316 Proceedings of the Royal Physical Society. 
remains of the trabecule are seen as a thin black streak on 
each side. 
IT have never, even to this day, seen the dorsal surface of 
the head of Palwospondylus; but to judge from sundry partial 
impressions, it is clear that in this anterior or trabeculo- 
palatine portion a median or cranial part was marked off 
from the palatine by a ridge on each side, and that this cranial 
part was, behind the nasal ring, widely open above, as in the 
Lamprey. The lateral or palatal fenestrated portion was 
distinctly concave above, and here the eye must pretty 
certainly have been placed. 
I have scarcely anything to add regarding the posterior 
part of the skull, which is presumably composed of the 
combined parachordals and ear capsules. Constantly, how- 
ever, in eroded heads, among its remains may be distinguished 
two oblong pieces converging forwards at an angle near the 
junction of the anterior and posterior parts of the cranium, 
as is seen in Figs. 1 and 2. No indication of such a con- 
figuration can, however, be seen in the head shown in Fig. 5, 
where the surface seems remarkably entire and uninjured. 
With regard to the problematical pieces x, their morpho- 
logical signification seems as obscure as before. They could 
not have been external plates, at all events, and may possibly 
be a remnant of an internal branchial support. 
Nothing new can be said of the rest of the skeleton as far 
as the tail-fin; but here a previously unobserved fact of con- 
siderable importance must be noted. The rods supporting 
this fin I have considered as neural and hemal spines, as they 
are apparently continuous with the neural and hemal arches, 
but it turns out that in none of my previous specimens were 
their extremities complete. A specimen, which I obtained last 
autumn from Mr Donald Calder of Thurso, shows that these 
rods or spines were considerably longer than they had been 
represented in any of my figures, and consequently that the 
fin was so much deeper. Buta point of greater importance 
than their length, is the fact that at least the neural spines 
are seen to dichotomise twice, just as in the case of the caudal 
spinous processes, the so-called fin-rays of the Lamprey. In 
Fig. 6 I have represented a portion of the caudal extremity 
