VERTEBRATA 183 Classification of CHAN I ATA. 1 Grade A. CYCLOSTOMA. Class I. Myxinoidca. II. Pet rom yzontia. Grade B. GNATHOSTOMA. Grade a. Branchiata heterodactyla. Class I. Pisces. II. Dipnoi. Grade b. Branchiata pentadactyla. Class. Amphibia. Grade c. Pentadactyla lipobranchia. Branch a. Monocondyla. Branch b. Amphicondyla. Class I. Rcptilia. Class. Mammalia. II. Avcs. Ances- If we now briefly consider what must have been the common tral form ancestral form from which these Craniata have proceeded, making ofCrani- use of such internal evidence as their structure affords, we find ata. that we get no further back than such an animal as would fit the description given above, with the exception that we should be warranted in substituting in the ancestor a pair of continuous lateral fins, with comb-like cartilaginous skeleton, in place of the two pairs of fins, or their total defect, seen in living Craniata. We get no clear suggestion from the study of Craniata themselves as to the meaning of the curious shape of the brain and its outgrowths (though the pineal outgrowth has recently been explained as an eye), nor as to the original genesis of the notochord. We should, however, be justified in representing that region which now cor responds to the hinder part of the skull and brain as more fully developed and segmented, so as to give a series of separate myotomes and perhaps separate nerves corresponding to the several furcal branches of the vagus ; and we may very well suppose that the num ber of pharyngeal gill-slits was larger in the ancestral than in any living form, though it seems improbable that in any true Craniate did each gill-slit correspond to a distinct muscular segment. An attempt to go further than this has been made by Dr Anton Dohrn by the method of hypothesis and subsequent corroborative inquiry into facts of minute structure and embryological history. Making use of the principle of degeneration, Dohrn started with the legitimate hypothesis that the branches of Vertebrata other than Craniata viz., Cephalochorda, Urochorda, and (though at the time he commenced his work their structure was not fully under stood) Hcmichorda were not to be regarded as permanent records of steps in the evolution of Craniata, but rather as greatly de generate offshoots from the ancestors of that group, which could throw but little light on the character of their non- degenerate ancestors. A second fundamental assumption which led Dohrn to his position was that the segmentation of the Craniates body-wall is a primitive and essential feature in their structure, and becomes more and more fully expressed instead of less developed the further we go back in their ancestry. Dohrn, in fact, assumes that what is called metameric segmentation is a phenomenon of structure which has occurred once only in the history of animal form, and that all segmented animals are genetically related and descended from a common segmented ancestor. Assuming this, he pointed to the existing Chaitopod Worms as most nearly representing at the present day the common ancestor of segmented animals. They have, as he pointed out, a high organization, little inferior to that of the lowest Craniata; they possess a well -developed ccelom, blood- vessels with red blood, a segmental series of nephridia (modified in some as gonaducts), segmental branchiae, and lateral locomotive organs ; not a few develop cartilage as a skeletal sup port ; and many show a concentration and fusion of segments to form a complex head, which resembles, so far, that of Craniata. The ventral in place of the dorsal position of the nerve-cord led Dohrn to accept De Blainville s conception that the dorsal and ventral surfaces are reversed in Vertebrata, as compared with Annelids, Crustaceans, and Insects, so that the Vertebrate is com pared to an Insect walking with its ventral surface upward. This led further to the notion that the mouth of the Clwtopod or Annelid, which penetrates the nerve-cord, or rather passes between its two divaricated lateral constituents in those animals, has in Craniata disappeared, its place being taken by a new mouth de rived from the modification of a pair of gill-slits. The remnant of the old mouth, which should, it the comparison instituted holds good, lie in Craniata somewhere on the dorsal surface of the cranial region, was sought by Dohrn in some of the peculiar and hitherto unexplained median structures of the brain: at one time the fourth ventricle with its deficient roof was suggested as thus to be explained, whilst subsequently the curious median structures, the pineal and pituitary bodies, were called in as possibly thus significant. Without pursuing further the elaboration of Dohrn s views, it must be at once noted that, whilst the legitimacy of the assumption of degeneration must be admitted, the