( 74) 



and more iioufilul muscles, and consetitu-ntly moit' strongly marked surfaces and 

 processes for their attachment. 



The foramen vutgnmii is oval, with its long axis vertical ; it is slightly Hattened 

 below and constricted in the middle by a i)air of small processes which arise just 

 within its margin and project forwards and inwards. It resembles that of Aptornis, 

 but differs from those of Ocijdrontus, C'dhuhm, Hijijotiieiiidin, and ErythromaclnhS, 

 which are as wide or wider than high, and have a nearly straight lower border. The 

 occipital surface above the foramen slojies more forwards than in the smaller Ocydro- 

 mine rails, apparently owing to the relatively smaller size of the cerebellum. In 

 this respect there is more resemblance with Aptorais. From the upper edge of 

 the foramen a romided occipital ridge, most strongly developed in the larger skulls. 

 runs u}) to the lambdoidal crest. This latter is strongly developed ; in its middle 

 portion it runs transversely in a .straight line, and does not dip down towards the 

 foramen mufjnura, as in Aptornis. Laterally it is confluent with the temporal 

 ridge and forms the hinder border of the temporal fos.sae; its lower portion con- 

 stitutes the outer liorder of the immense [)aroccipital ]irocesses. These ])roject 

 considerably below the occipital condyle, and are much larger than in any other 

 rail I liave examined, though Aptoi'nis approaches them in this respect. Their 

 ventral border is divided by a semicircular incision into a larger outer and a smaller 

 inner lobe. In Aptornis there is no trace of this arrangement : but in Oci/drohm-s 

 and Cabalus the small paroccipital processes are distinctly bifid. 



The cranial roof as far as the middle of the orbits is convex ; Imt anteriorly the 

 frontals are depressed in the middle line, the depression being deepest just behind 

 their junction with the premaxillae. .\mong the smaller rails this character is best 

 developed in Ocydroruus and Culxilns dieffenbachii, a nearly complete skeleton of 

 which is in the collection. 'J'he iuterorbital region of the frontals is much shoi'ter 

 and liroader in Diaphorapteri/.r than in the smaller Ocydromines ; in Porpkyrio, 

 on the other hand, it is considerably wider; while in Aptornis. owing to the 

 extension backwards of the na.sals and premaxillae, it is i>ractically wanting. 



The temporal fossae are \ery large and deep; they extend well up into the roof 

 of the skull, and are there separated by a space about equal to the width of the ft-ontals 

 between the orbits, but varying somewhat in different individuals. I'osteriorly the 

 temporal ridge is confluent with the lambdoidal crest, as above mentioned ; anteriorly 

 in runs on to the outer surface of the postorbital process, enclosing a sort of lobe of 

 the main fossa. In the smaller Ocydromines this anterior lolie is usually a separate 

 muscle impression, marked off from the tem])oral fossa by a more or less clearly 

 defined ridge. 



As in Oci/droiHKs, Gabalits, and llypulneuidiu, the u))per border of the orbit is 

 thick and truncate as far back as the postorbital process ; in Erylhromachus it is 

 rounded. The po.storbital processes are very large, l)ut do not extend downwards to 

 join the zygomatic processes of the squamosals, as in Aptornis; in Erylhromachus 

 these processes are extremely small. The posterior wall of the orbit bears a very 

 deep and clearly defined depression for the lachrymal gland, similar to that seen in 

 Ocydromus and Cabalus, but much more strongly developed. 



The interorbital vacuity is of much the same form as in Ocydromns : just above 

 its anterior margin is the aperture of the tube in which runs the olfactory nerve; in 

 the smaller rails the channel for this nerve remains completely open. 



The posterior portion of the basitemporal platform is not prominent and inflated 

 as in the smaller forms, hut is slightly concave from side to side, and lies about 



