MORPHOLOGY OF CYPRINODONT FISHES. 21 



The approximate symmetiy of the priapial ribs and cleithi-a in 

 Phcdlostethas, as compared with their marked asymmetry in Neo- 

 stethus, is no doubt due to the symmetrical attachment of the 

 priapium in the former (text-tig. 15), and its asymmetrical 

 attachment, owing to the great development of the glandular 

 groove, in the latter (text-fig. 4). 



With the presence of a toxactinium as a free appendage may 

 be correlated the freedom of the anterior part of the priapium in 

 Phallostethits, whereas in Neosteilius it is attached to the isthmus 

 right to the end. It seems probable that the toxactinium may 

 replace functionally the anterior part of the ctenactinium of 

 Neosteilius, and that the shortness of the ctenactinium of Phallo- 

 stethus may be connected with this. 



In both genera the ureter enters the priapium and runs down- 

 wards and across to the proctal side behind the intestine ; in 

 PhuUostethus it opens to the exterior just behind the anus, but 

 an Xeostethus, owing to the shortness of the free posterior part of 

 the priapium and the great size of the terminal coil of the vas 

 deferens, it finds the latter in its way and has to surmount it 

 before reaching the surface at some distance from the anus. 



In PhaUostethus, as compared with JSfeostethus, the great length 

 and coiling of the vas deferens, the separation of the first pair of 

 ribs from the vertebral column, and perhaps the presence of a 

 toxactinium, may be features of specialization ; in Neostethus the 

 development of the glandular gi^oove and the structures associated 

 with it may be similarly regarded. 



When I first described PhaUostethus I sviggested that the axial 

 bone of the priapium might be pelvic and the ctenactinium and 

 toxactinium fin-rays. This interpretation seems much less likely 

 to be correct when the structure of JVeostethus is considered also. 

 The antepleural cartilage of Xeostethas has developed in PhaUo- 

 stethus into a long bone with a cartilaginous core ; conversely, 

 the presence in Xeostethus of pulvinular and infrasulcar bones, 

 absent from PhaUostethus, seems to show that the skeletal ele- 

 ments of the priapium develop when and where they are wanted, 

 and are not to be homologized with any other parts of the 

 skeleton. Moreover, the pulvinular appendage of Neostethus 

 seems to represent a stage of development intermediate between 

 a simple dermal papilla and movable bony appendages, such as 

 the ctenactinium and toxactinium, and suggests that these may 

 have originated as outgrowths whose skeleton changed from con- 

 nective tissue to cartilage and then to bone as they grew longer 

 and formed more definite proximal articulations with the axial 

 bone. Their development in Neostethus hicornis supports this 

 view. 



In fact, the whole priapium seems to be an entirely new forma- 

 tion ; its appendages, bones, muscles, and glands are not to be 

 homologized with any structures found in the female fish or in 

 'Other Cyprinodonts. 



