Apbil 5, 1907] 



SCIENCE 



551 



tautonomy, is nearly always the first species. 

 In tlie case of Stephens's numerous bird 

 genera, a species is generally figured to illus- 

 trate the genus, and this species nearly always 

 heads his list. Naturally, in subdividing 

 these heterogeneous groups, the illustrated 

 species is taken as the type. In other in- 

 stances it frequently happens that where an 

 author proposes a genus to which he refers 

 several species, all the species except the first 

 were already the types, or congeneric with the 

 types, of other genera. Sometimes the diag- 

 nosis shows that the author based his genus 

 primarily on the first species, and subsequent 

 revisers, in dismembering the group, have had 

 the good sense to restrict the original genus 

 to this species. In the non-monotypic Lin- 

 nsean genera the currently recognized type in 

 44 per cent, of these genera is some other 

 than the first species. These facts suffice to 

 show that the type may be quite often the 

 first species by ' coincidence,' or without the 

 conscious application of the first species rule. 

 11. In place of ' general statements ' a few 

 facts and figures may now be presented re- 

 specting the comparative number of name 

 changes rendered necessary through the strict 

 application, respectively, of the ' elimination ' 

 and the ' first species ' rules in bringing the 

 nomenclature of the genera of the Check- 

 List of North American Birds to a proper 

 standard of accuracy. Mr. Stone stated in 

 his first paper' that the number of changes 

 would be practically the same under each, 

 namely, fifteen by elimination and sixteen by 

 the first species rule. As said above, I have 

 spent a large part odL the last three months 

 in determining the types of the genera and 

 subgenera of the Check-List by both elimina- 

 tion and the first species rule. The Check- 

 List was taken as it was left by the publica- 

 tion of the last ' Supplement ' in July, 1904. 

 The changes later found necessary by the 

 Check-List Committee, being as yet not 

 officially published, are not considered. Also, 

 in order to show just how much truth there 

 is in the allegations that no two eliminators 

 ever reach the same results, owing to the 



' Science, N. S., Vol. XXIV., p. 562. 



' almost infinite variety of ways ' in which 

 elimination may be applied, I have eliminated 

 from the same basis as the original A. O. U. 

 Co mm ittee in preparing the first edition of 

 the Check-List, and their successors in pre- 

 paring the second edition and its subsequent 

 supplements; that is, I have taken the genera 

 at the dates and from the same sources as 

 they took them, even in the few cases where 

 later investigation has shown that they origi- 

 nated earlier and with a different constitu- 

 ency. These later discoveries are considered, 

 however, in making up the statements for 

 comparison with Mr. Stone's statistics, with a 

 view to treating both phases of the subject 

 with perfect fairness. 



(a) First as to the difference in results 

 reached by different eliminators. I disclaim 

 any knowledge of just how Mr. Eidgway and 

 Dr. Stejneger reached their surprisingly ac- 

 curate results. I followed my own method, 

 strictly and consistently, and did not check 

 up my results with the Check-List till my 

 work was finished. The total number of 

 errors of elimination in the Check-List suf- 

 ficiently serious to affect generic nomen- 

 clature is 3, which result in changing the 

 names of 2 genera and 1 subgenus. One 

 additional change, affecting two genera, is due 

 to the application of the principle of tauton- 

 omy, a rule not formulated till many years 

 after the publication of the Check-List, and 

 this change of names is thus not chargeable 

 to the A. 0. U. Check-List Committee as an 

 error of elimination. The type in five other 

 genera is transferred from one species to 

 another strictly congeneric with it, in three of 

 the five cases through the application of the 

 rule of tautonomy, but in none of these five 

 cases is a generic name affected. There are 

 thus, all told, five errors of elimination, only 

 three of which affect the names of species. 



As bearing on the question of alleged di- 

 versity of results through elimination it may 

 be noted that my results not only agree closely 

 with the Cheek-List, but also in every case 

 with Mr. Eidgway's recent eliminations in 

 ' Birds of North and Middle America ' (Vols. 

 I.-IIL, 1901-1904), and also almost invariably 

 with those of the British Museum ' Catalogue 



