The presumption is, that in other cases the canal is formed by haemal arches with ribs remaining continuous with them. This view was strongly supported by the late Professor Balfour in his paper on Lepidosteus, published in connection with W. N. Parker (infra). In some Elasmobranchii, e.g. ScyIlium and some Teleostei, the anterior portion of the caudal canal is apparently inclosed by haemal arches, to which ribs are attached laterally. The explanation of this arrangement is as follows (Balfour): - The canal in this region is secondarily acquired: it is a prolongation forward of the posterior part which is normally developed, and it is formed by the growth of two processes - one from each haemal arch near its base, towards one another which meet and fuse, the true haemal arch being that portion of cartilage or bone which bears the rib.

In some instances, e.g. Pike, the right and left neural arches are not united by bone in the median line. It should be noted also that in the region of the trunk at least, the ribs have free ventral ends. There is no sternum.

Other structures to be noted are: (1) the ligamentum vertebrale superius which runs above the spinal cord and connects the successive vertebrae. (2) A pair of cartilages which appear in development between the ligament and the spinal cord: they project posteriorly, ossify or become surrounded by bone and become continuous ultimately with the arch. They are probably homologues of the intercalaria neuralia of the Elasmobranch which lie between the neural arches. They are present also in Ganoids. (3) The interspinal bones which support the azygos fins seen in the preceding preparation (Prep. 17). These according to Gotte, may be regarded as dissociated parts of vertebral spines, but this view is extremely doubtful. Dohrn has recently suggested, and Mayer has strongly supported his view, that the azygos fins are derived from paired rudiments; the anal and ventral element of the caudal fins, by the coalescence behind the primitive cloaca of the two ventro-lateral ridges from which the pectoral and pelvic fins originate: the dorsal fins and the dorsal element of the caudal fin from two similarly coalesced dorso-lateral ridges. If this is so, then the interspinal bones are really equivalents of the basal cartilages in the paired fins.

They are not connected with the vertebral column in most Elasmobranchii. In Dipnoi they remain articulated to the neural spines in the dorsal region, and to the haemal in the ventral region, where in other types the anal fin is formed. They appear in the young Teleostean just above the summits of the arches, but when the latter lengthen, come to lie between them. They are developed also in regions where the azygos fins are deficient; but it must be borne in mind that this system of fins is primitively represented by a continuous fold. (4) The set of bones known as epi-neurals, epi-centrals, epi-pleurals (Fleischgraten). They are formed in the intermuscular septa (myocommata), and never pass through a cartilage stage. Gotte has suggested that the series which lies on the ends of the inferior arches in the region of the trunk are ribs homologous with the ribs of Elasmobranchs owing to their position. But this view appears to have little to support it, and the mode of formation of the bones in question is against it.

It may be noted that the ribs of all fish except Elasmobranchii lie at the base of the myocommata, immediately below the peritoneum. In Elasmobranchii they extend outivards in the fibrous septum which separates the dorso-lateral and ventro-lateral muscle-masses, and therefore not below the peritoneum. Hence these ribs have been regarded as not homologous with the ribs of other fish. But in Lepidosteus the free ends of the ribs, the bodies of which are normally placed, bend outwards into this same septum. It is possible, therefore, that the ribs of Elasmobranchii have undergone a change of position.

Schmid-Monnard has recently investigated the origin of bony tissue in Teleostei. He finds as follows. (1) The first bone is always formed outside the cartilage. (2) All skeletogenous tissues may take part subsequently in the formation of bone, and whether cartilage or connective tissue they ossify either directly or indirectly, i.e. by means of osteoblasts. In most parts of the skeleton bone is derived exclusively from perichondrium. (3) The first-formed bone is invariably homogeneous, containing neither bone-cells nor connective fibres. Bone-corpuscles are to be found here and there in those fish in which, according to Kolliker, they are absent, and the bone resembles dentine, e. g. Pike, Perch, Lota vulgaris, Gadus aeglefinus, etc. But true dentine, i. e. tubular dentine, appears never to be formed. The great development and regular arrangement of Sharpey's fibres probably led to the supposition, coupled with the then imperfect methods of research at command.

Formation of vertebrae, arches and ribs in Teleostei, Gotte, A. M. A. xvii. 1879; Grassi, M. J. viii. 1882; in Ganoidei and Elasmobranchii, Gotte, A. M. A. xv. 1878; in Elasmobranchii, Balfour, Elasmobranch Fishes, London, 1878; in Lepidosteus, Id., and W. N. Parker, Ph. Tr. 173, 1882.

For discussion as to homologies of ribs and formation of caudal canal, see work last quoted.

Azygos fins. Mayer, Mitth. Zool. Stat. Naples, vi. 1885.

Formation of bone.Schmid-Monnard, Z. W. Z. xxxix., 1883, and Kolliker, P. R. S. ix. 1859. Cf. Kostler, Z. W. Z. xxxvii. 1882.