V0 i- 8 £ VI ] Correspondence. 375 



insular birds being obviously prevented from intergrading with main- 

 land forms, ought to be uniformly provided with binomial names. Any- 

 one who takes the trouble to examine the Check-List and the recent Sup- 

 plements thereto will find that this custom is by no means followed out, 

 as very many of the recently described insular forms from the California 

 islands, which differ but slightly from allied mainland birds, appear as 

 trinomials. In other words ' intergradation' has been disregarded and 

 'degree of difference'' recognized as the criterion for deciding the specific 

 or subspecific claims of a given form. 



We have, for instance, Carpodacns mcgregori from San Benito and 

 Carpodacus mexicanns dementis from Santa Barbara ; Aphelocoma insu- 

 laris from Santa Cruz, while the Helminth op hi la from the same island 

 is H. celata sordida. The Guadalupe Island birds are, I believe, without 

 exception, written as binomials, but the vast majority of the other insular 

 birds which have been described recently are listed as trinomials. 



In recent American mammalogy the tendency is in the other direction, 

 and not only are all island forms, no matter how slightly differentiated, 

 regarded as 'species' (i.e., binomials), but many peninsular and other 

 continental forms which may readily be expected to intergrade are 

 treated in the same way. In fact, the tendency among our mammalo- 

 gists seems to be to depart from trinomialism altogether. 



This to my mind is much to be regretted, and will work irreparable 

 dam'age to nomenclature. A trinomial name carries to the average stu- 

 dent just twice the information that a binomial would under these cir- 

 cumstances. For instance, to one who is not conversant with every 

 paper relating to modern mammalogy, how much more knowledge the 

 name Lynx canadensis subsolartus conveys than Lynx snbsolanus. The 

 former indicates at once a race of the Canada Lynx, the latter leaves him 

 in doubt whether the animal is related to the Canada Lynx or the Wild 

 Cat {L. ruffus). It seems that some modification of the A. O. U. Canon 

 relating to trinomials is desirable, especially as we seem to be deliber- 

 ately violating it, but this can surely be effected without abolishing this 

 extremely useful system. 



Animal forms (using this term for any recognizable species or sub- 

 species) are of four kinds: — (i). Those which exist side by side in the 

 same area without intergradation as the Hermit and Olive-backed 

 Thrushes. (2). Those which inhabit different areas and intergrade 

 where the areas join. These are obviously modified from one far rang- 

 ing form which is being broken up by different geographic environ- 

 ments. (3). Those which inhabit different areas, but which do not 

 intergrade and are often separated by wide gaps. (4). Island forms 

 which are often closely related to nearby continental forms, but are of 

 course completely isolated. 



By the A. O. U. Code trinomials can only be applied to forms coming 

 under category (2), and all others are treated as binomials. By common 

 usage in ornithology, however, we adopt trinomials for such forms under 



