§198. The general conclusion to which these various lines of evidence converge, is, then, that the shoot of a flowering plant is an aggregate of the third degree of composition. Taking as aggregates of the first order, those small portions of protoplasm which ordinarily assume the forms under which they are known as cells; and considering as aggregates of the second order, those assemblages of such cells which, in the lower cryptogams, compose the various kinds of thallus; then that structure, common to the higher cryptogams and to phænogams, in which we find a series of such groups of cells bound up into a continuous whole, must be regarded as an aggregate of the third order. The inference drawn from analysis, and verified by a synthesis which corresponds in a remarkable manner with the facts, is that those compound parts which, in Monocotyledons and Dicotyledons are called axes, have really arisen by integration of such simple parts as in lower plants are called fronds. Here, on a higher level, appears to have taken place a repetition of the process already observed on lower levels. The formation of those small groups of physiological units which compose the lowest protophytes, is itself a process of integration; and the consolidation of such groups into definitely-circumscribed and coherent cells or morphological units, is a completing of the process. In those coalescences by which many such cells are joined into threads, and discs, and solid or flattened-out masses, we see these morphological units aggregating into units of a compound kind: the different phases of the transition being exemplified by groups of various sizes, various degrees of cohesion, and various degrees of definiteness. And now we find evidences of a like process on a larger scale: the compound groups are again compounded. Moreover, as before, there are not wanting types of organization by which the stages of this higher integration are shadowed forth. From fronds that occasionally produce other fronds from their surfaces, we pass to those that habitually produce them; from those that do so in an indefinite manner, to those that do so in a definite manner; and from those that do so singly, to those that do so doubly and triply through successive generations of fronds. Even within the limits of a sub-class, we find gradations between fronds irregularly proliferous, and groups of such fronds united into a regular series.
Nor does the process end here. The flowering plant is rarely uniaxial—it is nearly always multiaxial. From its primary shoot there grow out secondary shoots of like kind. Though occasionally among Phænogams, and frequently among the higher Cryptogams, the germs of new axes detach themselves under the form of bulbils, and develop separately instead of in connexion with the parent axis; yet in most Phænogams the germ of each new axis maintains its connexion with the parent axis: whence results a group of axes—an aggregate of the fourth order. Every tree, by the production of branch out of branch, shows us this integration repeated over and over again; forming an aggregate having a degree of composition too complex to be any longer defined.
[Note.—A criticism passed on the general argument set forth in the foregoing sections, runs as follows:—“I have already pointed out that the process of evolution by which you believe the Liverworts with a distinct axis and appendages to have been produced from the thalloid forms is not founded on sound evidence either in comparative morphology or development. But even if we admit that such an integration of a proliferously-produced colony might have given rise to the leafy Jungermanniaceæ, there are even more weighty objections to the supposition that the same process produced the shoot structures of the flowering plants. In the first place the flowering plant-body is not homologous with the liverwort plant-body, since they represent different generations. The liverwort plant-body or gametophyte, i.e., the generation bearing sexual organs, is homologous with the prothallus of ferns and other Pteridophytes, and in the Flowering Plants with reduced structures contained within the spores (embryo-sac and pollen-grain) but still giving rise to sexual cells. The liverwort spore-capsule and its accessory parts (in fact everything produced from the fertilized egg) is homologous with the sporogonium of the mosses, and, as most botanists think, with the leafy plant-body of Pteridophytes and Phanerogams. This generation is called the sporophyte and from the spores which it produces are developed the gametophytes of the next generation. These generalizations were first established by Hofmeister, and all subsequent work has tended to establish them more firmly. The only doubtful question is (and the doubt is mainly, I think, peculiar to myself, certainly not being shared by the majority of botanists) whether the sporophyte of Mosses and Liverworts is really homologous with that of Pteridophytes and Phanerogams, whether it may not rather be regarded as a parallel development along another line of descent from the Green Algæ.
