§ 208. Insects, Arachnids, Crustaceans, and Myriapods, are all members of that higher division of the Annulosa[26] called Articulata or now more generally Arthropoda. Though in these creatures the formation of segments may be interpreted as a disguised gemmation; and though, in some of them, the number of segments increases by this modified budding after leaving the egg, as it does among the Annelids; yet the process is not nearly so dominant: the segments are usually much less numerous than we find them in the types last considered. In most cases, too, the segments are in a greater degree differentiated one from another, at the same time that they are severally more differentiated within themselves. Nor is there any instance of spontaneous fission taking place in the series of segments composing an articulate animal. On the contrary, the integration, always great enough permanently to unite the segments, is frequently carried so far as to hide very completely the individualities of some or many of them; and occasionally, as among the Acari, the consolidation, or the arrest of segmentation, is so decided as to leave scarcely a trace of the articulate structure: the type being in these cases indicated chiefly by the presence of those characteristically-formed limbs, which give the alternative name Arthropoda to all the higher Annulosa. Omitting the parasitic orders, which, as in other cases, are aberrant members of their sub-kingdom, comparisons between the different orders prove that the higher are strongly distinguished from the lower, by the much greater degree in which the individuality of the tertiary aggregate dominates over the individualities of those secondary aggregates called segments or “somites,” of which it is composed. The successive Figs. 170–176, representing (without their limbs) a Julus, a Scolopendra, an isopodous Crustacean, and four kinds of decapodous Crustaceans, ending with a Crab, will convey at a glance an idea of the way in which that greater size and heterogeneity reached by the higher types, is accompanied by an integration which, in the extreme cases, nearly obliterates all traces of composite structure. In the Crab the posterior segments, usually folded underneath the shell, alone preserve their primitive distinctness. So completely confluent are the rest, that it seems absurd to say that a Crab’s carapace is composed of as many segments as there are pairs of limbs, foot-jaws, and antennæ attached to it; and were it not that during early stages of the Crab’s development the segmentation is faintly marked, the assertion might be considered illegitimate.

That all articulate animals are thus composed from end to end of homologous segments, is, however, an accepted doctrine among naturalists. It is a doctrine that rests on careful observation of three classes of facts—the correspondences of parts in the successive “somites” of an adult articulate animal; the still more marked correspondences of such parts as they exist in the embryonic or larval articulate animal; and the maintenance of such correspondences in some types, which are absent in types otherwise near akin to them. The nature of the conclusion which these evidences unite in supporting, will best be shown by the annexed copies from the lecture-diagrams of Prof. Huxley; exhibiting the typical structures of a Myriapod, an Insect, a Spider, and a Crustacean, with their relations to a common plan, as interpreted by him.

Treating of these homologies, Prof. Huxley says “that a striking uniformity of composition is to be found in the heads of, at any rate, the more highly organized members of these four classes; and that, typically, the head of a Crustacean, an Arachnid, a Myriapod, or an Insect, is composed of six somites (or segments corresponding with those of the body) and their appendages, the latter being modified so as to serve the purpose of sensory and manducatory organs.”[27]

Thus even in the higher Arthropoda, the much greater consolidation and much greater heterogeneity do not obliterate all evidence of the fact, that the organism is an aggregate of the third order. Comparisons show that it is divisible into a number of proximate units, each of which is akin in certain fundamental traits to its neighbours, and each of which is an aggregate of the second order, in so far as it is an organized combination of those aggregates of the first order which we call morphological units or cells. And that these segments or somites, which make up an annulose animal, were originally aggregates of the second order having independent individualities, is an hypothesis which gathers further support from the contrast between the higher and the lower Arthropods, as well as from the contrast between the Arthropods in general and the Annelids. For if that masking of the individualities of the segments which we find distinguishes the higher forms from the lower, has been going on from the beginning, as we may fairly assume; it is to be inferred that the individualities of the segments in the lower forms, were originally more marked than they now are. Reversing those processes of change by which the most developed Annulosa have arisen from the least developed; and applying in thought this reversed process to the least developed, as they were described in the last Chapter; we are brought to the conception of attached segments that are all completely alike, and have their individualities in no appreciable degree subordinated to that of the chain they compose. From which there is but one step to the conception of gemmiparously-produced individuals which severally part one from another as soon as they are formed.

§ 209. We must now return to a junction whence we diverged some time ago. As before explained under the head of Classification, organisms do not admit of uniserial arrangement, either in general or in detail; but everywhere form groups within groups. Hence, having traced the phases of morphological composition up to the highest forms in any sub-kingdom, we find ourselves at the extremity of a great branch, from which there is no access to another great branch, except by going back to some place of bifurcation low down in the tree.

