Mr. Owen on the Thylacotherium. 55 



nearly a right angle from the rougher articular surface of the elongated symphysis. 

 It may be supposed that this symphysial articular surface (PL V. fig. \, e), which 

 at once determines the side of the jaw, might be obscured in a plaster-cast, but 

 there is no possibility of mistaking it in the fossil itself ; it is long and narrow, and 

 is continued forwards in the same line with the gently convex inferior margin of 

 the jaw, which thus tapers gradually to a pointed anterior extremity, precisely 

 as in the jaws of the Didelphys, as well as in other Insectivora, both of the marsu- 

 pial and placental series. Its lower margin presents a small but pretty deep notch, 

 which possesses every appearance of a natural structure ; and a corresponding but 

 shallower notch is present in the same part of the jaw of the Myrmecobius (PI. V. 

 fig. 2,/). In the relative length of the symphysis, its form and position, the jaw 

 of the Thylacotherium corresponds with that of the Didelphys, Myrmecobius, and 

 Gymnurus. A greater proportion of the convex articular condyle is preserved in 

 this than in the foregoing specimen, and it projects backward to a greater extent. 

 The precise contour of the coronoid process is not so neatly defined in this as in 

 the first specimen of Thylacotherium, but sufficient remains to show that it had the 

 same height and width. 



The exposed surface of the coronoid process is slightly convex : had it been the 

 outer side, as M. de Blainville asserts, it would have been concave. The surface 

 of the ascending ramus of the jaw is entire above the angle, whence we may con- 

 clude, that if the process from the latter part had been continued directly back- 

 wards, it would also have been entire ; but the extremity of the angular process is 

 broken off, proving it to have originally inclined inwards, or towards the observer: as, 

 however, the greater part of the angle (c) is entire, it could not have been inflected 

 to the same extent as in the Didelphys, Dasyurus, or the Phascolotherium, next to be 

 described. A groove is extended from the lower end of the articular condyle for- 

 ward to the orifice of the canal for the dental artery, where it divides ; the upper 

 branch terminates in the dental orifice ; the lower and larger division {d) is continued 

 forward near the lower margin of the jaw, and is gradually filled up half way to- 

 wards the symphysis : — no one could mistake this smooth vascular groove for an 

 articular fissure. There is a broader and shorter groove in the corresponding part 

 of the jaw of the Myrmecobius (PI. V. fig. 2, d), and a narrower groove in that of the 

 Wombat. The alveolar wall of the posterior grinders makes a convex projection, 

 characteristic of the inner surface of the ramus of the lower jaw. The posterior 

 grinder in the present jaw (PL V. fig. I) is fortunately more complete than in the 

 first example, and shows a small, middle, internal cusp, with part of a large external 

 cusp, both projecting from the crown of the tooth in nearly the same transverse 

 line. The enamel covering the internal cusp, which is vertically fractured, is beau- 

 tifully distinct from the ivory, and considerably thicker in proportion to the size 



