424 Mr. H. Seebohm on the 



3. The coracoids are not furnished with subclavicular 

 processes. 



4. The vomer is ankylosed with the maxillo-palatines. 



5. They possess basipterygoid processes which spring from 

 tlie body of the basisphenoid^ and articulate with the ptery- 

 goids as near the quadrate as possible. 



6. In the articulation of their dorsal vertebrae they are 

 heterocoelous. 



7. They have never more than one dorsal vertebra in front 

 of the ankylosed sacral vertebrae^ which is without a ventral 

 process. 



8. The humerus is pneumatic. 



The 4th character is diagnostic of the Crypturi, but they 

 are also easily diagnosed by their combination of the 1st and 

 3rd characters. 



The Tinamous divide with the Penguins the honour of 

 being the oldest family in this order. The 3rdj 4th, and 5th 

 characters are all struthious. The sternum is^ however, in 

 no respect Ostrich-like. It resembles that of the Phasia- 

 nidse in having a well-developed keel, springing from a very 

 narrow median xiphoid process, but differs from it in having 

 the forked and flattened lateral processes represented by a 

 long slender unforked rounded process on each side. 



There can be little doubt that the Crypturi are more nearly 

 allied to the Gallinte than to any other group of birds. 



Galling. 



If the Gallinse be restricted to the three families Phasia- 

 nidae, Cracidaj, and Megapodidse, their diagnosis becomes 

 very simple. They all agree in having the following cha- 

 racters : — 



1. In the bifurcation of their nasals they are holorhinal. 



2. The ankylosed sacral vertebrae are preceded by a free 

 vertebra, in front of which are four ankylosed dorsal vertebrae. 



3. The coracoids are not furnished with subclavicular 

 processes. 



4. The basipterygoid processes are always present, and 



