43 [Vol. XXXV. 



query (?). Mr. Grant goes so far as to say that 'Hhe de- 

 scription of S. dactylatra given by Lesson is unrecognizable.''^ 

 This is, however, not the case, for although the description 

 of the plumage, "^pure white^ wings and tail black," would 

 apply approximately to three species occurring on or around 

 Ascension Island (the type-locality o£ S. dactylatra) , the 

 possession of a "semicircular naked throat patch ^'' excludes 

 S. capensis, which narrows it down to two. The statement 

 that the beak is horn-colour at once eliminates the red- 

 billed Sula piscatrix, called by Lesson, on the same page^ 

 S. erythrorhyncha. Therefore the oldest name for the 

 "^ Masked Gannet " of the Atlantic Ocean is Sula dacty- 

 latra Less., and this name must be used in the place of 

 S. cyanops Sund. 



Now, on going into the question of local races_, I have 

 been able to distinguish the following five formSj including 

 a new subspecies : — 



1. Sula dactylatra dactylatra Less. 



Bill horny blue-grey, very slender ; feet and legs yellow. 

 Range. Ascension Island and S. Atlantic Coasts. 



2. Sula dactylatra melanops Hartl. 



Bill greenish-yellow, slender; feet and legs slaty-blue to 

 dull black. 



Range. Western Indian Ocean (Bed Sea and Islands 

 north of Madagascar) . 



3. Sula dactylatra personata Gould. 



Bill yellow, very stout and large ; feet and legs greenish 

 blue. 



Range. Western Pacific. 



4. Sula dactylatra californica, subsp. n. 



Bill bright yellow, very thick ; feet and legs orange. 

 Ra?ige. Coasts of California and Central America. 

 Type in the Tring Museum : ? ad. San Benedicto 

 Island, 7. xii. 01. 



Obs. In addition to the colour of the soft parts this race 



