574 
ME. W. K. PAEKEE ON THE STEUCTUEE AND DEVELOPMENT 
In the third stage of Siredon (Plate 23. tigs. 1, 2, tr .) the trabeculae wer e joarachordally 
related to the notochord, taking in full half of its cranial part ; but in front they still 
only reached to the vomers, not up, even, to the nasal sacs. 
In the next stage (Plate 22. tigs. 4, 5) the parachordals are well developed and quite 
distinct ; the trabeculae, also, have grown forward into the frontal wall ; not, however, 
by the direct elongation of the primary rods, which have now grown only to the middle 
of the nasal sacs ; but a large flat transverse “ internasal plate,” ending in lunate horns, 
finishes the skull in front. 
There is evidently some difference in the method in which the highly modified axial 
skeleton ends, even among the Caducibranchiate Amphibia ; for in Seironota (Plate 29. 
figs. 1 & 2. tr., c.tr.) the paired rods grow directly forward, only gently lessening in 
size, right into the frontal wall; there they send out a small facial lobe, the trabecular 
cornu. 
At this stage the internasal plate has no existence ; it is formed afterwards by a 
commissural growth bridging over the space between the two rods. 
Already the hind part of each trabecula has coalesced with its counterpart of the 
other side; so that, now, the point of the notochord reaches only halfway to the 
pituitary space ; this state of things corresponds with that of the fifth stage of the 
Axolotl (Plate 22. figs. 6, 7). The hind part of the trabeculae is flat; each, being thin, 
bends outward and then turns forward, thickening at the side, and gently lessening 
in width. 
The arcuation of each bar is not great ; and the two are some distance apart, even 
close to the frontal wall. The hinder half of each has arisen into a sphenoidal crest, 
which curves inwards, above, fitting itself to the dura mater. There is just a perceptible 
groove between the rod and its cornu in front ; for the process is thick and is hooked 
outward and backward. 
Instead of the proper parachordals, coming up close to the notochordal part of the 
trabeculae, as in the fourth stage of Siredon (Plate 22. figs. 4, 5), there is an ample space 
running athwart the floor of the skull obliquely between them, entirely composed of 
membrane. These spaces are two thirds the size of the parachordals or moieties of the 
investing mass ( iv .). 
Each moiety is concentric in shape, turning its concavity to the ear-capsule ; but the 
opposite side is scooped, for there is a rudiment now of the ascending part — the occipital 
wall. 
Below -(fig. 2), the halves of the basal plate run somewhat under the notochord, and 
closely embrace it above (fig. 1) ; thus the axial rod can be fully seen only on this 
aspect. This rod is composed of two parts, that are more sharply defined than in any 
other species I have examined as yet. 
The foremost part of the notochord is a high, blunt-topped cone ; the second is 
roughly hourglass-shaped, that is to say, it is exactly like the notochordal axis of the 
succeeding vertebrae, whose arches are formed of a pair of cartilages, manifestly the serial 
homoloques of the parachordals that form the occipital ring. 