There exist such similarities of shape and structure between the larval forms of low Molluscs and those of Annelids and Rotifers, as to show that there was an early type common to them all; and its probable characters, suggested by comparison, seem to imply that it had arisen from some cœlenterate type, intermediate between the Cnidaria and the Ctenophora. But there is this noteworthy difference between the molluscan larva and the allied larvæ, that it gives origin to only one animal and not to a group of animals, united or disunited. No true Mollusc multiplies by gemmation, either continuous or discontinuous; but the product of every fertilized germ is a single individual.

It is a significant fact that here, where for the first time we have homogenesis holding throughout an entire sub-kingdom, we have also throughout an entire sub-kingdom no case in which the organism is divisible into two, three, or more, like parts. There is neither any such clustering or branching as a cœlenterate or molluscoid animal usually displays; nor is there any trace of that segmentation which characterizes the Annulosa. Among these animals in which no single egg produces several individuals, no individual is separable into several homologous divisions. This connexion will be seen to have a probable meaning, on remembering that it is the converse of the connexion which obtains among the Annulosa, considered as a group.

A Mollusc, then, is an aggregate of the second order. Not only in the adult animal is there no sign of a multiplicity of like parts that have become obscured by integration; but there is no sign of such multiplicity in the embryo. And this unity is just as conspicuous in the lowest Lamellibranch as in the highest Cephalopod.

It may be well to note, however, more especially because it illustrates a danger of misinterpretation presently to be guarded against, that there are certain Molluscs which simulate the segmented structure. Externally a Chiton, Fig. 188, appears to be made up of divisions substantially like those of the creature Fig. 189; and one who judged only by externals, would say that the creature Fig. 190 differs as much from the creature Fig. 189, as this does from the preceding one. But the truth is, that while 190 and 189 are closely-allied types, 189 differs from 188 much more widely than a man does from a fish. And the radical distinction between them is this:—Whereas in the Crustacean the segmentation is carried transversely through the whole mass of the body, so as to render the body more or less clearly divisible into a series of parts which are similarly composed; in the Mollusc the segmentation is limited to the shell carried on its upper surface, and leaves its body as completely undivided as is that of a common slug.[28] Were the body cut through at each of the divisions, the section of it attached to each portion of the shell would be unlike all the other sections. Here the segmentation has a purely functional derivation—is adaptive instead of genetic. The similarly-formed and similarly-placed parts, are not homologous in the same sense as are the appendages of a phænogamic axis or the limbs of an insect.

§ 210. In studying the remaining and highest sub-kingdom of animals, it is important to recognize this radical difference in meaning between that likeness of parts which is produced by likeness of modifying forces, and that likeness of parts which is due to primordial identity of origin. On our recognition of this difference depends the view we take of certain doctrines that have long been dominant, and have still a wide currency.

Among the Vertebrata, as among the Mollusca, homogenesis is universal. The two sub-kingdoms are like one another and unlike the remaining sub-kingdoms in this, that in all the types they severally include, a single fertilized ovum produces only a single individual. It is true that as the eggs of certain gasteropods occasionally exhibit spontaneous fission of the vitelline mass, which may or may not result in the formation of two individuals; so among vertebrate animals we now and then meet with double monsters, which appear to imply such a spontaneous fission imperfectly carried out. But these anomalies serve to render conspicuous the fact, that in both these sub-kingdoms the normal process is the integration of the whole germ-mass into a single organism, which at no phase of its development displays any tendency to separate into two or more parts.

Equally as throughout the Mollusca, there holds throughout the Vertebrata the correlative fact, that not even in its lowest any more than in its highest types, is the body divisible into homologous segments. The vertebrate animal, under its simplest as under its most complex form, is like the molluscous animal in this, that you cannot cut it into transverse slices, each of which contains a digestive organ, a respiratory organ, a reproductive organ, &c. The organs of the least-developed fish as well as those of the most developed mammal, form but a single physiological whole; and they show not the remotest trace of having ever been divisible into two or more physiological wholes. That segmentation which the vertebrate animal usually exhibits throughout part of its organization, is the same in origin and meaning as the segmentation of a Chiton’s shell; and no more implies in the vertebrate animal a composite structure, than do the successive pairs of branchiæ of the Doto, or the transverse rows of branchiæ in the Eolis, imply composite structure in the molluscous animal. To some this will seem a very questionable proposition; and had we no evidence beyond that which adult vertebrate animals of developed types supply, it would be a proposition not easy to substantiate. But abundant support for it is to be found in the structure of the vertebrate embryo, and in the comparative morphology of the Vertebrata in general.

