PROFESSOR OWEN ON THE GENUS DINORNIS. 161 



bases of" ' parapophyses,' the fore or preaxial articular surface (ac, fig. 20) may be 

 said to extend thereupon ; but the parapophysis, ^;s, in Dinornis maximus, may better 

 be held to spring out external to the preaxial surface than in Struthio. The outer 

 border of the prezygapophysis (fig. 20, az) now begins to rise, and gives a more inward 

 or medial aspect to the oblique articular surface. The postzygapophyses (fig. 2\,pz,iiz) 

 show a corresponding change in the contrary sense, but they do not extend postaxially 

 beyond the centrum in so great a degree as in Struthio ^. A marked difference between 

 the vertebr8e here compared is in the greater height and greater breadth of the neural 

 spine in Dinornis ; but the chief distinction is shown by the coexistence in Struthio of 

 an independent or movable pleurapophysis with the first appearance of the single and 

 simple hypapophysis. In Dinornis such condition of the hypapophysis is associated with 

 a continued confluence of the riblet, pi- In other words, the single hypapophysis marks 

 the first dorsal vertebra in Struthio ^ and the last cervical in Dinornis (fig. 18, hy). 



The character of the fifteenth cer^'ical in the series of the skeleton of D. maximus 

 is that of the sixteenth cervical in the neck-series of Button's skeleton of D. ele- 

 phanto])us, and this is followed by a seventeenth cervical, or one with anchylosed pleur- 

 apophyses, beyond which there are seven vertebrae for a dorsal series. But in both the 

 sixteenth and seventeenth cervicals the neural spine is bifid ; the ridges from the 

 hyperapophyses converge to the base of a single neural spine only in the first of the 

 series of vertebrae in which the pleurapophyses retain their independence and mobility. 

 If my series of cervicals in D. maximus be, as it seems by characters of juxtaposition, 

 the correct number, Hutton's specimen of the skeleton of D. eJephantopus has two 

 additional cervicals, in all seventeen, instead of fifteen as in Apteryx. These remarks 

 are based on a photograph of the skeleton in the Otago Museum. 



In a cervical vertebra (figures 22, 23, 24) of D. giganteus, which I regard as homo- 

 logous with the fifteenth or last cervical of i*. maxinms, the neural spine (fig. 23, ns) 

 retains, as in D. elephantopus, its bifid character, but the parial portions are relatively 

 less developed, and their connecting bar (ib. h) has begun to rise, indicating, as it were, 

 a rudiment of the single and longer neural spine in D. maximus. The hypapophysis 

 (figs. 22, 24, hy) is single, and its base is supported by the ridges from the pleur- 

 apophyses representing the ' catapophyses ' of Mivart. The posterior hypapophysial 

 tubercles (ib. ky') are better marked than in figure 1 9 [D. maximus). 



The neck-vertebrae in every species of Dinornis in which I have been able to deter- 

 mine them correspond, with unimportant modifications, with those above described and 

 figured, and in like degree differ from their homologues in Struthio. 



In the well-marked class of Vertebrates characterized by the many cervical vertebrae, 

 these, as a rule, are small ; but in Apjteryx, and especially in Dinornis, they are excep- 

 tionally large. Some of those in Dinornis maximus almost equal in size the neck- 

 vertebrae of the horse. 



' Comp. fig. 18, pz, with Mivart, loc. dt. p. 40G, fig. 35. '■ Mivart, he. cit. p. 408, fig. 40, hy. 



