KIMMERIDGIAN DINOSAURS. 571 



phery slightly concave toward the palmar border {/>), which is broken away, the articular 

 surface being continued a short way upon the ulnar {u) side of the shaft for junction with 

 the second metacarpal. 



The articvilar surface is pitted with small deepish depressions, as in most great 

 Saurians, where the joint surfaces seem to have been more syndesraosal than synovial. 

 The transverse and ancono-thenal diameters of the proximal surface are equal, each being 

 3 inches G lines ; but, had the ulnar border been entire, the transverse diameter would 

 have somewhat exceeded the other. 



The short thick shaft of this bone is three-sided ; one side extends obliquely from 

 the ancono-ulnar (fig. 3, an) angle to the radio-palmar (rp) angle, with a transverse 

 convexity ; the second, or palmar, side (fig. 4, p) is less convex across ; the third, or 

 ulnar side, is flat across at the middle part, and somewhat concave near the two expanded 

 ends of the bone. All these surfaces are concave lengthwise, the palmar one least so ; 

 but the proximal half of this (fig. 4, P> l>') has been crushed. 



The distal articular expansion (fig. 6), almost flat transversely at its anconal part (a), 

 begins to be concave at the middle of the distal surface (J), and this deepening to the 

 palmar one [p) divides the joint there into a pair of convex trochlear condyles. The 

 radial (,-, fig. G) of these, when entire, would have been the most prominent of the two. 



The metacarpal (PI. 71, figs. 1 and 2) which supported the fourth digit has a 

 proximal articular surface of a more definite triangular figure (PI. G7, fig. 5) ; the 

 anconal border (a) being the longest, and the angle between the radial [r) and ulnar {u) 

 borders being rounded off". The articular surface is continued upon both these sides of the 

 shaft, but fui'ther for the articulation with the mid-metacarpal than for that with the fifth. 



The anconal surface (PI. 71, fig. 1) of the shaft is almost flat and lies more on the 

 plane of that surface of the entire metacarpus than in the marginal metacarpal above 

 described (fig. 3). The radial and ulnar surfaces of fig. 1 converge palmadtothe narrow 

 convex palmar surface whicl) forms the rounded angle of the proximal triangular tract 

 (ib., fig. G, ur,p). Both radial and ulnar surfaces of the shaft are concave lengthwise and 

 across (ib., fig. 2, r). The transverse concavity of the distal articular surface is 

 feebly indicated, and the bifid character of the joint is scarcely marked, though fractured 

 surfaces suggest that a pair of low palmar prominences may have been broken away ; 

 but the joint is much less trochlear than in the first metacarpal (ib., fig. G). 



A metacarpal of similar type to the preceding has suffered too great mutilation of 

 both ends to serve for profitable description ; it is not a corresponding metacarpal of the 

 right fore-foot, but may be either a second or third, thougii from the slight su[)eriority of 

 length I should judge it to have been the second metacarpal of the same left fore-foot as 

 the subjects of PI. 71 belonged to. 



A metacarpal with a subtriedral shaft, aiul an oblique twist at its basal half through 

 an extension radiad of the radial angle, upon which angle the flat proximal articular 

 surface has extended for the metacarpal on that side, is evidently a fifth metacarpal bone. 



