R. C. PUNNETT 
315 
a dorsal root in the adult. A dorsal root may be exhibited by the last spino- 
occipital nerve, though in most cases it is lacking. The arrangement of the 
cartilages in this region is somewhat irregular. When the dorsal root of z is 
present the hinder part of the skull is longer than usual, and, whilst always 
passes through the skull, issues through a separate intercalary cartilage. In 
other words, where there is least condensation in this region of the skull, the 
basidorsal cartilage corresponding with nerve z is fused with the cranium, whilst 
the interdorsal remains separate and transmits zf^ if this be present. When 2'' is 
absent the interdorsal may remain as a separate cartilage, or it may fuse with 
the basidorsal next behind it. An idea of the amount of variation in this region 
may be obtained from Figs. 4 and h on Plate I. The existence of such variations 
serves to greatly strengthen the theory of vertebral condensation, though the 
point with which we are here concerned is that they are not of sufficient magni- 
tude to materially confuse the nomenclature of this region. The anterior limit 
of the whole vertebrae is capable of rigid demarcation. 
The junction of the whole and half vertebrae is in most cases easily de- 
termined, and the two most rostral of the half vertebrae each show a basi- 
dorsal and an interdorsal cartilage on either side. Over the 1st, 3rd, 5th,... 
half vertebrae there are no nerve foramina ; they occur only over the even- 
numbered half vertebrae. From this normal condition variations of two sorts 
occur. There may be only two pairs of cartilages over the first two half vertebral 
centra taken together (Fig. 6 a, Plate I), a very large basidorsal and a small 
interdorsal, which latter, as usual, affords a passage for the dorsal root. In fact, 
we have here a whole vertebra in which the centrum has become divided, 
without division of the cartilages forming the neural arch. For purposes of this 
paper, such a condition has to be reckoned as two half vertebrae. The second 
form of variation is an incomplete division of the centrum of the last whole 
vertebra. The division may be incipient only, or it may be fairly well marked 
on one side whilst absent on the other. In such cases, which are comparatively 
rare, the centrum exhibiting such incomplete division has been reckoned as that 
of the last whole vertebra. 
2. Rolf vertebrae. The rostral limit of the half vertebrae has already been 
discussed. The caudal limit in Spinax is quite well defined, and little difficulty 
was experienced in deciding the number when the cleaned skeleton of the tail-fin 
was examined with a simple lens, though the higher powers of a dissecting 
microscope were used in the case of the embryos. It has been assumed that all 
the vertebrae, from the junction with the whole vertebrae to the tip of the tail, 
fall into the same category, viz. that of half vertebrae. 
Ridewood ('99, p. 55) has recently disputed this on the ground that towards 
the caudal end the myotomes are twice as numerous as in the rostral part of the 
series. As this point is of great importance for the present paper I may here 
give the reasons which have led me to dissent from the view taken by Ridewood. 
40—2 
