376 Correspondence. \^ 



(4) as show only a slight deviation from the allied continental type, and 

 we also treat very many forms as trinomials which from lack of material 

 we are undecided whether to place in (2) or (3). 



Personal opinion must necessarily govern such cases, no matter what 

 Code we set up, just as it must govern all cases where 'degree of differ- 

 ence' is adopted as our criterion. Considering the great diversity of 

 custom at present, it seems to me time that we came to some definite 

 agreement on the matter, and our practice shows that ' degree of differ- 

 ence ' must influence us in certain cases. 



To my mind (A), binomials should be applied to all forms which 

 occur together without intergradation, no matter how slight the differ- 

 ences, and (B) trinomials, to geographic races which intergrade, or which 

 differ so slightly that there is every probability of intergradation, and to 

 slightly differentiated island forms. Only such geographic races should 

 be considered as species (binomials) as are markedly different, and of the 

 intergradation of which there is no probability. In other words, where 

 intergradation is probable, give it the benefit of the doubt. This practice 

 is nearly that followed by the A. O. U. Committee, but is at variance 

 with that of many of our mammalogists, with whom the custom seems 

 to be to call everything a species until intergradation is proven ; which 

 will speedily result in the adoption of a binomial name for every geo- 

 graphic variation — a most undesirable state of affairs and a distinct 

 retrograde step in nomenclature. 



Witmer Stone. 

 Academy of Natural Sciences, Philadelphia. 

 Sept. 6, 1S99. 



{Without at present attempting to discuss the question of the ap- 

 plication of binomials and trinomials in its broader aspects, there is 

 one fact in connection with the naming of insular forms which Mr. 

 Stone and other writers 1 on this subject have apparently not considered. 

 In challenging the propriety of giving a trinomial name to an insular 

 form on the ground that the nature of its range renders geographical 

 intergradation with its nearest ally impossible, they evidently have not 

 given due allowance to the possibility of intergradation through indi- 

 vidual variation. 



Island forms, as all systematists know, because of their isolation are 

 often separated on the basis of characters too slight to warrant similar 

 action if they were inhabitants of the mainland. Hence it frequently 

 happens that among a large series of a given form from a certain island 

 there will be found a number of individuals indistinguishable from this 

 form's representative on the mainland or on a neighboring island, and 

 vice versa. Thus, for example, when we examine large series of 

 Pyrrhulagra noctis or Dendroica petechia from the West Indies we find 

 a complete intergradation of the extremes and, at the same time, average 

 differences among the series from the different islands of sufficient 

 importance to be recognized trinomially. — Frank M. Chapman.] 



1 Cf. William Palmer, The Nidologist, III, 1896, 91. 



