may even be present in the same animal. If we follow the process
of progressive change still further back, we reach a stage in which
cephalization had practically disappeared, and where even
metameric segmentation was in a much less advanced condition. The
tadpoles of Ascidians, and still more remotely Balanoglossus, although
still less than Amphioxus to be regarded as actual ancestral vertebrate
types, give images of some of the many phases in which the
ancestral type may have been exhibited. It is needless to say that
the creatures exhibiting such a stage in the ancestry of the
Vertebrates would have formed simply one in the vast series of marine
coelomate types which the anatomy of the Invertebrates shows us
to have existed. Its distinguishing features would have been the
presence of gill-slits, of the skeletal rod, known as the notochord,
and of the dorsal tubular nervous system. We cannot make even
profitable guesses as to the exact conditions under which these
features, or the corresponding features of other coelomate types,
arose in the kaleidoscopic differentiation of form, but consideration
of the general morphology of the nervous system enables us to see
the Chordate ancestor in its true perspective amongst other coelomic
groups. In the Coelentera the nervous system appears as a diffused
layer of cells and fibres, underlying, and in close connexion with,
the epidermis. This diffused layer may thicken in special regions,
forming rings round apertures, radial bands, and so forth, whilst
in the intervening areas it disappears. In the different groups
of Coelomates specialized bands and strands have formed in this
way from a primitive diffuse system, giving rise to the nervous
patterns distinctive of the various groups, whilst a second process,
that of inward migration from the epidermis, produces further
changes. In the Turbellaria there have been formed two
ventrolateral cords with variously placed anastamoses; in the Trematodes,
two ventral, two lateral and two dorsal cords with variously placed
anastomoses, and in the Cestodes two lateral and in some cases
one dorsal cord. In the Nemertea the primitive continuous
sub-epidermal sheath is retained with two lateral and sometimes one
dorsal thickening. In the Nematodes there are one dorsal, one
ventral and at each side two lateral thickenings, sometimes separated
cords, sometimes mere sub-epidermal bands, whilst the traces of a
circum-oesophageal ring may be regarded as another specialization
of the primitive complete sheath. In Balanoglossus there is a
continuous sheath with a dorsal and ventral band, the latter in
certain regions showing traces of a tubular structure. In Annelids
and Arthropods there are two ventral bands tending to unite in the
median ventral line, and a circum-oesophageal collar. In the
Chordates there is a continuous dorsal band, which secondarily migrates
inwards and becomes tubular. In almost any of these types, as
the individual becomes more integrated, there is a tendency for
the nervous matter of the specialized areas to become still further
massed; and in bilaterally symmetrical animals with forward
progression and the beginning of cephalization a specially important
mass forms something comparable with a brain in special relation
with the sense organs of the primitive head. If the problem of
vertebrate origin be considered from the wide point of view of
comparative anatomy, it becomes no more difficult nor remarkable
than the differentiation of any other type amongst simple, marine,
unsegmented, or little segmented, wormlike creatures. It is obvious,
however, that such a theory of origin cannot expect confirmation
from the geological record, as it supposes a differentiation of the
main chordate characters in a stage too simple to leave fossil remains.
Reference must be made, however, to definite theories of the origin of Vertebrates which have been successively urged by anatomists. A. Dohrn, if not the inventor, was the most ingenious advocate of the Annelid theory. He recognized the fundamental importance of segmentation in vertebrate structure and sought for a highly segmented ancestor. Partly influenced by Ray Lankester's studies on degeneration, he held that the apparently simplest living members of a group may give misleading clues with respect to the ancestral line, and he devoted much brilliant anatomical and embryological work to develop the thesis that Amphioxus and the Tunicates were degenerate offshoots from a higher vertebrate stock. He took a Chaetopod worm as the closest living representative of the stock of all segmented animals, and in particular of the Vertebrates, laying stress on the segmentation, the large coelom, the segmental excretory tubules, the vascular system with red blood, the segmentally disposed branchiae, the lateral organs of locomotion, and the tendency to form a distinct head. The chief difficulty was the nervous system, and this he explained by accepting an idea propounded many years before by De Blainville, that the dorsal surface of Vertebrates was homologous with the ventral surface of Annelids and Arthropods. He assumed that the ancestral type was a marine creature in which reversal of surface was of little physiological moment. He supposed that a new mouth had been formed, probably by a coalescence of a pair of gill-slits on what was to be the ventral surface of the vertebrate, and that the old invertebrate mouth with the downward turn of the anterior end of the alimentary canal, between the diverging ends of the ventral nerve cords, was to be sought for in the roof of the vertebrate brain, possibly the pineal body. Dohrn's theory has failed to find acceptance for many reasons, of which the chief are the difficulty as to reversal of surfaces, the knowledge that segmentation occurs independently in many groups of animals and in different organs, greater knowledge of the vascular, excretory and nervous systems, and in particular the discovery that the pineal body was a degenerate eye. F. M. Balfour from the first refused to accept Dohrn's theory and suggested that the dorsal position of the nerve cord in Vertebrates could be accounted for by supposing that the primitive condition was a lateral cord at each side such as were then known to occur in Nemertines, and that these cords had fused dorsally in Vertebrates, ventrally in Annelids. A. A. W. Hubrecht soon afterwards discovered the existence of a continuous nerve sheath in Nemertines, and he and Ray Lankester suggested a Nemertine origin for Vertebrates, and homologized the notochord with the proboscis sheath, Ray Lankester, in particular, pointing out that the tubular condition of the vertebrate nervous system was secondary, that it consisted essentially of a dorsal band, which sank inwards, and that the canal might have been at first an epidermal water canal. These authors were emphatic in laying stress on the view that no actual Nemertine could be supposed to represent the vertebrate ancestor, but that the Nemertines were to be taken merely as showing the kind of material out of which the vertebrate structure, and in particular the vertebrate nervous system, might have arisen. The view adopted in this article as given above, is in reality an extension of the Hubrecht-Lankester theory.
