430 



SCIENCE. 



[N. S. Vol. XXI. No. 533. 



diversity of iisage or gi-eater strenuosily of 

 opinion than on this, although the tendency 

 is, or was formerly, to follow one of two 

 courses, either to take the first species as the 

 type, or to determine the type by the prin- 

 ciple of elimination, under certain reasonable 

 restrictions. Of late the latter has been the 

 course favored by the greater part of those 

 systematists who have any special regard for 

 rules of nomenclature. Two qualifications of 

 the strict rule of determining the type by 

 elimination have been widely accepted. One 

 is that when a genus containing a number of 

 species is divided, and the name of one of the 

 species is chosen as the name of a new genus, 

 the type of that genus shall be the si)ecies the 

 name of which has been selected as the name 

 of the genus — a perfectly logical, unequivocal 

 proceeding, open to no reasonable objection. 



A second exception is that of the A. O. U. 

 Code, which provides that if a " genus con- 

 tains both exotic and non-exotic species — from 

 the standpoint of the original author — and 

 the generic term is one originally applied by 

 the ancient Greeks and Romans, the proc- 

 ess of elimination is to be restricted to the 

 non-exotic species." In this way the name is 

 retained in nearly its ancient sense, and its 

 transference to an irrelevant association is 

 prevented. This exception comes in mainly, of 

 course, in connection with Linnaean and Bris- 

 sonian names, and is akin to that other rule, 

 more or less tacitly held in the minds of many 

 systematists, that the type of a Linnsan genus 

 should be the best known European or officinal 

 species originally included within it. 



Canon XXI. of the A. O. U. Code is: 

 " When no type is clearly indicated the author 

 who first subdivides a genus may restrict the 

 original name to such part of it as he may 

 judge advisable, and such assignment shall not 

 be subject to subsequent modification." This 

 was not a new rule when announced by the A. 

 O. U. in 1886, but was a part of the British 

 Association Code originally promulgated in 

 1842, and reaffirmed by nearly every later code 

 down to 1905, when three revolutionary ichthy- 

 ologists came forward with the following as 

 their Canon X. : " The type of a genus can be 

 indicated by the original author only. * * * 



In every case, the determination of the type 

 of a genus shall rest on evidence offered by 

 the original author, and shall be in no wise 

 affected by restrictions or modifications of the 

 genus in question introduced by subsequent 

 authors, nor shall the views or the dates of 

 subsequent authors be considered as affecting 

 the assignment of the type of a genus " ! For 

 such a reactionary and far-reaching proposi- 

 tion there should certainly be most convincing 

 and satisfactory reasons, for it involves the 

 overthrow of the consistent usage of the 

 majority of systematists for the last half cen- 

 tury, and invites at least temporary chaos 

 in the place of what seemed permanent sta- 

 bility. The proposed new ruling should leave 

 nothing to personal opinion, but should pro- 

 vide a rule of unquestionable applicability to 

 all cases. 



The argument for the new proposition is as 

 follows : " It is believed that the principle that 

 a generic name must be fixed by its original 

 author is one of vital importance in nomencla- 

 ture. All processes of fixing types by elimina- 

 tion or by any other resting on subsequent 

 literature, lead only to confusion and to the 

 frittering of time on irrelevant questions. 

 The method of elimination can not be so 

 defined as to lead to constant results in dif- 

 ferent hands. In general it is much more 

 difficult to know to what types subsequent 

 authors have restricted any name than to hnow 

 ivhat the original author would have chosen 

 as his type. Most early writers who have 

 dealt with Linnaean species have consciously 

 or unconsciously encroached on the Linnaean 

 groups rather than made definite restrictions 

 in the meaning of the generic names." 



In determining types and the tenability of 

 names it is notorious that the systematist is 

 and must be guided by what an author has 

 done and not by what he may have intended 

 to do, no matter how evident the unaccom- 

 plished intention may be. Rules, to be effect- 

 ive, must be rigidly enforced, regardless of 

 personal preference in favor of some particu- 

 lar result. But the foregoing is a proposition 

 to override rules and usages that have brought 

 nomenclature to a reasonable condition of 

 stability respecting a wide class of cases it is 



