102 Mr. H. Seebolim on the 



least two important characters from each of the other genera. 

 In tlie following key each genus is treated as a family, but 

 the only distinction between the Plotidse and the Phalacro- 

 coracidae is that in the latter the two ventral processes on 

 the 7th to the 13th cervical vertebrae do not meet to form a 

 carotid arch, but they bend towards each other almost suf- 

 ficiently to do so. In fact the difference between Phala- 

 crocorax and Plotus appears to be less than that between 

 Phalacrocorax and Fregata or Phaethon. In the latter the 

 ventral processes are short and parallel. 



Clavicle ankylosed to ) "^ 

 stermira. 



Pelecanidce 

 SuUdce . . . . 

 Flotidce. . . 



Several cervical vertebrae 

 v^'ith carotid arches. 



Phalacrocoracidce . [ Dorsal vertebrae with ven- 

 Nasal aperture large. 1 C tral processes. 



Palatines not co- > Phaethontidcp .... 

 alesced. ) -' 



In the following key three perhaps less important cha- 

 racters are used to diagnose the families, but Plotus is re- 

 garded as part of the Phalacrocoracidse. Which of the two 

 modes of treating Plotus is the more scientific must be left 

 for future investigation to determine. 



All the dorsal verte- 

 brae heteroccelous. 



Humerus not pneu- 

 matic. 



P}iaetho7itid(e. 

 PelecmiidcB . . . 



Subclavicular process arti- 



Fregatid<B )■ culating Avith scapular, 



but not with clavicle. 

 I Phcdacrocoracidce 



\ Stdidce. 



It is an interesting fact that the egg of Phaethon is spotted, 

 and consequently differs from the eggs of all the other genera 

 of the Steganopodes, but the significance of this fact is les- 

 sened by the knowledge that Platalea is equally aberrant in 

 its suborder. 



Phalacrocorax, Plotus, and Sula are so absolutely similar 