The theory of vertebrate origin that has been most elaborately expounded is W. H. Gaskell's hypothesis that they are descended from Arthropods. Gaskell accepts Dohrn's view of the importance of segmentation and of the degeneracy of Amphioxus and Tunicates, but rejects the conception of a reversal of surfaces. He takes the larval stage of a Cyclostome as the most generalized living representative of the essential vertebrate type, and selects Limulus, the kingcrab, in a very general way, as the closest living representative of such an Arthropod type as might have been the vertebrate ancestor. The starting-point of Gaskell's theory is the conception of the vertebrate nervous system as a band of nervous tissue which immediately underlies and gradually grows up round a distinct epidermal tube, the tube which forms the vesicles of the brain and the central canal of the spinal cord. Ray Lankester had already applied this to the Nemertine theory, but Gaskell urges that it affords an immediate comparison with Arthropod structures. The ventral mouth of Limulus leads vertically upwards through a ring of nervous tissue, the circumoesophageal commissure, into an expanded stomach, and from this the digestive tube runs back to the anus immediately dorsal to the ventral nerve chain. For Gaskell the infundibulum is the Arthropod oesophagus, the ventricles of the brain are the stomach, and the spinal canal leading back to fuse with the anus at the neurenteric canal is the Arthropod digestive tract. In the Vertebrate a new digestive tract has been formed, probably from a structure corresponding to the branchial chamber of Arthropods. The lateral halves of the ventral nervous system of the Arthropod, where they diverge on either side of the oesophagus, represent the crura cerebri of Vertebrates, whilst the supra-oesophageal ganglia represent the fore-brain. Gaskell has instituted an elaborate comparison, extending to very minute details of structure, and finds remarkable analogies between the organs of Arthropods and structures in the Vertebrates. From the palaeontological side, he points out that at the time when the earliest known Craniates were abundant, large Arthropods, essentially like Limulus, were also abundant. He thinks it probable that Vertebrates arose from a dominant invertebrate group, and points to many resemblances in detail between the Silurian Arthropods Palaeostraca and the Craniate Ostracoderms of the same horizon.
Bibliography.—F. M. Balfour, Monograph on the Development of Elasmobranch Fishes (1878); W. Bateson, “Balanoglossus,” in the Quarterly Journal of Microscopical Science (1884, 1885, 1886); J. Beard, “The System of Branchial Sense-Organs and their Associated Ganglia in Ichthyopsida,” in the Quart. Journ. Micr. Sci. (1885); A. Dohrn, “Studien zur Urgeschichte des Wirbelthierkörpers,” in the Mitth. Zool. Sta. (Naples, 1882, 1884, 1885, 1886, 1904); W. H. Gaskell, The Origin of Vertebrates (1908); C. Gegenbaur, “Die Metameric des Kopfes,” in the Morph. Jahrb. (1888); Grundzüge der Vergleichenden Anatomie (various editions from 1870); E. S. Goodrich, volume on “Vertebrata Craniata,” in Lankester's Treatise on Zoology (1909) (a notable review of the subject, to which the writer of this article is specially indebted); B. Hatschek, “Die Metameric des Amphioxus und des Ammocoetes,” in Vertr. Anat. Ges. (Wien, 1892), and Anat. Anz. (1893); A. A. W. Hubrecht, “On the Ancestral Form of the Chordata,” in the Quart. Journ. Micr. Science (1883); “The Relation of the Nemertea to the Vertebrata" (ibid. 1887); A. Kowalevsky, “Le Développement de l'Amphioxus lanceolatus,” in Arch. Sci. Phys. Nat. (1866); “Entwickelungsgeschichte der einfachen Ascidien,” in St Petersb. Acad. Sci. (1867), and summarized in the Quart. Journ. Micr. Sci. (1870); N. K. Koltzof, “Metameric des Kopfes von Petromyzon Planeri,” in Anat. Anz. (1899); “Entwickel. d. Kopfes von Petromyzon Planeri,” in Bull. Soc. Imp. Nat. Moscow (1902); E. Ray Lankester, “Notes on Embryology and Classification,” in Quart. Journ. Micr. Science (1877); article “Vertebrata,” in Ency. Brit. (9th ed.); A. M. Marshall, “The Segmental Value of the Cranial Nerves,” in the Journ. Anat. and Phys. (1882); H. V. Neal, “Segmentation of Nervous System in