“Hence we must look for the origin of the shoot-structure of flowering plants in the sporophytes of the Pteridophytes, from which group there is no reason to doubt that the phanerogams have arisen in descent. The various groups of Pteridophytes vary much in the organization of these shoot-systems, as a mental glance at the types exhibited by the Ferns, Horse-tails, Club-mosses, Ophioglossaceæ, and the isolated Isoetes will convince you at once. It may be that some of these groups are independent in descent, i.e., that the Pteridophyta are polyphyletic, and the current hypothesis with regard to the phanerogams is that they have arisen by two, if not three, separate lines of descent from different groups of Pteridophytes (this is indicated in the classificatory diagram on p. 377 of vol. I). I should not, however, care to pin my faith to these or to any such lines of ancestry. Still I think we must look for the ancestors of the Flowering Plants among the Pteridophytes, and the latter always have a good distinction between axis and appendages. The problem of the evolution of these differentiated sporophytic shoots is undoubtedly the great outstanding problem of morphology. Various attempts have been made to solve it, of which probably the most important is the theory of Profs. Bown and Campbell, who derive the Pteridophytes from some Liverwort like Anthoceros, but the sporophyte of course from the sporophytic portion of the plant (not much more than a spore-capsule), the prothallus of the Fern representing the vegetative thallus of Anthoceros. I am not wholly convinced by these undoubtedly ingenious hypotheses, in support of which an immense amount of facts have been collected; but my position would, I know, simply ‘put us to ignorance again’ on this question.
“I have discussed this at some length in order to bring out clearly the immense difficulty of constructing a wellgrounded theory of the origin of the differentiated shoot-system of the higher plant. I confess I don’t think it can be done at all with the materials at present at our disposal. Of course it is just possible to suppose that some ancestral sporophyte had the structure of a proliferous thalloid liverwort gametophyte, and that from it was evolved the phanerogamic shoot in the ways you suggest. This gives us absolutely no clue, however, to any Pteridophytic shoot, which ought to be intermediate (more or less) between the hypothetical ancestor and the Phanerogam, and is furthermore, as far as I can see, not supported by an atom of evidence of any kind. It is true that your theory fits in well with the phenomena exhibited by phanerogamic shoots themselves, but this fact you will see must lose much of its significance if the hypothesis lacks foundation.
“With regard to your method of explaining the fundamental characters of ‘Exogens’ and ‘Endogens,’ this of course is part of the same hypothesis; but I may point out that since Von Mohl and Sanio, between 1855 and 1865, showed (1) that the growth at the stem apex of a monocotyledon was not endogenous, and (2) that the ‘thickening ring’ near the apex of a dicotyledon was not to be confused, as had been done up till then, with the ring of secondary meristem or true cambium, which arose lower down, and only in woody or practically woody stem, the terms ‘Exogen’ and ‘Endogen’ have necessarily fallen into disuse, since they imply a false conception of what happens. Both monocotyledons and dicotyledons have a ‘thickening ring,’ which gives rise to the primary vascular cylinder of the stem. When the stem is of considerable thickness, as in Palms, &c., it grows by the active cell-division of its outer layers, so that both classes are ‘exogenous’ in this sense; while the addition of a centrifugal zone of secondary wood is confined to certain Dicotyledons (Trees, shrubs, &c.).
“The distinction between the embryos, moreover, is not absolute. The single cotyledon is usually terminal in monocotyledons, but not always (Dioscoraceæ have lateral cotyledons), but the plumule may push through it (Grasses) or make its exit sideways (Palms), or be formed at the side (Alisma); and Dicotyledons very similarly.
“The occurrence of completely sheathing leaves in grasses is perhaps correlated with the absence of cambium, but grasses are an aberrant type among monocotyledons, and secondary thickening is only found in very few genera of this class, so that the correlation is, so to speak, negative and indirect.... It is clear that the greater part of the discussion will have to be re-written.”
For the reasons assigned in the preface I cannot undertake to re-write the discussion, as suggested. It must stand for what it is worth. All I can do is here to include along with it the foregoing criticisms.