° ' 19 ig j Correspondence. 447 



is, in the majority of cases, very small. I would say that whilst the proba- 

 bility is small it is not negligible. The persistence of Larus argentatus 

 smithsonianus in our lists through acceptance of unverified authority, is 

 evidence that this is a practical as well a theoretical source of error. 

 The use of the specific term where the subspecific differentiation is not 

 actually verified is a precaution against perpetuated error that is very 

 cheaply applied. By its very nature it can cause no confusion, for the 

 greater the subspecific certainty, the less necessity there is for naming it. 

 When the probability becomes certainty, the subspecific title may be a 

 convenience but it ceases to be a scientific necessity. The further the 

 probability departs from certainty the more advisable subspecific designa- 

 tion becomes but the more cautious we should be in applying it. It would 

 be interesting to know how often since their original description these 

 "American" races have been examined and compared with adequate 

 extralimital material. It would not be surprising if many of them failed 

 to stand modern tests and were found to rest on faith rather than fact. 



Your parallel between these subspecific cases and the New-world Tit- 

 lark is another question. Doubt unquestionably exists with species as 

 with subspecies and if there were as easy a way of generalizing upon them 

 I would advocate its adoption. As there is no such neutral course I do not 

 see that we can do other than get along as well as we can with an imperfect 

 system. Specific difficulties of this kind are comparatively few whilst 

 subspecific ones are legion and that we cannot correct the few former is 

 no argument against progress with the many latter when it can be obtained 

 by simple methods, already to hand, which have received the sanction of 

 leading authorities and the A. O. U. 



An objection that has been generally advanced against records made in 

 the manner under discussion is that they are "unquotable." I cannot see 

 that a definite subspecific identification made in a concrete statement in 

 small print is any less quotable than when made by inference in a heading 

 in heavy type. Where such identifications are not definite they certainly 

 can not be quoted as definite. The fault, if fault there is here, lies in the 

 writer, not in the method of presentation, for the latter certainly allows a 

 freedom of action that has great advantages. 



If there is no room in scientific literature for other than final results; 

 if no interest lies in specific facts and distributions unaccompanied by 

 subspecific identities, such identifieations, definite specifically but indefinite 

 subspecifically, as I have published lately are to be condemned. If how- 

 ever, we admit that a species is worthy of study as a species, and that 

 statements of evidence uninfluenced by preconceptions can be presented 

 as the basis for future generalizations rather than as finalities such lists 

 fill a valuable place in scientific investigation. In attempting an orderly 

 arrangement of our material we have had a tendency towards forcing of 

 facts into pigeon holes prepared for them. The attempt to make every 

 specimen agree with a preconceived scheme is not for the advancement of 

 science. To call intergrading, worn, undeveloped, mutating or wandering 



