184 VERTEBRATA strable that two originally lateral cords have coalesced neutrally to form the Annelid s ventral nerve-chain. The comparison of the Vertebrates nervous system with that of the Nemertines had already been made by the present writer, as cited by Hubrecht (//) in connexion with the latter s discovery of a complete sub-epidermal nerve-tunic in those worms. Hubrecht has more recently on two occasions (12 and 13) developed an in teresting and important comparison of Nemertine and Vertebrate structure. He has in the first place suggested that the notochord of Vertebrata is nothing more than a modified survival of the pro boscidean sheath of the Nemertines, whilst the oral invagination of the epidermis, in connexion with the hypophysis cerebri of the Vertebrate, may be a last remnant of the proboscis itself. More conclusively he has drawn attention to the median dorsal nerve of Nemertines as corresponding to the Vertebrate cerebro-spinal nerve- cord, whilst the great lateral nerve-cords of Nemertines, and the lateral ganglia in which they expand anteriorly, are compared to the lateral ganglia of the cephalic region of Craniate Vertebrata and the nerve of the lateral line (see fig. 6). The comparison is strengthened by the existence of a metameric series of transverse nerves in the Nemertine, which correspond in respect of their meta merism and their connexion with a dorsal median trunk, with the spinal nerves of Craniata. Hubrecht is careful to insist that he does not regard the Nemertines as representing the direct ancestry of Vertebrata ; but he points out that from the primitive condition of an elongate animal, with a plexiform nerve-tunic, it is readily conceivable that a form was developed in which the nervous tissue was concentrated in three cords, a median dorsal and two lateral, and from such a form we can derive the Craniates condition by excessive development of the median tract and relatively small development of the lateral cords, whilst the Nemertines condition would be attained by the converse process. The tubular condition of the cerebro-spinal nerve-cord of Vertebrata, it may here be re marked, is now very generally regarded as being in its origin a purely developmental feature. It was primitively separated from the epidermis by delamination and in-sinking, and the mode of formation by invagination of a canal has been substituted in accord ance with a general embryological law of growth, which is that bulky structures originating beneath a surface from the cells form ing that surface take up their position in embryonic growth by in vagination of the parent surface. The tubular form, having thus started, seems to have been utilized during one phase of Vertebrate evolution for the respiration of the nervous tissue, by the introduc tion through an anterior unclosed pore of a current of water, which escaped by the neuranal canal (as in larval Amphioxus). There is a wide gap between any form presenting an approach to a Nemertine Worm and the most simple Craniate Vertebrate which can be imagined still provided with the organization characteristic of all Craniata. To pass from such a Worm-like animal to a Craniate, we have to account for and introduce, amongst other new develop ments, (1) a greatly increased metamerism, showing itself in the segmentation of the muscles of the body-wall and in the repetition of the nephridia ; (2) the characteristic sense organs; (3) the lateral and median longitudinal folds or continuous fins ; (4) the carti laginous rods and bars of the skeleton ; (5) the gill-slits, even if we admit the notochord to be represented by the proboscidean sheath. It remains to inquire whether the structure of the other Verte brata throws light on this long hypothetical passage from the simple Worm phase to the elaborate Craniate, or suggests any other ancestry. THE CEPHALOCHOKDA. Char- Cephalochorda are Vertebrata in which there is no anterior acters of dilatation of the nerve-tube to form a brain (see fig. 6) and no Cephalo- specialized skeletal brain-case. The notochord extends from one chorda, extremity of the elongate body to the other as a tapering uncon- stricted rod, passing anteriorly some distance in front of the nerve-cord. The longitudinal muscles of the body-wall are divided by transverse fibrous septa into a series of segments (sixty-two in Amphioxus lanceolatus), the more anterior of which are in front of the mouth and not in any way fused to form a head or cranial structure. Dense connective tissue (differing but little from car tilage) forms an unsegmented sheath to the notochord and an unbroken neural canal above it, in which the nerve-cord lies. The same tissue forms a series of metamerically repeated fin rays, which support the base of a median fin extending along the entire dorsal surface. The fin is continued ventrally from the caudal ex tremity as far forward as the anus, but without fin rays. Tvo lateral up-growths of the body-wall (the epipleura) extend one on either side from the head as far back as the anus. Each of these is divided into three regions, (1) an anterior, which forms the praoral hood ; (2) a median, which forms the wall of the great branchial chamber, the two folds meeting one another and coales cing in the ventral mid-line, excepting where they leave a posterior median aperture, the atriopore ; and (3) the post-atrioporal prreanal ventral fin (extending between atriopore and anus), which is formed by the complete coalescence of the two folds behind the atriopore. FIG. 7. Amphioxus lanceolatus, Yarrell (Rranchiostoma liibricum, Coste). (Original drawings.) (1) Lateral view of adult, to show general form, tha myomeres, fin rays, and gonads. A, oral tentacles (28 to 32 in full-grown animals, 20 to 24 in half-grown specimens) ; B, prceoral hood or prseoral epi- pleur ; C, plicated ventral surface of atrial chamber ; D 1 , D 1 ?, D -6 ; gonads, twenty-six pairs, coincident with myotomes 10 to 36 ; E, metapleur or lateral ridge on atrial epipleur ; F, atriopore, coincident with myotome 36 ; G 1 , G J5 , G 34 , double ventral iin rays, extending from myotomes 37 to 52, but hav ing no numerical relation to them ; H, position of anus, between myotomes 51 and 52 ; I, notochord, projecting beyond myotomes ; K 7 , Iv 2 ?, K.*> 2 , myo tomes or muscular segments of body- wall, 62 in number ; L ll >o, L-30, L 25;J , dor sal fin rays, about 250 in number, the hard substance of the ray being absent at the extreme ends of the body (these have no constant numerical relation to the myomeres); M, notochord as seen through the transparent myotomes, the thin double-lined spaces being the connective-tissue septa ami the broader spaces the muscular tissue of the myotomes ; N, position of brown funnel of left side (atrio-coelomic canal) ; O, nerve-tube resting on notochord. (2) Dissection of Amphioxus. By a horizontal incision on each side of the body a large ventral area has been separated and turned over, as it were on a hinge, to the animal s left side. The perforated pharyngeal region has then been detached from the adherent epipleura or opercular folds (wall of atrial or branchial chamber) by cutting the fluted pharyngo-pleural membrane d, and separated by a vertical cut from the intestinal region, a, Edge of groove formed by adhesion of median dorsal surface of alimentary canal to sheath of notochord ; b, median dorsal surface of alimentary canal ; c, left dorsal aorta ; cc, single dorsal aorta, formed by union of the two anterior vessels ; cc , same vessel resting on intestine ; d, cut edge of pharyngo-pleural folds of atrial tunic, really the original outer body-wall before the downgrowth ol epipleura ; d , atrial tunic (original body- wall) at non-perforate region, cut and turned back so as to expose peri-enteric coelom and intestine r ; , up standing folds of body-wall (pharyngo-pleural folds) on alternate bars of per forate region of body ; /, atrio-ccelomic canals or brown funnels (collar-pores of Balanoglossus); g, cavity of a gonad-sac ; m, cut musculature of body-
wall ; n, anus ; o, post-atrioporal extension of atrial chamber in form of a