PHYLOGENY OF THE PAL.EOGNATILE AND NEOGNAT1LE. 203 



In Rhea the nasal must be regarded as holorhinal, though the maxillary process is 

 absent. Inasmuch as this process, when present, serves as the boundary, caudad, of 

 the anterior nares, its place must be regarded as being filled by a vertical spike of bone 

 arising from the maxillo-palatine process to the anterior border of the lachrymal. In 

 Struthio this vertical spike — in Rhea separating the anterior nares from the lachrymo- 

 nasal fossa — is represented by a ligament which is attached dorsad to the free end of 

 the maxillary process of the nasal. 



In Struthio the nasal is holorhinal (PI. XLII. fig. 3). Its maxillary process is 

 well developed, and continued downwards on the maxillo-palatine by ligament. 



The Tinamidce may be regarded as holorhinal, though the nasal cleft has extended 

 backwards so far as to lie level with the free end of the nasal process of the premaxilla. 

 The maxillary process is not wanting as in Rhea, though reduced to a mere vestige. 

 The nasal cleft is bounded externally by a vertical, bony rod, representing the 

 maxillary spike found in Rhea. It differs therefrom in being rod-like, very slender, 

 and is perfectly free. Above, it articulates, by means of a short inturned process, with 

 the anterior border of the lachrymal and the vestigial maxillary nasal process, whilst 

 ventrally it articulates with the extreme postero-external border of the maxillo-palatine 

 process where it joins with the rod-shaped portion of the maxilla. This vertical rod 

 was regarded by Kitchen Parker as the maxillary process of the nasal. It is, however, 

 in all the skulls I have examined, quite distinct from this. 



The lachrymal, in Dromceus, sends backwards a very long, curved, rod-like process 

 to overhang, and thereby greatly increase the size of the orbit (PI. XLII. fig. 2). It 

 sends downwards a large antero-posteriorly flattened limb, the outer border of which 

 is produced still further downwards into a short rod-like spine which articulates with 

 the quadrato-jugal bar. The body of this ventral process of the lachrymal is perforated 

 by a large foramen for the passage of the lachrymal duct. The inner border of this 

 ventral process articulates with the antorbital plate. 



In Casuarius the lachrymal differs from that of Dromceus in the greater relative 

 shortness of its external orbital process, which is scarcely longer than the internal. A 

 further point of difference lies in the fact that the external orbital process passes 

 insensibly forwards and downwards into a laterally compressed plate, which, after 

 forming the boundary caudad of the posterior narial aperture, turns sharply backwards, 

 and becomes the antero-posteriorly compressed ventral process bounding the lachrymo- 

 nasat fossa as described in Dromceus. From this it differs in that the laminated portion 

 is continued downwards to the level of the quadrato-jugal bar, instead of stopping short 

 of this, and sending downwards thereto a spine from its outer border. The foramen for 

 the lachrymal duct is smaller than in Dromceus. Internally it articulates with the 

 antorbital plate. 



In Rhea the lachrymal resembles that of Dromceus. The supra-orbital process is 

 large. The lachrymo-nasal process is also antero-posteriorly compressed. The foramen 



