"The complete and accurate repetition of phyletic by biontic development is falsified and altered by secondary adaptation, in that the bion[372] during its individual development adapts itself to new conditions: accordingly the repetition is the more accurate the greater the resemblance between the conditions of existence under which respectively the bion and its ancestors developed" (ii., p. 300).
The last two propositions, it will be observed, are taken over almost verbally from F. Müller.
Now we have seen that the natural system of classification gives a true picture of the genealogical relationships of organisms, that the smaller and larger classificatory groups correspond to greater or lesser branches of the genealogical tree. If ontogeny is a recapitulation of phylogeny, we must expect to find the embryo repeating the organisation first of the ancestor of the phylum, then of the ancestor of the class, the order, the family and the genus to which it belongs. There must be a threefold parallelism between the natural system, ontogeny and phylogeny (ii., pp. 421-2).
It will be observed that there is here implied an analogy between the biogenetic law and the law of von Baer, for both assert that development proceeds from the general to the special, that the farther back in development you go the more generalised do you find the structure of the embryo; both assert, too, that differentiation of structure takes place not in one progressive or regressive line, but in several diverging directions.
But the analogy between the biogenetic law and the Meckel-Serres law is even more obvious, and the resemblance between the two is much more fundamental. It is a significant fact that in his theory of the threefold parallelism Haeckel merely resuscitated in an evolutionary form a doctrine widely discussed in the 'forties and 'fifties,[373] and championed particularly by L. Agassiz,[374] a doctrine which must be regarded as a development or expansion of the Meckel-Serres law.[375] It is the view that a parallelism exists between the natural system, embryonic development, and palæontological succession. Actually, as Agassiz stated it, the doctrine applied neither to types, nor as a general rule to classes, but merely to orders. It was well exemplified, he thought, in Crinoids:—"The successive stages of the embryonic growth of Crinoids typify, as it were, the principal forms of Crinoids which characterise the successive geological formations. First, it recalls the Cistoids of the palæozoic rocks, which are represented in its simple spheroidal head; next the few-plated Platycrinoids of the Carboniferous period; next the Pentacrinoids of the Lias and Oolite with their whorls of cirrhi; and finally, when freed from its stem, it stands as the highest Crinoid, as the prominent type of the family in the present period" (p. 171).
The Meckel-Serres law, it will be remembered, expressed the idea that the higher animals repeat in their ontogeny the adult organisation of animals lower in the scale. Since Haeckel recognised clearly that a linear arrangement of the animal kingdom was a mere perversion of reality, and that a branching arrangement of groups more truly represented the real relations of animals to one another, he could not of course entertain the Meckel-Serres theory in its original form. But he accepted the main tenet of it when he asserted that each stage of ontogeny had its counterpart in an adult ancestral form. Such ancestral forms might or might not be in existence as real species at the present day; they might or might not be discoverable as fossils. That they had real existence either now or at some past epoch Haeckel never doubted. In his construction of phylogenetic trees he was so confident in the truth of his biogenetic law that he largely disregarded and consistently minimised the importance of the evidence from palæontology.
The biogenetic law differed from the Meckel-Serres law chiefly in the circumstance that many of the adult lower forms whose organisation was supposed to be repeated in the development of the higher animals were purely hypothetical, being deduced directly from a study of ontogeny and systematic relationships. The hypothetical ancestral forms which the theory thus postulated naturally took their place in the natural system, for they were merely the concrete projections or archetypes of the classificatory groups.
The transcendentalists, of course, conceived evolution, whether real or ideal, as a uniserial process, whereas Haeckel conceived it as multiserial and divergent. It is here that the superficial agreement of the biogenetic law with the law of von Baer comes in.
We might almost sum up the relation of the biogenetic law to the laws of von Baer and Meckel-Serres by saying that it was the Meckel-Serres law applied to the divergent differentiation upheld by von Baer instead of to the uniserial progression believed in by the transcendentalists.
How near in practice Haeckel's law came to the recapitulation theory of the transcendentalists may be seen in passages like the following, with its partial recognition of the Échelle idea:[376]—"As so high and complicated an organism as that of man ... rises upwards from a simple cellular state, and as it progresses in its differentiating and perfecting, it passes through the same series of transformations which its animal progenitors have passed through, during immense spaces of time, inconceivable ages ago.... Certain very early and low stages in the development of man, and other vertebrate animals in general, correspond completely in many points of structure with conditions which last for life in the lower fishes. The next phase which follows on this presents us with a change of the fish-like being into a kind of amphibious animal. At a later period the mammal, with its special characteristics, develops out of the amphibian, and we can clearly see, in the successive stages of its later development, a series of steps of progressive transformation which evidently correspond with the differences of different mammalian orders and families."[377]