52 MR W. E. AGAR ON THE DEVELOPMENT OF THE 



prechondral quadrate. As this is a very evanescent structure, it will be best to follow 

 its fate at once. At this stage (about 31) there is no bone present either in the tooth 

 or in the connective tissue (fig. 10). In stage 31 +bone has been deposited in the tooth 

 rudiment and spreads a short distance back along the inner side of the connective tissue 

 strand. [Graham Kerr # has described the development of the teeth in Lepidosiren, 

 and showed that they develop from a continuous rudiment, and are not formed by the 

 fusion of separate denticles as in Ceratoclus (Semon)]. In stage 34 the bone has grown 

 back from the tooth along the connective tissue to the inner side of the quadrate, which it 

 overlaps (fig. 11). The above series refers to Protopterus, but would apply almost equally 

 well to Lepidosiren. In the earliest larva of this genus, however, in which ossification 

 was found, this was not taking place in the tooth germ but dorsal to this, in the part of 

 the bone which in the adult forms the ascending process (cf. fig. 16, PI. III.). The early 

 appearance of calcareous matter in this region may probably be ascribed to the fact that 

 this part of the bone is the thickest in the adult. Almost simultaneously with its 

 appearance the bony trabeculae in the tooth are formed, and become continuous with it, 

 and the whole bone begins to grow back. It is significant that the connective tissue is 

 not merely ossified, but that the bone grows back from the tooth along its inner side 

 (fig. 12). This is also shown by the fact that whereas in stage 31 the connective tissue 

 strand passes directly into the anterior face of the quadrate, the bone when formed, 

 that is, about stage 32 + , overlaps this on its inner side and is only loosely attached to 

 it here (cf. figs. 10, 11). The connective tissue never shows any sign of chondrification, 

 so if it represents the palato-pterygoid cartilage, it must be in a very vestigial condi- 

 tion. It is replaced by bone before the quadrate itself is out of the prochondral stage. 



This bone shows no indication in its development of being composed of separate 

 palatine and pterygoid elements, although, according to Gt)NTHER,t the suture between 

 the two constituents is traceable in the adult Ceratodus. 



At a slightly later stage than that figured in figs. 7 and 13, the hinder parts of 

 the skull basis appear in the form of a pair of neural arches, which will subsequently 

 form the occipital arches. Their bases are prolonged forward for a short distance on 

 each side of the notochord as the occipital plates. 



The only other changes to be noticed are the appearance of the palato-pterygoid 

 bone (just described), the parasphenoid, and the splenial. The parasphenoid takes the 

 form of two longitudinally running bones, stretching from the region of the attachment 

 of the quadrate to some way in front of the first neural arch (future occipital arch). 

 The bones are widely separated from each other by the hypophysis, which is only just 

 losing contact with the pharynx. In front they are flattened horizontally, but further 

 back take the form of almost circular rods. The parasphenoid is paired at first in 

 Protopterus also. 



The mandibular tooth is now forming, and a short flat process is growing back from 

 its base along the inner side of Meckel's cartilage — the rudiment of the splenial. 



* Quart. Journ. Micr. Sci., vol. xlvi. t Phil. Trans. Roy. Soc, 1871. 



