( 7o) 



on a level with the lower edge of the occipital condyle, froiu which it is separated 

 hv a very slightly marked precondylar fossa. At the sides it is raised into the very 

 large and prominent mammillary processes (ra.) to which ^Ir. Forbes has drawn 

 attention : they extend below the paroccipital processes, and are separated from them 

 by a deep cleft at the bottom of which the vatpis foramen, opens. In the smaller rails 

 the mammillary processes are very slightly developed ; in Aptornis they are carried 

 far below the level of the occipital condyle, owing to the peculiar arrangement of the 

 precondylar region, which is directed downwards nearly at right angles to the long 

 axis of the skull. The anterior portion of the basitemporal platform is somewhat 

 inflated, and is united laterally with the prominent pretemporal wing. 



The rostrum resembles that of Ocijdromits. At its base, immediately in front 

 of the eustachian opening, there is a small longitudinal cleft, a]i]iarently the remnant 

 of the pituitary s})ace. 



The form of the tympanic cavity is very similar to that of Ctihalas or Ocydroiuan, 

 differing only in the deepening of its posterior wall consequent upon the increased 

 size of the paroccipital process. The pretemporal wing above mentioned forms a 

 large rounded prominence, internal to the articular surface for the inner head of 

 the quadrate. 



The ethmoidal region also closely resembles that of Ocydronius, but the 

 antorbital plate is somewhat less developed. In all the skulls I haxe examined the 

 lachrymal has lieen lost. 



The beak is very long and decurved, but both the length and curvature are 

 subject to considerable individual variation. The relation of the premaxillae and 

 nasals to one another and to the froutals is the same as in Ocydromus. The narial 

 opening, though large, is less in proportion to the size of the skull than in the 

 smaller Oeydromines : and the increased length of the beak in the fossil is entirely 

 due to the elongation of the premaxillae in front of the nostril, thisprenarial portion 

 being rather more than half the total length of these bones measured along the facial 

 process. In this elongation of the beak Cfihalus sylvestris, among living (Jcydromine 

 rails, approaches the fossil most nearly. The palatal surface of the united premaxillae 

 is channeled by a deep longitudinal groove, narrowing somewhat from behind 

 forwards. 



Themaxillo-palatine platesand the jialatines are very Ocydromine, but the scroll- 

 like portion of the latter bones is njore decurved and inroUed at the edges. i\I]-. Forbes 

 has jiointed out that the posterior ends of the palatines do not meet in the middle 

 line beneath the rostrum, but this may be merely due lo the loss of the pterygoids 

 allowing them to spring apart in the dry skull. The vomer, which is very rarely 

 preserved (it is present in a specimen in the Natural History Museum), is relatively 

 shorter than in Ocydroiuus, and does not api)ear to have anchylosed with the i)alatines. 



The quiulrate has the articular surface for the mandilile similar to that of 

 Ocydromiis or Cabalus sylvestris, but its orbital process is proportionately larger. 



The mandible much resembles those of Ajjhanapieryx and Erythromuchus, but 

 there are some well-marked differences. .\nteriorly it is much more sharply 

 decurved (in this respect, however, as in Ivrylhrouiachnii, there is considerable 

 variatioMj, and the symphysial portion ^e(■ln^ to be wider and niucb less com])ressed 

 than in Aphanapteryx. 



The articular region is proportionately larger and more uiassi\e than in that 

 genus; the outer articular surface projects beyond the ramus, and the inner is 

 broader and ajiparently more continuous with tlie outer. The posterior border of the 



