( "9 ) 



The posterior region is unlike that of any other ralline sternum I have examined, 

 and is subject to consideralile individual variation. Instead of the short median and 

 deep lateral notches, bounded externally by long slender lateral processes which 

 project far beyond the median ones, there are here a median and two pairs of shallow- 

 lateral notches, bordered by three pairs of short processes. ]5otli the notches and 

 the processes are very irregular and inconstant in form ; in some specimens the 

 external notch is converted into a fenestra, and the inner pair of processes are almost 

 obsolete. 



The coracoid is very similar to that of Ocydromits, but is relatively shorter, and 

 is thicker in proportion to its length ; it also has a much smaller, indeed almost 

 rudimeutarv procoracoid process, while the acrocoracoid is larger and more swollen ; 

 its lower extremity is wider, and it has a fairly prominent lateral (hyosternalj process. 

 There is a supra-coracoid (subclavicular) foramen, below which the inner border is 

 sharp and somewhat irregular; the inner angle is produced into a sharp angular 

 process. The posterior surface of the bone is hollowed by an extensive sterno-coracoid 

 impression. 



The dimensions of the coracoids associated with the figured sternum are : — 



Length, 40 mm.; greatest breadth, 16 mm. ; least ditto, 5 mm. 



Expressed, according to Fiirbringer's method, in terms of the average length of 

 a dorsal vertebra (11 mm.), its length is roughly 3-5, its breadth 1-5. 



The scapida differs only from that of Ocydromus in being somewhat less curved 

 in having a somewhat more expanded distal end and a more massive articular end, 

 with a smaller acromium process. Its length, in the same individual as that to which 

 the coracoids above described belong, is 65 mm., its greatest breadth (at the 

 articular end) 12 mm.; or, expressed in terms of the length of a dorsal vertebra, 

 6 and 1 nearly. 



As already mentioned, the scapula and coracoid above described Ijelong to the 

 same individual, and the most interesting point about them is tliat, when they ai-e 

 placed in their natural position relatively to one another as nearly as their articular 

 surfaces allow one to judge, they make an angle of nearly 130" with one another, 

 a much larger one than occurs in any other carinate liird except perhaps Aptornia. 

 In most C'ariimt(te the angle is less than a right angle, but in Ocydromus and Bidux 

 rather greater. The coraeo-scapular angle of Hesperornis is aliout the same as in 

 Diaphorapteryx. Much importance has Iseen attached to the form of this angle, 

 since in all the Ratitae it is very obtuse (from 130 — 160 degrees), while in the 

 flying birds it is, as already stated, less than a right angle ; but iu such forms as 

 Dutphorapieryx, Ocydronias, and Didus, in which the power of flight has been lost, 

 the aceompan3-ing increase in this angle appears to be merely the mechanical result 

 of the reduction in length of the coracoid or of the increase in the size of the thorax, 

 or of both. It is possible that, owing to the difficulty of determining tlie precise 

 position of the bones with regard to one another, the size of the angle may be found 

 to be somewhat overestimated; but I do not fliiuk this is the case to any serious 

 extent. 



The humerus has the characteristic twisted appearance of tiiat of the 

 Ocydromiue rails. It is about eLpial in length to tliat of Aphanapteryx hroechi, 

 recently figured by Xewtou & Gadow,* but it is considerably stouter and lias more 

 massive articular ends ; the shaft is also less curved, and the median tubercle is 

 larger. A blunt ectepicondylar process projects beyond tlie border of the bone. In 



• Z.IW ciL, ?1. XXXV. Figs. 17, IS. 