second assumption, viz., that metameric segmentation is a character bringing all forms showing it into a special genetic continuity, cannot be accepted, "i The classes here enumerated are described in separate articles, whilst Cyclostoma and Dipnoi are included in the article ICHTHYOLOGY. The property of repeating units of structure, so as to build up a complex of many similar parts united to form one individual, is a very general one in organic forms, and is exhibited in various con ditions by both animals and plants. Its simplest expression is found in cell-structure and the binary division of cells. It shows itself as affecting larger masses of structure in the arborescent colonies of Ccelentera, in the radial or antimeric composition of Echinoderms and of Compound Ascidians, and in the linear or metameric segmentation of Worms, Arthropods, and Vertebrates. There is abundant evidence that this property is a general one, which may assert itself at any period in the history of a group of animals, and does not imply special unity of origin in forms which exhibit it. As pointed out in the article HYDKOZOA, merogenesis the name applicable to this phenomenon generally may take an extreme and complete character, leading to the separation and independence of the units of structure produced ; in that case it may be termed eumerogenesis. Or the process may be very partial, occurring only during a period of embryonic growth, and subse quently ceasing, so that later growth obscures or obliterates it altogether (dysmerogenesis). There is no ground for assuming that either one of these extremes is fundamental or original. Any mechanical or nutritional condition may lead to merogenesis in an organism in which the tissues have a certain reproductive capacity, or have not acquired final differentiation ; and it will depend upon the balance of advantage, determined by natural selection, whether the segmentation (supposing the merogenesis to take the linear form) results in the separation of segment-buds, or in the formation of an annulate body, or leaves traces of its occurrence only in certain tissues and organs. The Cestoid Worms present within the range of a single group almost every grade of eumerogenesis and dysmero genesis (Caryophyllseus, Ligula, Taenia). In the otherwise amero- genetic Moltusca, Chiton and the pearly Nautilus show dysmero genesis in certain organs, whilst the Planarian Worms frequently exhibit eumerogenesis in their bud-segmentation (to be compared with that of the Annelid Ctcnodrilus described by Zeppelin, 9) and the elongated Nemertines only slight traces of dysmerogenesis. If we deny Dohrn s assumption with reference to segmentation, we are no longer led in the direction of the Annelids (Chsetopods) in our search for the ancestry of the Craniate Vertebrata. The fact that the notochord is the forerunner of the segmented vertebral column, and is itself never segmental, instead of being a difficulty, acquires directive significance. The fact that the nerve tube is dorsal, and not ventral, no longer requires the large assump tion that animals have reversed their habitual carriage, but suggests that the Craniates ancestor had a dorsal median nerve, which has increased in size and importance so as to become the nerve-tube of existing forms. m The explanation of the curious struc ture of the brain will have to be found otherwise than in the assump tion of a perforating pharynx, an as sumption which the recent discovery of the true nature of the pineal body has rendered untenable in the latest form advocated by its ingenious author, whose speculations, nevertheless, de serve the fullest re cognition as having stimulated inquiry and guided observa tion. Balfour (ro) in 1878 refused to a- dopt Dohrn s views, and considered it probable that the dorsal position of the nerve - cord in Vertebrata could be accounted for, with out any assumption of a substitution of a pair of gill-slits for the original mouth, AMPHIOXUS Fio. 6. Comparison of nervous systems of a Ncmertino, a primitive Craniate, and Ampli ioxus. m, Median dor sal nerve, which becomes the myelon in the Craniate and Amphioxits, acquiring an anterior enlargement in the former; I, lateral nerve (right and left), absent by degeneration in Amphioxiis Ig, ganglia of lateral nerve, forming a single large lobe on each side in the Nemertine, and broken into a metameric series in the Craniate ; , roots of vagus nerve of the Craniate ; dr, dorsal roots of nerves given off from myelon or median dorsal nerve ; vr, ventral roots of these nerves, here represented as separate nerves; 0, mouth. (After Hubrecht.) by assuming that primitively the nerve-cord consisted of two lateral cords, as seen at the present time in the Nemertine Worms, and that these cords
have coalesced dorsally in Vertebrata, just as it is clearly demon-