METRIORHTNCHUS. 163 



length as they pass back in the series ; they are compressed from above downwards, 

 terminate in an oval facet, and are directed downwards to a varying degree. The zyga- 

 pophyses are large, with flat oval surfaces. The neural spines increase in length 

 from before backwards ; they are nearly vertical and are somewhat thickened at the 

 upper end, especially towards the posterior side. 



The exact number of dorsal (text-fig. 62) and caudal (text-figs. Go, 66) vertebrae 

 is doubtful, there being much uncertainty as to the completeness of the specimens 

 preserved. The greatest number of rib-bearing dorsals preserved in any skeleton is 

 seventeen, and there were probably one or two lumbars. Fraas, in Geosaurvs, finds 

 that there were sixteen dorsals and two lumbars, a number which is adopted by 

 Arthaber for Metriorhynchus; probably there were sixteen or seventeen rib-bearing 

 dorsals and one or two lumbars. In recent Crocodiles it is usual to regard as the first 

 dorsal the first vertebra of which the rib joins the sternum, but in these fossil forms, 

 in the absence of the sternum, this distinction is not possible ; therefore the first 

 dorsal is considered to be that vertebra in which the parapophysis is wholly or partly 

 borne on the neural arch, although in modern types this does not occur till some 

 distance back in the dorsal series. In fact, it is not possible to distinguish clearly the 

 different regions of the presacral portion of the column. The centrum of the first 

 dorsal (according to the above definition) differs little in form from that of the last 

 cervical, except that the parapopbyses have passed up on to the neural arch. Further 

 back in the series the centra become more elongated and cylindrical, and are more 

 constricted in the middle. In uncrushed specimens there is no keel on the mid-ventral 

 line, but in compressed specimens such a ridge is frequently present. The articular 

 ends are gently concave and the vertical diameter is a little greater than the transverse ; 

 there is a slight flattening beneath the neural canal. 



The neural arch unites with the centrum by broad pedicles, which, however, in the 

 hinder part of the series at least, do not extend to the posterior end of the centrum. 

 In the first dorsal the pedicle of the arch bears near its antero-ventral angle a blunt 

 parapophysial prominence, above and, for the most part, behind which arises the 

 strong transverse process, the outer end of which bears the surface for the tubercle of 

 the rib. In this vertebra the transverse process is little compressed from above 

 downwards, but further back in the series this compression increases and the process 

 widens, its anterior edge being thin, the posterior thickened and grooved, at least near 

 the base. The parapophysis in a few of the dorsals behind the first may form a distinct 

 rounded process, rising from the anterior border of the base of the transverse process 

 {e. g., in the small many-toothed species, M. Iceve), or it may merely form a step-like 

 prominence on the anterior border of the process. As it is followed backwards in the 

 series, this step does not appear to pass outwards towards the diapophysial surface of 

 the end of the process as in Steneosaurus, but, on the other hand, the diapophysial 

 surface, by the shortening of the process, to some extent approaches the parapophysis. 



t2 



