562 



SCIENCE. 



[N, S. Vol. XXIV. No. 618. 



with either. Furthermore, many early au- 

 thors had no conception of a type species and 

 here it seems to me our selection must by any 

 method be an arbitrary one. 



(&) It is claimed that because we have fol- 

 lowed * elimination ' so long in certain groups 

 — as in North American birds, for instance — 

 we should be unwarranted in reversing our 

 method, because it would involve an immense 

 number of changes in generic names which 

 have been fixed by elimination. 



This is a serious question and one which I 

 have looked into very carefully. As a test I 

 have consulted the original publication of 391 

 genera and subgenera of North American 

 birds as given in the A. O. U. Check List, the 

 works in which the other 19 occurred being 

 inaccessible. I find 9 based upon diagnoses 

 without citation of species, while 5 are nomina 

 nuda or not used in a generic sense; 173 are 

 monotypic, 59 have their types designated by 

 their authors and 21 are fixed by tautonomy, 

 leaving 124 composite genera with no type 

 stated. 



Of these the type as accepted in the A. O. U. 

 Check List is the first species in 92 cases and 

 some other species in 32 cases. Of the latter 

 16 are Linnsean genera where the type is arbi- 

 trarily fixed (see antea), reducing the number 

 that would be changed by adopting the ' first 

 species' rule to 16. In two of these the new 

 type would be congeneric with the old, so that 

 there would be no change in the generic name 

 and three must be changed in any case for 

 reasons of priority. Some of the remainder, 

 however, involve the change of two names 

 each and the total change incident to the 

 adoption of the first species rule would be 

 10 generic and 4 subgeneric names. 



But let us look further. How consistently 

 has elimination been applied? There are 92 

 genera in which the first species is taken as 

 the type in the A. O. U. Check List, pre- 

 sumably as the result of elimination; but was 

 elimination employed in each case ? Let us see. 



In 25 cases the original species were all con- 

 generic, and elimination being impossible the 

 first species was selected as the type. In 5 

 only a partial elimination was possible and the 

 same plan was adopted. 



In 21 cases elimination, as I understand it, 

 fixes the type on the first species, as accepted 

 in the Check List, but in 12 cases it fixes it 

 upon some other species. There are also two 

 cases where the tautonomy rule will compel 

 a change and one where the designation of a 

 type by a subsequent author has been over- 

 looked, while 18 are Linn^an genera and 8 I 

 can not decide positively by elimination. 



In all 12 generic and 3 subgeneric names 

 will certainly be changed if elimination is 

 consistently applied, and the types of 5 other 

 genera will change but fortunately fall upon 

 congeneric species. 



It may be claimed that I did not eliminate 

 properly in all these instances, but in all con- 

 fusing cases I have followed the practise of 

 Dr. J. A. Allen, who was one of the framers 

 of the code and who has freely and cordially 

 advised me in this matter. 



It will be seen from the above that far from 

 causing an overturning of our ornithological 

 nomenclature the adoption of the first species 

 rule will cause less change than our adherence 

 to elimination. 



As a further test I have examined the bird 

 genera of the world from 1758 to 1820, com- 

 prising 513 names exclusive of Linnaeus. Of 

 these 282 are monotypic, 94 have their types 

 fixed by tautonomy, 18 are based upon diag- 

 noses only, leaving 119 composite genera in 

 which no type is indicated. The selected type, 

 according to the British Museum Catalogue, is 

 the first species in 102 cases and some other 

 species in 17 cases.* The task of working out 

 the results by elimination I have not ventured 

 to attempt. 



It remains now to show the various ways in 

 which 'elimination' is applied in practise. 

 Tor this purpose I prepared the following 

 series of questions which were intended to 

 cover the elementary principles of elimination : 

 Question I. 



Genus A, 1850. 



Species l = type of B 1860. 



2 = type of C 1870. 



3 = type of D 1880. 



* In these the action is usually arbitrary, 

 seldom or never the result of ' elimination.' 



