N. ROSÉN, STUDIES ON THE PLECTOGNATHS. 7 



very much. The vertebral column is bent by the inflation. 

 In the spina bifida runs a longitudinal muscle, the contrac- 

 tion of which straightens the vertebral column when the in- 

 flation is ended. The development of this muscle — and 

 with this the length of the spina bifida — increases with the 

 power of inflating the body. 



I have indicated above that the dorsal arches with all 

 probability have been formed by the coalescence of one cra- 

 nial and one caudal portion. This opinion of mine is based 

 on the study of sections of young specimens of Spheroides, 

 measuring 18 mm. in length. In this stage the vertebra is 

 already quite distinct and ossified. It is possible, however, 

 to distinguish two centres of each dorsal hemiarch, one cra- 

 nial and one caudal. In these parts the neural arch is much 

 thicker, and lacunes and small remains of cartilage show that 

 two cartilaginous arches originally existed. These låter on 

 ossified and fused with each other. The question arises then 

 whether these portions belong to the same sclerotome or not. 

 An examination shows that the myosepts insert to the joints 

 between the vertebrae, and that intersegmental blood-vessels 

 run just behind the myosepts. The spinal nerves penetrate 

 the centrum of the neural arches, between the two original 

 portions. From all this it is evident that the dorsal arches 

 have heen formed hy the coalescence of two j^ortions, one cra- 

 nial and one caudal^ belonging to the same sclerotome. (Fig. 

 B, I). The vertebra develops in the same mode as in 3Iola. 

 The ventral arches are of great interest. Cranial and caudal 

 portions can be distinguished just as in the dorsal arches 

 (Fig. B, I). Each vertebra bears two ventral arches on each 

 side, joined with each other at their bases. They represent, 

 of course, the original cranial and caudal arches of the same 

 sclerotome. A peculiarity is that they form a haemal canal 

 (Fig. B, II), except on the first vertebra. A haemal canal 

 in the praecaudal region occurs both in some selachians and 

 in some ganoids; but I do not think that such a canal has 

 hitherto been described in any teleost. In all the forms 

 which possess a praecaudal haemal canal, this is formed by 

 haemal processes. Distinct ventral arches are present in all 

 those forms. The haemal processes are quite correctly, I 

 think, considered as separated portions of the ventral arches. 

 In Spheroides we meet with a somewhat peculiar state of 