Embryologists teach us that the primordial relations of parts are most clearly displayed in the early stages of evolution; and that they generally become partially or completely disguised in its later stages. Hence, were the vertebrate animal on the same level as the annulose animal in degree of composition—did it similarly consist of segments which are homologous in the sense that they are the proximate units of composition; we ought to find this fundamental fact most strongly marked at the outset. As in the annelid-embryo the first conspicuous change is the elongation and division into segments, by constrictions that encircle the whole body; and as in the arthropod embryo the blastoderm becomes marked out transversely into pieces which extend themselves round the yelk before the internal organization has made any appreciable progress; so in the embryo of every vertebrate animal, had it an analogous composition, the first decided change should be a segmentation implicating the entire mass. But it is not so. Sundry important differentiations occur before any divisions begin to show themselves. There is the defining of that elongated, elevated area with its longitudinal groove, which becomes the seat of subsequent changes; there is the formation of the notochord lying beneath this groove; there is the growth upwards of the boundaries of the groove into the dorsal laminæ, which rapidly develop and fold over in the region of the head. Rathke, as quoted and indorsed by Prof. Huxley, describes the subsequent changes as follows:—“The gelatinous investing mass, which, at first, seems only to constitute a band to the right and to the left of the notochord forms around it, in the further course of development, a sheath, which ends in a point posteriorly. Anteriorly, it sends out two processes which underlie the lateral parts of the skull, but very soon coalesce for a longer or shorter distance. Posteriorly, the sheath projects but little beyond the notochord; but, anteriorly, for a considerable distance, as far as the infundibulum. It sends upwards two plates, which embrace the future central parts of the nervous system laterally, probably throughout their entire length.” That is to say, in the Vertebrata the first step is the marking out on the blastoderm of an integrated structure, within which segments subsequently appear. When these do appear, they are for some time limited to the middle region of the spinal axis; and no more then than ever after, do they implicate the general mass of the body in their transverse divisions. On the contrary, before vertebral segmentation has made much progress, the rudiments of the vascular system are laid down in a manner showing no trace of any primordial correspondence of its parts with the divisions of the axis. Equally at variance with the belief that the vertebrate animal is essentially a series of homologous parts, is the heterogeneity which exists among these parts on their first appearance. Though in the head of an adult articulate animal there is little sign of divisibility into segments like those of the body; yet such segments, with their appropriate ganglia and appendages, are easily identifiable in the articulate embryo. But in the Vertebrata this antithesis is reversed. At the time when segmentation has become decided in the dorsal region of the spine, there is no trace of segments in the parts which are to form the skull—nothing whatever to suggest that the skull is being formed out of divisions homologous with vertebræ.[29] And minute observation no more discloses any such homology than does general appearance. “Remak,” says Prof. Huxley, “has more fully proved than any other observer, the segmentation into ‘urwirbel,’ or proto-vertebræ, which is characteristic of the vertebral column, stops at the occipital margin of the skull—the base of which, before ossification, presents no trace of that segmentation which occurs throughout the vertebral column.”

