DEVELOPMENT OF THE SKELETON OF THE TTJATARA. 15 



them, as in Tupinambis 1 . This notwithstanding, since the anterior element of the whole 

 series (that lying between the skull and first vertebra) is the first intercentrum, pre- 

 cision and uniformity are gained by enumerating it in correlation with the centrum 

 behind it, and so on for the whole series which follow as has been done, for example 

 by Gadow, for the chevrons, though not with that uniformity which is desirable. The 

 method mostly adopted, especially in illustrations, of associating with the numeral 

 indicative of a given chevron that denoting the vertebra in front of it, as has been done 

 with descriptions of the skeleton of the adult Sphenodon from Giinther to Osawa, leads 

 to confusion and must be given up. 



Passing now to the intercentra, a marked advance is seen to have taken place in 

 those of the anterior twelve body-segments and the tail. At Stage P the intercentral 

 bodies are everywhere paired. At Q, fusion in the mid- ventral line has come about in the 

 anterior region (rf. PL I. fig. 8, i.p'.J. In the case of the first four bodies, the contour 

 of the fused product gives no indication of its paired origin, but from the fifth to about 

 the twelfth (as determined by very careful examination of serial longitudinal sections) 

 a median depression remains. These facts are embodied in column Q of our table on 

 p. 27, and perusal of them would seem to indicate that fusion of the paired intercentra 

 takes place antero-posteriorly. 



At Stage P, owing to the absence of demarcation between the cranial and vertebral 

 portions of the differentiating tissues, the enumeration of the centra had to be 

 determined by reference to the sacral region, i. e. the segments had to be worked out 

 in postero-anterior succession. The numerous practical difficulties arising from this 

 necessity, leave us in doubt whether the first and second intercentra are or are not 

 represented at this stage. Hence the 1 in our table, column P (p. 27). This con- 

 sideration would seem to throw some doubt upon the paired origin of these first 

 two intercentra; but, judging from the fact that from segments 12 to 5 the unpaired 

 intercentra at Stage Q still show traces of paired origin, and that in segments 4 and 3 

 these elements are proved to be beyond doubt originally paired, we deem it fair to 

 conclude that the same conditions apply to segments 2 and 1. 



It may be remarked that at this stage (Q) the union of the first centrum with 

 adjacent parts is effected. While it fuses (fig. 8 a) completely with the second inter- 

 centrum (i j/.), it does so only peripherally with the second centrum, and that so as to 

 enclose at* a central mass of undifferentiated tissue. 



The neural arches have at this stage completely united in the dorsal middle line and 

 are fully chondrified. 



Passing to the caudal region, the intercentra of its fourth and succeeding vertebral 

 segments, by an astonishing rapidity of growth, have acquired a characteristic elongation, 

 each extending downwardly into (PL I. fig. 7, i. p.' 1 ) the substance of the tail, but 

 without meeting its fellow. Although they are chondrified, and at this stage differ- 

 entiated in advance of their homologues the paired intercentra in front, their cells 



1 Cf. Boulenger, G. A. : op. cit. p. 114. 



