292 J, K. Thacher — Median and Paired J^)'ns. 



nale in Cerafodus* is also, without the slightest doubt, homologous 

 with the chord which lies entirely above the neural arch in the Elas- 

 mobranchs. The peculiar fibrous character is almost exactly the 

 same in each. We have seen that this cord in the sharks is in all 

 probability homologous with the fibrous tatty ridge-pole of the 

 neural canal in Petroniyzon. Thus while in the latter the arches of 

 one side and the other are entirely separate, in the sharks they have 

 spread beneath the ligament so as to meet, and in the higher 

 Gnathostomes they have also joined above it, or nearly joined, for 

 the origin of a is still to be discussed. 



There are two possibilities with regard to the neui-al spine a. Either 

 it is formed by the union of a median fin-ray with the neural arches, 

 the ray thus constituting the keystone of the arch, or else by the 

 union of the neural rods from each side and their prolongation dorsad. 



But the junction between <i and h is quite close ; the neural spines 

 correspond in number and j)osition with the lateral parts of the arch ; 

 while fig. 61 shows conclusively the absolute independence of neural 

 spines and primordial median fin-rays. 



The second of the two possibilities is then the true one. Thus 

 neither are median fin-rays derived from neural spines, nor neural 

 spines, where they occur, from primordial fin-rays. 



But the cartilaginous supports of the median told in the Dipnoans 

 are very long and segmented. They are simply elongated neural 

 spines and are not primordial fin-rays in any homological sense. 

 If they were formed by the reduction in number of the primordial 

 fin-rays and their coalescence with the neural sjjines it is impossible 

 that we should not have here and there an extra one, and some evi- 

 dence in the case of others of such a junction. But there is nothing 

 of the kind, either in the descriptions of Gtinther in the case of Gera- 

 toduSjj or in those of Owen J and Peters§ in that of Protopterus annec- 

 tens. or in those of Bischofi"|| in that of Lepidosiren paradoxa. Gtln- 



* Griinther's Description of Ceraiodus, Pliil. Trans., vol. elxi, pt. ii, PI. XXXVIII, 

 Pigs. 3-9, e. 



f Phil. Trans., vol. clxi, pt. ii, 1871. G-iinther, Description of Oeratodu.s. In Giinther's 

 fig. 2, PI. XXX, the proximal joint of the 14tli neural spine seems to bear two 

 ' ' interneurals," one orad of the other. But as no notice is taken in the text of this, 

 which would be a very noteworthy fact, if it were fact, and as the description of these 

 parts there given is such as would demand a notice of this exception, it is evident 

 that it must be an inaccuracy in the figure. 



:|: Trans. Linnean Soc, vol. xviii, pt. iii. Owen, Description of Lepidosiren annedens. 



§ Miiller's Archiv., 1845. Peters, Ueber einen dem Lepidosiren annectens verwandten, 

 Fiach von Quellimane. 



II Ann. Sc. Nat., xiv, 1840. Bischoif, Sur le Lepidosiren iiaradoxa. 



