650 
MR. W. K. PARKER ON THE STRUCTURE AND 
these tracts do not reach to the notochordal apex, which is embraced by the apices of 
the trabeculae. All trace of the composite nature of the cartilage there, is, however, 
obliterated, and from the “ atlas ” to the front wall of the face all is one continuous 
sheet of cartilage. The large primary, pituitary space (Plate 55. fig. 1) is, indeed, filled 
in here ; but at the middle a faint line of imperfect chondrification can be seen, and 
thus the distance from this line to the small crescentic “subocular fenestra” ( s.o.f .) 
gives the breadth of the trabecular band ( tr .). This breadth is doubled in the inter- 
nasal ( i.n.l .) region, for here the converging trabeculge have coalesced, back to back. 
The trabeculae, as in Dactyletlira, do not end in the front wall of the face ; their 
coalesced internasal portion ends there in a transverse, slightly emarginate plate ; but 
externally the two bars are free again, and those distal, or “hypo-trabecular” continuations 
of the trabeculae are the “ recurrent cornua ” ( c.tr .). These cornua are much more 
suddenly bent backwards than even in Dactyletlira ; in the higher types, with much 
more compressed faces, these horns lie near each other, and often more or less coalesce 
(“ Pig’s Skull,” plate xxxvi. fig. 1, c.tr., rc.c.). Here the cornua end in a rounded spatula, 
which lies upon and almost coalesces with the “ prenarial lamina” ( p.n.l .), which is, 
here, a projection from the pterygopalatine band (ppg.), and becomes the prepalatine 
spur. Pipa, in these things, is in essential agreement with Dactyletlira ; but the head is 
much more flattened out in the latter than in Pipa. The well-marked lateral part of 
the trabecular floor next to the subocular space (s.o.f.) is similar in both types ; in both 
it forms a flap with a convex outer edge. The recurrent trabecular cornua not only lie 
on the conjugational part of the next arch, but they also close in upon the bars of 
which they form the distal ends. Thus is formed, outside the “ internasal plate,” by 
this looping of each bar, the space by which the outer nostril opens into the fore palate 
(i.n.). In Dactyletlira (Plates 56 & 57) the outer nasal space is formed in the same 
manner; but the next bar (p.n.l., p.pg.) assists to finish the loop, and the space itself 
is transverse, whereas in Pipa it is almost longitudinal. In the latter the transverse 
part of the trabecular internasal plate, in front, is of much less extent, and the inner 
cells of the upper lip are not yet chondrified, so that here we miss not only the 
tentacles, but also the long, transverse, upper labial, from which they spring. 
Coming to the first facial arch, the mandibular, we find all its proximal part in close 
relation with the great trabecular sheet of cartilage. Moreover, here the morpholo- 
gical stage is one much further advanced than in the youngest larvse of Dactyletlira , 
and more in accordance with the fourth stage (Plate 58. figs. 2 & 3) ; this is further 
evidence of rapidity of development; at any rate it shows quick morphological change 
in these parts, the condition of the embryo, as an individual, being so immature. 
The whole pier (the free part of the mandible has been removed, see fig. 4) is a 
multilobular plate, whose inner lobes are confluent with the trabecular margin 
(Plate 60. fig. 3,pd., tr.). The hinder of these (yd.) is a wide flat plate of cartilage 
which runs in obliquely between the auditory capsule and the posterior end of the 
trabecula ; with this it has completely coalesced, with the other partially, on its outside. 
