Petrels from the North-East Pacific Ocean. 579 



we cannot exactly determine the proportion of the tarsus 

 to the middle toe and claWj but the tarsus seems longest. 



We would then recognise five genera, Hydrobates, Halo- 

 cyptena, CyinocJiorea, Oceanodroma, and Banuermania. The 

 first three agree in coloration and structure and are closely 

 allied, the last two resemble these in structure but show a 

 great discord in coloration and have probably little direct 

 relationship. To put it iu other words, we would give the 

 difference in coloration present more value than the differ- 

 ence in structure observed, or we would consider Halocyptena 

 more closely allied to Cymochorea than Oceanodroma is to 

 that genus. 



We would then recognise : — 



Genus Hydrobates 



Sole species Hydrobates jJelugica (L.). 

 Genus Halocyptena Cones. 



Sole sjjecies Halocyptena microsoma Cones. 

 Genus Cymochorea Cones. 



Species CymocJtorea leucorhoa (Vieillot) & subsp. 

 „ castro (Harcourt) & subsp. 



,, monorliis (Swinhoe) & subsp. 



,, liomochroa (Coues). 



„ macrodactyla (Bryant). 



,. melania (Bonaparte). 



„ markhami (Salvin). 



„ owstoni, n. sp. 



Genus Oceanodroma Reichenbach. 



Sole species Oceanodroma furcata (Gmelin). 

 Genus Bannermania Mathews & Iredale. 



Sole species Bannermania hornbyi (Gray). 



The names missing from the above list are 0. beali Emerson, 

 O. beldinyi Emerson, 0. kaedingi Anthony, and O. tristrami 

 Salvin. All these have been referred to Cymochorea, but Ave 

 are quite unable to determine the first three, while the last 

 is specially dealt with later. 



When Emerson (Condor, viii. 1906, pp. 53-55) intro- 

 duced his two forms he gave us numerals indicating the 



