Prof. R, S. Traquair — On a new Carboniferous Fish. 533 



region, and becoming gradually attenuated to a fine point in the 

 caudal, but in specimens free from distortion a very distinct upward 

 curvature of the axis of the body is seen near the tail. There being 

 no traces of vertebral bodies visible, the notochord must have been 

 persistent, and the empty space which it occupied is well seen, 

 especially in the caudal region. The neural arches and spines are 

 well ossifi.ed ; the latter are slender in form, and are long in front, 

 getting gradually shorter as we proceed backwards. The form 

 of the neural spines is most distinctly seen in the caudal region, 

 where each, springing from a slightly expanded neural arch, passes 

 very obliquely upwards and backwards, and becoming thin in the 

 middle, ends in a laterally-flattened, expanded termination, to which 

 one of the next set of elements, the dorsal fin-supports, may in 

 many instances be clearly seen to have been articulated : indeed, 

 I believe that this must always have been the case, though in the 

 present fossils these elements are often displaced from each other, as 

 in Fig. 2, PI. XIV. These succeeding elements, the dorsal " inter- 

 spinous bones," or fin-supports, well shown both in Figs. 1 and 2 

 of the plate, are likewise slender in form, expanded at both ends, 

 and pass, like the neural spines which support them, obliquely 

 upwards and backwards. They are frequently slightly curved, with 

 a forwardly directed concavity. They are rather short and delicate 

 in front, attaining their greatest length and stoutness opposite the 

 beginning of the caudal region, whence backwards they again 

 become smaller, till towards the end of the tail they can no longer 

 be distinguished from the neural spines. The ribs are, as in 

 Ph. Andersoni, very long and prominent, and are well seen in Fig. 1. 

 The haemal spines follow on the ribs as we pass from the abdominal 

 to the caudal region, and with the appended supporting bones of the 

 anal and inferior lobe of the caudal fins, agree essentially with the 

 corresponding elements on the neural side of the notochord. The 

 ribs, neural and heemal spines, and median fin-supports seem to 

 have been hollow, the central cavities, originally occupied by car- 

 tilage, being now filled with carbonate of lime. 



Fins.^—A comparison of the fi.ve specimens under examination 

 shows that the dorsal fin was, as in Ph. Andersoni and in the recent 

 Lejjidosiren, perfectly continuous with the upper lobe of the caudal. 

 In specimen No. 1 (Fig. 1, PI. XIV.) the dorsal is seen distinctly to 

 commence only three-quarters of an inch behind the head. This dorso- 

 caudal fin is low at its commencement, but gets broader posteriorly, 

 attaining its greatest breadth opposite the beginning of the caudal 

 region, after which it becomes rapidly attenuated as it proceeds to 

 the tip of the tail. On the h^mal aspect of the body there is some 

 evidence in nearly all the specimens, but especially in No. 4, of a 

 small, narrow, separate anal fin in front of the lower lobe of the 

 caudal. This lower lobe of the caudal is not so broad as the upper 

 lobe, and, proportionally, is considerably less deep than the corre- 

 sponding part in Ph. Andersoni; like the upper lobe, it becomes 

 finely attenuated posteriorly. The tail thus formed is beautifully 

 diphycercal, as seen in Fig. 2 ; there being, however, as above 



