594 Mn. It. Ti. wniTEnousE on the [-^^ir. 5, 



in tlic figure) ; proceeding back wards, as they become smaller, 

 three .se<^iiients are shown, then two, and finally oidy one can be 

 traced. This segmentation is due, I believe, to the formation of the 

 hvpural by the coalescence of radial with luenial arch ; anteriorly, 

 the three segments, the pi-oximal (;).), median (?n.), and distal (rf.), 

 of the radials are present, the distal still remaining free ; from 

 before, backwards, the fusion becomes more peifect and finally 

 all trace of it is lost This composite nature is a feature which 

 substantiates the claim of these hypaxial supports to the 

 title hypural. The dermotrichia constituting the ventral lobe 

 considei'ably overlap their endoskeletal supports. Dor^ally, the 

 arches {n.a.) and ladials {d.c.r.) I'cmain distinct throughout the 

 tin and ai-e less regtdar than the hypaxial elements. The dorsal 

 radials are almost hidden from view by the forking proximal ends 

 (removed in the figure) of thestout fulcra which fiinge the margin 

 of the doi-sal lobe. The whole of the elements at the extremity of 

 the tail are hidden by the dermotrichia. 



PoLYODox SPATULA. (Chondrostei,Polyodontida?.) (Plate XLVII. 

 fig. 3.) 



Externally the tail of Pohjodon is very deceptive ; its symmetry 

 of form is most marked, and it resembles in outline the deeply 

 cleft tail of some Teleosts. It is well-known, however, that on 

 dissection it proves to be an excessively heterocercal type, the 

 caudal axis being directed upwards a,nd extending to the extremity 

 of the dorsal lobe. No centra are formed. Almost throughout 

 the tail-fin the neural arches (n.a.) are single elements, and are 

 somewhat irregidar in size, pai'ticularly towards the extremity of 

 the axis. Kesting upon these arches are the radials ((/.c.r.) ; the 

 first four or five are slender in build, but almost throughout the 

 fin they resemble the neural aiches themselves. They, too, are 

 very iri'egular both in disposition and size, and here and there 

 they show signs of having been more numerous formerly. The 

 various elements at the extremity of the axis are difficult to dis- 

 tinguish, owing to their small size and tendency to fusion ; the 

 same difiiculty is experienced ventrally in this region. The radials 

 nvv the immediate suj)ports of the fulci-a (_/.). 



Turning to the ventral elements, a very regular series presents 

 it.self, in striking contrast to the dorsal side : theie is a regular 

 series of hyi)urals, very large at the root of the fin where the large 

 dermotrichia of the ventral lobe are supported. As in Acipeiiser, 

 very many of the.se hypurals show di.stinct segmentation, thus 

 giving evidence of being comjjosed of four successive elements, the 

 distal of which, in the ca.se of the first seven sxipports, is distinct 

 ami separate. Tiie interpretation seems to be that each hy{)ural 

 is composed of ha-mal aix-h and i-adial fused, the tri-segnuMital 

 nature of the latter Ijeing retained in several cases. Naturally the 

 l)e.st examples of this are given by the anterior hypurals, where 

 the supporting elements are larger. The tri-segmental arrange- 

 ment gradually gives way to a lii- and finally a uni segmental 



