10 ARKIV FÖR ZOOLOGI. BAND 10. N:0 8. 



description seems to be almost exclusively based on studies 

 en Diodon. However, this may be, his paper has quite a 

 descriptive character and does not contain any genetical 

 views. Regaist writes, p. 292: »praecaudal vertebrae without 

 parapophyses». As has been shown, ventral arches (== par- 

 apophyses) are present. 



The nu7nber of vertebrae is about 20. The last caudal 

 vertebra is symmetrical. 



Diodontidae. 



Descriptions are given by Dareste, Hollard and Regan, 

 but without any genetical views, nor have I myself had an 

 opportunity of making ontogenetical studies. A spina bifida 

 is present. Ribs are absent. The number of vertebrae is 

 about 20. The last caudal vertebra is symmetrical. 



Ostraciontidae. 



The number of vertebrae is about 15. As Hollard 

 showed, the ventral arches are very small. The spinal pro- 

 cess in very short or absent. No ribs. The last caudal 

 vertebra is symmetrical. 



General Remarks on the vertebral column. 



It is impossible of course to characterize the vertebral 

 column of all the plectognathous families in relation to other 

 teleosts until more extensive genetical researches have been 

 made. There are, however, some peculiarities to which atten- 

 tion may be called. The number of vertebrae is in all groups 

 very small. It has been shown that the neural arches of Mola 

 and Spheroides have been formed by the coalescence of two 

 original portions, one cranial and one caudal, belonging to the 

 same sclerotome. In Spheroides the ventral arches have a si- 

 milar diplospondylous origin, which can be seen even on verte- 

 brae of adult speciraens. The ventral arches of Spheroides 

 form, even in the praecaudal region a haemal canal except on 

 the first vertebra. In several forms a spina bifida is present. 



