120 PEOFESSOE OWEN ON THE GENUS DINOENIS. 



enables me to refer with confidence the sternum to that species, most probably to the 

 same individual as the pelvis; and the comparison of the sternum with the femur 

 described and figured p. 371, pi. 43. figs. 5, 6, 7, Zool. Trans, vol. vii, gives the fol- 

 lowing proportions of length : — sternum 6 inches 6 lines, femur 7 inches 6 lines. The 

 length of the sternum is here taken from the fore border of the coracoid groove to the 

 tapering hind end, which is all but entire. 



The inferiority of size of the sternum figured in vol. iv. pi. 4. figs. 5-8 to the sternum 

 of Aptornis defossor is greater than that of the femur of Aptornis otidiformis as com- 

 pared with that of Apt. defossor. But the relative size of the smaller sternum to the 

 femur of Aptornis otidiformis is more like the relative size of the sternum of Aptornis 

 defossor to the femur of that species than is the relative size of the smaller sternum to 

 the femur of Notornis. Seeing therefore, in the two existing Eallines here compared, 

 that as the wings decrease and the legs increase in size the sternum becomes shorter in 

 proportion to the femur, I am inclined to believe that the smaller stenium (Trans. Zool. 

 Soc. vol. iv. pi. 4. figs. 5-8) has belonged to Aptornis otidiformis rather than to Notornis. 

 This view derives further support from the fact that, with the decreasing relative size 

 to the femur, there is a progressive simplification of the sternum in the recent Rallines 

 {Tribonyx, Ocydromus), by which the still more simple type of the bone in Aptornis 

 defossor is approached. The slender lateral processes, for example, in Tribonyx ven- 

 fralis come off immediately behind the costal border, diverge and extend backward 

 beyond the body of the bone, with slightly expanded terminations. In Ocydromus 

 australis the slender lateral processes come oflf at some distance from the costal borders, 

 near the hind part of the body of the sternum, are consequently much shorter, and 

 have no terminal expansion. The keel, which is well developed in Trihonyx ventralis, 

 is much reduced in Ocydromus australis^; it is almost obsolete in Aptornis defossor. 



As the wings of Notornis are relatively less than those of the Weka, and the legs 

 relatively larger, it is not likely that the sternum would bear the same proportions of 

 length to the femur as the sternum figured in torn. cit. pi. 4. figs. 5-8 bears to the 

 femur figured in torn. cit. pi. 2. fig. 3. Whether, however, the sternum of the smaller 

 species of Aptornis would be shorter in proportion to the femur than it is in Aptornis 

 defossor may be questioned. 



Demonstration on this point still waits the acquisition, so long desiderated, of the 

 entire skeleton of Notornis. 



This, at least, is certain, that the larger sternum, belonging to Aptornis defossor, 

 repeats all the generic characters of that of which the reference oscillates between 

 Aptornis otidiformis and Notornis mantelli. 



As in it, the present larger sternum is of an elongate triangular form, the base 

 anterior and concave, the curve of the emargination being interrupted by the pair of 



' Descriptive Catalogue of the Osteological Series, Mus. Coll. of Surgeons, 4to, 1853, p. 238. no. 1280 

 {Brachijpteryx). 



