252 UNDERWOOD: REVIEW OF THE GENERA OF FERNS 
4. While homonyms that may in the future interfere with sta- 
bility are not to be admitted as valid genera, the extreme and in- 
flexible application of the law to those cases where interference is 
clearly impossible is neither necessary nor desirable. Such cases 
as that of Angiopteris Hoffm. will not be displaced because of a 
eru of this nature. 
. In general, the laws of evidence and interpretation usually 
Piece will be followed with the view of ascertaining an author's 
intended meaning. Mere technicalities based on legal quibbles 
have no place in a rational system of.nomenclature. Here, as 
elsewhere, it is “ the letter that kills but the spirit that maketh alive.” 
x 
+ > 
With these principles in view we proceed to a brief historical 
tice. The method of ‘‘ residues” works on-the principle that the last species remain- 
ing in the genus from those originally named by its author when the genus was founded 
shall constitute the type of the genus and shall hold the generic name. This method, 
moreover, has the disadvantage of being variously interpreted by different people. 
While it may be true that early writers did not always name the most typical species. 
first, it is equally true that the last species to remain in the genus after division has taken 
place is less likely to be a typical form; in fact, if doubtful or little known species are 
placed in a genus originally, these are the very ones which are unlikely to be taken away 
from the original group to form other genera. The system of ‘‘residues,’’ moreover, 
has a shifting instead of a fixed type, varying as more and more of the original species 
are taken out to form other genera, thus leaving a smaller and smaller residue. 
of the many reasons for eS e pene of the first species under a 
genus, stead of that of ** residues ” are the wing : 
s more direct, simple, and ads in its application. 
2. Iti is strictly in accord with a system which recognizes priority of publication as 
a fundamental principle. It conforms with the practice generally in use for recognizing 
the type of a species originally described from composite material. 
. It is in accord with a starting point for genera (1753) which in the absence of 
generic — can only base its original genera on species instead of generic 
descriptio: 
Iti tis d universal application, while the principle of residues, besides being capa- 
ble of various interpretations, cannot be applied to the many cases in which several 
tations on the groups involved in Polystichum Roth (1800), Aspidium Swz. (180 1), 
itn cc (1801), and € Rich. (1803) —generic groups based on 
varying 
a ki is the least unjust to el deber of genera based originally on several species. 
6. It anchors a generic name toa species with which it must rise or fall, according © 
species is or is not a part of a distinct generic group, and prevents it.from 
being shifted along to a smaller and smaller group, or even of being shifted to a group 
_of species of which its original describer knew nothin ng. 
