NOVEMBEB 2, 1906.] 



SCIENCE. 



561 



sible, fortified by rules that will leave no 

 chance for personal opinion and subsequent 

 alteration. 



With this idea in mind I have given the sub- 

 ject of ' type-fixing ' much thought and study, 

 taking for my especial investigation the genera 

 of North American birds, a group which for 

 twenty years has been constantly under the 

 scrutiny of a committee on nomenclature and 

 which has been subjected to about as much 

 changing as any group of genera with which 

 I am acquainted. 



As to the merits of the two principal meth- 

 ods of fixing types, my investigations lead me 

 to strongly favor the plan of selecting the 

 first species mentioned.^ 



Its advantages are: 



(a) Personal opinion is eliminated, two per- 

 sons can not reach different conclusions. 



(h) The question is settled independently 

 for each genus, the result does not depend 

 upon the fijiing of the type of some other 

 genus. 



(c) The possibility of change in a generic 

 name rests solely upon the question of pri- 

 ority, and the discovery of an error in the 

 usually accepted date of a publication has no 

 bearing upon the types of genera. 



(d) It is necessary to consult only the orig- 

 inal reference to ascertain the type of the 

 genus. 



Contrasting with these my objections to the 

 method of ' elimination ' as embodied in the 

 A. O. U. Code: 



(a) It permits the greatest range of per- 

 sonal opinion in the method of its application 

 and the almost endless combinations of prin- 

 ciples which it presents. 



(b) In ascertaining the type of one genus 

 it is often necessary to eliminate one or more 

 others first and an error in one operation 

 affects the others; in fact the genera stand in 

 an interdependent series and a change in the 



* In the case of Linnsean genera I realize that 

 no good can come of enforcing this rule, but we 

 have practical unanimity of opinion on the types 

 of these genera, and I see no reason why we 

 may not accept them arbitrarily just as we ac- 

 cept the genera themselves as our starting point. 

 The A. O. U. Code moreover does not demand con- 

 sistent elimination for Linnaean names. 



type of one may affect a number of others. 



(c) The discovery of an error in the date 

 of a publication affects not only the priority 

 of the genera therein described but also every 

 operation of elimination where these genera 

 have been involved; and the types of other 

 genera will be altered simply because the type 

 of one of these genera has been taken out at 

 the wrong date. 



(d) To ascertain the type by elimination it 

 is necessary to consult every work in which 

 genera have been erected upon any of the 

 included species; also every work where some 

 subsequent author may have specifically select- 

 ed a type for the genus. It is manifestly 

 impossible to be sure when one has exhausted 

 the latter literature. 



Dr. C. W. Stiles's method, as detailed in 

 his paper on ' The Determination of Generic 

 Types,' ^ seems to me to be the perfection of 

 the ' elimination ' idea and while better than 

 that, inasmuch as it is more complete and 

 more logical, it is open to objection in even 

 greater degree on account of its necessary 

 complexity. 



While I have the greatest admiration for 

 Dr. Stiles's handling of this subject, I can not 

 see how his method can be generally adopted. 

 The systematist can not afford to waste time 

 in studying the application of nineteen rules 

 and recommendations containing thirteen sec- 

 ondary suggestions in fixing the type of a 

 genus. What he must have is simplicity and 

 definiteness. To use a mathematical simile 

 he wants elementary arithmetic rather than 

 calculus. 



Now as to the arguments advanced in favor 

 of elimination. It is claimed that: 



(a) It upholds the work of our predecessors 

 by accepting the genera that they have from 

 year to year separated off from the original 

 composite genus, so that the residue must be 

 what they regarded as the type of the original 

 genus. This argument however, amounts to 

 little or nothing, as in the past many men were 

 working wholly independently of one another 

 and by ' elimination ' we inextricably confuse 

 two or more independent lines of work, arriv- 

 ing at results which are probably not in accord 



'Bull. 79, Bureau Anim. Indust., U. S. Dept. 

 Agric. 



