MR. ST. GEORGE MIVART ON THE VERTEBRATE SKELETON. 385 
of fishes this undifferentiated condition of the paraxial and hypaxial parts obtains. 
Nevertheless, sometimes both elements coexist in full perfection, as is well shown by 
Dr. Cleland himself, whose instructive woodcut of Naseus fronticornis I have had copied 
(Plate LITI. fig. 9). In this fish we have on each side a diapophysis (or tubercular pro- 
cess) with an upper rib (fig. 9, £, tr), and a parapophysis (or capitular process) from which 
an aponeurosis (fig. 9, dotted line), in the position of a lower rib, runs down to embrace 
the visceral cavity while the hypaxial arch plainly and manifestly dips into that cavity 
(like the caudal chevron bone of a crocodile) in the line of suspension of the primitive 
embryonie internal laminæ. Dr. Cleland declines to recognize a relation of homology 
between this arch of Naseus fronticornis and the chevron bone of Menobranchus, because 
in the latter animal the visceral cavity “lies entirely below the inferior caudal spine, 
whose top only reaches its upper margin." But this objection, if valid, would destroy 
the homology between the chevron bones of Menobranchus and those of the crocodile, as 
deseribed by Professor Goodsir, at least of such of them as are quite posterior to the cau- 
dal leg-muscles; and it seems to me to imply no more than that the area of the internal 
lamine has become narrowed for a certain space, certainly not any essential difference in 
the nature of the parts—nothing being more common than the application of serially 
homologous parts to diverse uses. 
In Murena Helena we have a good example (Plate LIII. figs. 16, 17, 18 & 19) of a rudi- 
mentary and undifferentiated condition of these parts. In the vertebrz immediately behind 
the head (fig. 16) we have a transverse process on each side, and a single median inferior 
process. This latter process, as it lies directly in the line suspending the viscera, must be 
regarded as hypaxial (fig. 16 x). Further backward it divides and approximates toward 
the transverse process, with which it gradually coalesces (fig. 17, PH). This single process 
is now the serial homologue of both the hypaxial and paraxial parts of the more anterior 
vertebræ, i. e. itis parhypaxial. It is short and does not extend beyond the walls of the 
groove for the aorta, which groove is formed into a canal by membrane extending trans- 
versely beneath it. From the ends of these parhypaxial processes the membrane extends 
which bounds the pleuro-peritoneal cavity externally, and in which there are no hard 
parts; but if hard parts did exist there, they would be ribs. Further backward the 
parhypaxial process becomes again differentiated into its elements (fig. 18); and still fur- 
ther back the hypaxial ones, meeting below, form inferior arches (fig. 19) which are the 
general homologues of the Mammalian and Reptilian chevron bones, and the serial 
homologues of the inner parts of the single more anterior processes. Did the trans- 
versely extending subaortic membrane of the trunk ossify, then such median ossification 
would be plainly hypaxial, and would connect together (if it extended completely across) 
the two parhypaxial processes. 
Of such an essentially complex nature are the subcaudal arches of most fishes. 
These arches are in general evidently the bent-down serial homologues of the paraxial 
elements of the trunk; but they are often also the serial homologues of the margins 
of a subcentral abdominal aortic groove, and so far are hypaxial, as also are the trans- 
verse bars of bone sometimes extending beneath the caudal vessels. 
In certain cases, however, we have (as in Murena Helena) a more complete differenti- 
VOL. XXVII. 3 F 
