 
culties to be met in a satisfactory delimitation of the species, mainly 
on account of their great size and the consequent incompleteness of 
herbarium material. Most collectors seem to have thought it of 
little importance to collect the different characteristic parts of the 
plants. Naturally several species have been redescribed under new 
names; and most of them have been very inadequately described, 
usually with scant attention to venation, which, however unsatis- 
factory it may be as a generic character in this group, yet seems to be 
fairly constant for the recognition of species. Latterly there has 
accumulated at Washington and New York a large amount of material 
which affords the main basis for the present paper. It would have 
been impossible, however, to bring the work to a satisfactory con- 
clusion without the courteous assistance of the authorities of several 
European herbaria. Acknowledgments are therefore gratefully 
extended to the Director of the Royal Gardens, Kew; to Mr. A. B. 
Rendle, Keeper of the Herbarium, British Museum (Natural His- 
tory); to Mr. Carl Christensen, of the Botanisk Museum, Copen- 
hagen; to Prof. H. de Willdman, Curator, Jardin Botanique de I’ Etat, 
Brussels; to Dr. H. Christ, Basel; to Dr. E. Rosenstock, Gotha; and 
especially to Prof. Dr. I. Urban, Assistant Director of the Royal 
Botanical Garden and Museum, Berlin. 
From data and specimens thus made available, and with the help 
derived in several instances from Doctor Underwood’s unpublished 
notes, it has been possible to bring to completion that which appeared 
at first an almost hopeless task. It has been found that, with ample 
material, definite lines of demarcation may be drawn about as sharply 
among the species of this group as in related genera of the Cyatheaceae. 
The segregation of the species, indeed, has offered far less difficulty 
than the purely taxonomic problem of reapplying several of the early 
names in their original sense and of indicating with certainty their 
later synonyms. That several of these names are now to be applied 
in a sense wholly or partly different from their employment by recent 
writers will occasion at most a slight and temporary confusion, since the 
species of this genus have been so illy defined in recent years and their 
limits so misunderstood and so persistently disregarded that there 
is, in fact, no recognized treatment of the group to be disturbed thereby. 
For the same reason a rather large proportion of the species here 
recognized must be described as new. Within the past 15 years 
but two species of the subgenus Cnemidaria have been described from 
North America as new, and one of these (H. bullata), from Grenada, 
proves to be synonymous with an early species. It is entirely probable 