Consider next the evidence supplied by comparative morphology. In preceding sections (§§ 206, 208) it has been shown that among annulose animals, the divisibility into homologous parts is most clearly demonstrable in the lowest types. Though in decapodous Crustaceans, in Insects, in Arachnids, there is difficulty in identifying some or many of the component somites; and though, when identified, they display only partial correspondences; yet on descending to Annelids, the composition of the entire body out of such somites becomes conspicuous, and the homology between each somite and its neighbours is shown by the repetition of one another’s structural details, as well as by their common gemmiparous origin: indeed, in some cases we have the homology directly demonstrated by seeing a somite of the body transformed into a head. If, then, a vertebrate animal had a segmental composition of kindred nature, we ought to find it most clearly marked in the lowest Vertebrata and most disguised in the highest Vertebrata. But here, as before, the fact is just the reverse. Among the Vertebrata of developed type, such segmentation as really exists remains conspicuous—is but little obscured even in parts of the spinal column formed out of integrated vertebræ. Whereas in the undeveloped vertebrate type, segmentation is scarcely at all traceable.[30] The Amphioxus, Fig. 191, is not only without ossified vertebræ; not only is it without cartilaginous representatives of them; but it is even without anything like distinct membranous divisions. The spinal column exists as a continuous notochord: the only signs of incipient segmentation being given by its membranous sheath, in the upper part of which “quadrate masses of somewhat denser tissue seem faintly to represent neural spines.” Moreover, throughout sundry groups of fishes and amphibians, the segmentation remains very imperfect: only certain peripheral appendages of the vertebræ becoming defined and solidified, while in place of the bodies of the vertebræ there still continues the undivided notochord. Thus, instead of being morphologically composed of vertebral segments, the vertebrate animal in its primitive form is entirely without vertebral segments; and vertebral segments begin to appear only as we advance towards developed forms. Once more, evidence equally adverse to the current hypothesis meets us on observing that the differences between the parts supposed to be homologous, are as great at first as at last. Did the vertebrate animal primordially consist of homologous segments from snout to tail; then the segments said to compose the skull ought, in the lowest Vertebrata, to show themselves much more like the remaining segments than they do in the highest Vertebrata. But they do not. Fishes have crania made up of bones that are no more clearly arrangeable into segments like vertebræ, than are the cranial bones of the highest mammal. Nay, indeed, the case is much stronger. The simplest fish possessing a skeleton, has a cranium composed of cartilage that is not segmented at all!

Besides being inconsistent with the leading truths of Embryology and Comparative Morphology, the hypothesis of Goethe and Oken is inconsistent with itself. The facts brought forward to show that there exists an archetypal vertebra, and that the vertebrate animal is composed of archetypal vertebræ arranged in a series, and severally modified to fit their positions—these facts, I say, so far from proving as much, suffice, when impartially considered, to disprove it. No assigned, nor any conceivable, attribute of the supposed archetypal vertebra is uniformly maintained. The parts composing it are constant neither in their number, nor in their relative positions, nor in their modes of ossification, nor in the separateness of their several individualities when present. There is no fixity of any one element, or connexion, or mode of development, which justifies even a suspicion that vertebræ are modelled after an ideal pattern. To substantiate these assertions here would require too much space, and an amount of technical detail wearisome to the general reader. The warrant for them will be found in a criticism on the osteological works of Prof. Owen, originally published in the British and Foreign Medico-Chirurgical Review for Oct. 1858. This criticism I add in the Appendices, for the convenience of those who may wish to study the question more fully. (See Appendix B.)

Everything, then, goes to show that the segmental composition which characterises the apparatus of external relation in most Vertebrata, is not primordial or genetic, but functionally determined or adaptive. Our inference must be that the vertebrate animal is an aggregate of the second order, in which a relatively superficial segmentation has been produced by mechanical intercourse with the environment. We shall hereafter see that this conception leads us to a consistent interpretation of the facts—shows us why there has arisen such unity in variety as exists in every vertebral column, and why this unity in variety is displayed under countless modifications in different skeletons.[31]

§ 211. On glancing back at the facts brought together in these two chapters, we see it to be probable that there has gone on among animals a process like that which we saw reason to think has gone on among plants. Minute aggregates of those physiological units which compose living protoplasm, exist as Protozoa: some of them incoherent, indefinite, and almost homogeneous, and others of them more coherent, definite, and heterogeneous. By union of these nucleated particles of sarcode, are produced various indefinite aggregates of the second order—Sponges, Polycytharia, Foraminifers, &c.; in which the compound individuality is scarcely enough marked to subordinate the primitive individualities. But in other types, as in Hydra, the lives of the morphological units are in a considerable degree, though not wholly, merged in the life of the integrated body they form. As the primary aggregate, when it passes a certain size, undergoes fission or gemmation; so does the secondary aggregate. And as on the lower stage so on the higher, we see cases in which the gemmiparously-produced individuals part as soon as formed, and other cases in which they continue united, though in great measure independent. This massing of secondary aggregates into tertiary aggregates, is variously carried on among the Hydrozoa, the Actinozoa, the Polyzoa, and the Tunicata. In most of the types so produced, the component individualities are very little subordinated to the individuality of the composite mass—there is only physical unity and not physiological unity; but in certain of the oceanic Hydrozoa, the individuals are so far differentiated and combined as very much to mask them. Forms showing us clearly the transition to well-developed individuals of the third order, are not to be found. Nevertheless, in the great sub-kingdom Annulosa, there are traits of structure, development, and mode of multiplication, which go far to show that its members are such individuals of the third order; and in the relations to external conditions involved by the mode of union, we find an adequate cause for that obscuration of the secondary individualities which we must suppose has taken place. The two other great subdivisions, Mollusca and Vertebrata, between the lower members of which there are suggestive points of community, present us only with aggregates of the second order, that have in many cases become very large and very complex. We find in them no trace of the union of gemmiparously-produced individuals. Neither the molluscous nor the vertebrate animal shows the faintest trace of a segmentation affecting the totality of its structure; and we see good grounds for concluding that such segmentation as exceptionally occurs in the one and usually occurs in the other, is superinduced.

