| a.c. Alimentary canal. | sp.g. Spinal ganglion. | d.p. Neural muscle-plate. |
| n.c. Nerve cord. | n. Notechord. | v.p. Haemal muscle-plate. |
He gets his facts from an elaborate study of the process of budding in the Naidæ, making the somewhat risky assumption that regeneration takes essentially the same course as embryonic development.
He succeeds in showing—to his own satisfaction at least—that in the formation of new segments in Nais and Chætogaster a strand of cells appears between the alimentary canal and the nerve-cord, and that from this axial strand the hæmal muscle-plates grow out dorsally round the alimentary canal and the neural muscle-plates ventrally round the nerve-cord (see [Fig. 14]).
This strand of cells, he concludes, must clearly be the notochord, and the type of development is obviously the double-symmetrical met with in Vertebrates.
The nervous system Semper found to develop in the buds of Nais and Chætogaster by an ectodermal thickening, just as in some Vertebrates. The cerebral ganglion was formed by the ends of the nerve-cord growing up round the œsophagus and fusing with the paired "sense-plates" which develop from the ectoderm of the head. The cerebral ganglion is accordingly only secondarily hæmal in position, and there is no need therefore to seek in Vertebrates for the homologue of the œsophageal commissures of Annelids, as, for instance, Schneider did.
Since the mouth opens on the neural surface in Annelids and on the hæmal surface in Vertebrates, Semper considers that they cannot be equivalent structures, and he finds the homologue of the Vertebrate mouth in a little pit on the hæmal surface of the head in the leech Clepsine (also in the true mouth of Turbellaria and the proboscis-opening in Nemertines). The primitive Annelid mouth, however, does not appear in the embryogeny of Vertebrates, for the great development of the brain crowds it out of existence.
The homologues of the gill-slits Semper finds in two little canals in the head of Chætogaster, which open from the pharynx to the exterior. In Sabellids he describes an elaborate system of gill-canals, with a supporting cartilaginous framework which forms a real Kiemenkorb or gill-basket, comparable with that of Amphioxus.
Gill-slits, notochord, relation of nervous system, mesonephric tubules, are thus common to Annelids and Vertebrates—what further proof could one desire of the close relationship of these groups? Yet Semper enters into refinements of comparison, seeing, for instance, in the lateral portions of the ventral ganglia ([Fig. 14], sp. g.) the homologues of the spinal ganglia of Vertebrates, and comparing the lateral line of sense organs in Annelids with the lateral line in Anamnia.
He will not admit that Amphioxus and the Ascidians show a closer resemblance to Vertebrates than his beloved Annelids. Amphioxus, he thinks, is not a Vertebrate, and Ascidians, though sharing with Annelids the possession of a notochord, gill-slits, and a "dorsal" nervous system, yet are further removed from Vertebrates than the latter by reason of their lacking that essential characteristic of Vertebrates, metameric segmentation.
Not content with establishing the unity of plan of Annelids, Arthropods, and Vertebrates, Semper tries to link on the Annelids, as the most primitive group of the three, to the unsegmented worms, and particularly to the Turbellaria. His speculations on this matter may be summed up somewhat as follows:—The common ancestor of all segmented animals is a segmented worm-like form, not quite like any existing type, resembling the Turbellaria in having two nerve strands on the dorsal side and no œsophageal ring, potentially able to develop either the Vertebrate or the Annelid mouth, and so to give origin both to the Articulate and to the Vertebrate series. The common ancestor alike of unsegmented worms and of all segmented types is probably the trochosphere larva, which in the Vertebrates is represented by the simple Keimblase or blastula.