OF THE TELEOSTEAN FISH SEBASTES MARINUS. 



511 



angle of the quadrate. This supposition is strengthened by a 

 comparison with the conditions found in the chondrocranium of 

 the larval Chf.pea, in which the upper jaw also arises at a late stage 

 and grows back to join the quadrate in a slender connection. A 

 somewhat similar process occurs in Amiurus and Syngnathus, in 

 each of which the palatine arises separately. 



It appears that the quadrate, symplectic, and hyomandibular 

 have all arisen from the original temporal cartilage. In such a 

 case it might be expected that the three elements were formed 

 separately in the earliest stages, then became fused to form the 

 temporal and finally separated again. In studying the earliest 

 stages in which tlie temporal can be distinguished I have not 

 been able to trace any division. But in these early stages the 



Text-figure 9. 

 A 



,-Ep.Br 



Stage 4. Dorsal view of brancliial bars. 



cartilage is in an immature condition, and it is very hard always 

 to be certain whether it is continuous or not. It would also be 

 difficult to reconstruct the details of the process by which the 

 differentiation of the hyomandibular, symplectic, and quadrate 

 has come about. It is worth noting that in the 5*5 mm. stage the 

 hyomandibular part of the temporal is very short and that it has 

 become greatly lengthened in the intervening period. The long 

 hyomandibular is a feature correlated with the rather unusual 

 depth of the head in the adult Sebastes. 



The stylohyal is now a small cartilage articulated with the 

 hyomandibular and ceratohyal. The latter is in the form of a 

 very large plate bearing a, posterior process for the articulation 

 with the stylohyal. The anterior pai-t of the ceratohyal tends to 