[Note:—A critic calls in question the statement on p. 121 respecting the Amphioxus. At the outset, however, he admits that in the Amphioxus “the central nervous system and the notochord are not segmented.” In the Annelid, however, the central nervous system is segmented, and there is segmentation of the part which, as a supporting structure, is analogous to the notochord in respect of function—the outer part which represents the exo-skeleton in contrast to the endo-skeleton. He goes on to say that “the gut is not involved [in the segmentation] and exhibits in Amphioxus just as it does in worms differentiations entirely independent of the segmentation of the mesoblast.” Part of this statement is, I think, not congruous with all the facts. In Protodrilus, one of the lowest of the Archiannelida, “the intestine is moniliform, there being a constriction between each segment” and the next. (Shipley.) Complete segmentation of the intestine is obviously impossible, since, were the canal divided into portions by septa, no food could pass. But the fact that the gut has these successive expansions and constrictions, corresponding to the successive segments, and giving to each segment a partially-separate stomach, shows that segmentation has gone as far as consists with the carrying on of the lives of the segments. No such partial segmentation exists in the Amphioxus. Thus, then, three fundamental structures—the directive structure, the supporting structure, and the alimentary structure—are respectively simple in the lowest vertebrate and segmented, or partially segmented, in the lowest Annelid. Again, while it is said that the gill-clefts exhibit segmentation, it is admitted that this has no relevance to any constitutional segmentation: “they are segmented on a plan of their own” irrespective of other organs. Another allegation is that the ovaries of Amphioxus are segmented. Their segmentation, however, like that of the gills, is isolated, and may be considered as illustrating those repetitions of like parts seen in supernumerary vertebræ in various creatures—a repetition which becomes habitual if the resulting structure is advantageous to the species. On the statement that while the Amphioxus has no rudiments of a renal system the Elasmobranch embryo has such rudiments, which are as distinctly segmented as the nephridia of a worm, two comments may be made. The first is that if in these Vertebrates the nephridia bear a relation to the general structure like that which they do in Annelids, then one would expect to find the segmental arrangement shown in the lowest type, as in Annelids, rather than in a type considerably advanced in development. Should it be replied that in the Amphioxus an excretory system had not yet arisen, though one is required for the higher organization of an Elasmobranch, then the answer may be that since the segmental arrangement in the Elasmobranch corresponds with that of the myotomes, it has no reference to any primordial segmentation, since the myotomes have been functionally generated. The second comment is that whereas the nephridia of the Annelid have independent external openings, the nephridia in the Elasmobranch have not. These discharge their secretions into certain general tubes of exit common to them all; showing that each of them, instead of being a member of a partially independent structure, is united with others in subordination to a general structure. That is to say, the segmentations are far from being parallel in their essential natures. The assertion accompanying these criticisms, that there is “no difference in principle between the segmentation of Amphioxus and Annelid” is difficult to reconcile with the visible contrast between the two. Whatever local segmentations there are in an Amphioxus appear to me quite unlike “in principle” to those which an Annelid exhibits. Could its portion of gut be duly supplied with nutriment, the segment of a low Annelid could carry on its vital functions independently. In the parts of the Amphioxus we see nothing approaching to this. Cut it into transverse sections and no one of them contains anything like the assemblage of structures required for living. The Amphioxus is a physiological whole, and in that respect differs radically from the Annelid, each segment of which is in chief measure a physiological whole. No occurrence of local segmentation in the Amphioxus can obliterate this fundamental contrast.

An accompanying contrast tells the same story. On ascending from the lowest to the highest annulose types we see a progressing integration, morphological and physiological; so that whereas in a low annelid the successive parts are in large measure independent in their structures and in their lives, in a high arthropod, as a crab, most of the parts have lost their individualities and have become merged in a consolidated organism with a single life. Quite otherwise is it in the vertebrate series. Its lowest member is at the very outset a complete morphological and physiological whole, and the formation of those serial parts which some think analogous to the serial parts of an Annelid, begins at a later stage and becomes gradually pronounced. That is to say, the course of transformation is reversed.]