515 



jaw are maintained to full growth, the remark on the minor depth of the horizontal 

 ramus (p. 271) may not be applicable to a specimen less " mutilated and abraded" than 

 is the type of the species (Plate XLIL). 



The nototherian modifications of the lower border of the hinder half of the ramus, 

 exemplified by the anterior (d) and posterior (a) inflections in fig. 4, are well shown in 

 the subject of the present description. 



PLATE CXXV. 



Nototherium inerme. 



Fig. 1. Outer side view of a mutilated right mandibular ramus and teeth. 

 Fig. 2. Under surface of the same. 

 Fig. 3. Upper surface of the same. 

 Fig. 4. Inner side view of the same. 



All the figures are of the natural size. 



§ 3. Atlas Vertebra. — In the description of the fragment of an atlas of a large 

 mammal with marsupial characters of that bone, I expressed [ante, p. 209) a doubt 

 whether the ring of the vertebra " had been completed below by bone ;" but by reason 

 of other demonstrable modifications traceable in the fragment, I was led to refer it to a 

 Diprotodon australis, and to figure it in Plate XXVIII. fig. 2. 



I have since received an almost entire atlas confirming that determination, and one, 

 also nearly entire, of a Nototherium (Plate CXXVL), both of which demonstrate the 

 fact that, as in some species of JIacropus (Plate LXVII. figs. 1-3), of Phascolomys 

 (Plate XOVIII. figs. 3 & 4), and in certain other Marsupials*, that vertebra was not 

 completed below by bone. 



The atlas in the two cited genera of large extinct Marsupials presents so close a 

 similarity that a description, with references to the figures of the smaller one (in Plate 

 CXXVL), will suffice to indicate their leading characters. 



Each neurapophysis, after developing the large articular surfaces (z, z, fig. 1 ) for the 

 occipital condyle in front, and the prezygapophysis (z\ z', fig. 2) for the axis vertebra 

 behind, terminates below these surfaces in a short, thick, rough tuberosity (ib. fig. 1, mp). 

 The pair leave an interspace of 1 inch 7 J lines, which received the permanently detached 

 centrum, here, as usual in Hematotherms, confluent, as an " odontoid process," with the 

 succeeding centrum (Plate XXVIII. fig. 1, ca). 



I know not at present any mammal showing this non-completion by bone of the ring 

 of the atlas save in the limits of the Marsupial order. 



The unossified interval (fig. 2, h) is relatively greater in Nototherium than in lJipro- 



* Art. Marsupialia, Cyclopaedia of Anatomy, vol. iii. 8vo, 1841, p. 277, fig. 98 (Perameles), fig. 99 (Pfuis- 

 colaretos). I did not then foresee the applicability of the character, here first defined, to a determination of the 

 marsupial nature of a mammal little less than an elephant. 



55 



