1920.] Birds of the Canary Islands. 545 



Formation of Insular Races. 



In recent years vei'v close attention has been paid to the 

 investigation of ge()2ra|)hical races — sul)species as we term 

 them to-day. As a result of tliis seareli for new forms (which 

 has been sadly abused l)y some, who have thus brought the 

 careful work of many systematic workers under the lash of 

 those ornithologists who have clung to the older traditions) 

 the trinomial systen^ of nomenclature has gradually come 

 more and more into general nse. This is not the place to 

 enter into a discussion as to the merits or demerits of the 

 system, but in a paper dealing with insular races the utility 

 of the trinomial system is indisputable and must be apparent 

 to anyone who studies the question with an unbiased mind. 

 Whereas under tiie old binomial system the geographical 

 races of the (/anarian avifauna were given tlie status of 

 species, we now^ realise that many of the birds in these 

 islands are really only insular forms of a continental parent 

 race from which they have sprung. In some cases the 

 distinction may be only small, in others much greater 

 ditiercntiation will be found, depending on the amount of 

 variation Avhicli has taken place; but in both cases the bird 

 can be considered a subspecies (and for the purposes of 

 classification it is nsnnlly more convenient to do so), and as 

 such is named trinomially. Tiie dividing line between a 

 subspecies and a sj)ccies is difficult to define, and this is 

 where the champion of the binomial system generally com- 

 mences liis argument ! 



All field naturalists and uiauy ])urely systematic orni- 

 thologists know how many ami varied are the factors which 

 combine to bring about variation in a bird. It may be 

 quite by chance — for a few wanderers blown out of their 

 usual course may have been compelled to take refuge on 

 some oceanic island and there have come under the influence 

 of the very factors which would most likely tend to produce 

 variation in their case. 



This is exactly what has happened in certain cases in the 

 (.'anaiy Islands, and although those who have studied the 



