two stages of development ultimately diverged so widely in structure that the transition from one to the other could no longer be made at a single moulting, and a period of rest had to be interpolated, in order that the transformation of the body could take place. In this way arose the stage of the resting and fasting pupa, a 'cœnogenetic' modification of the last larval stage, _not a recapitulation of an ancestral form_, but a stage which has been interpolated, or better, has 'interpolated itself' into the ontogeny on account of the widely different adaptations of the early and the final stages. This is a perfectly clear idea, and Haeckel's distinction between palingenesis and cœnogenesis is undoubtedly justified. But it is quite a different matter to be able to decide whether a particular stage or organ has arisen palingenetically or cœnogenetically with the same certainty as in the case of the insect-pupa, or even with any degree of probability, and we must admit that in very many cases, perhaps even in most cases, it is impossible. This is so chiefly because pure palingenesis is hardly likely to occur now; the ancestral stages were bound to be modified in any case if they were to be compressed into the ever-shortening ontogeny of later descendants, and particularly so if they were to be shunted back into embryogenesis. In the latter case they would not only be materially shortened, and, as I have already shown, modified by the mutual adaptations of the different developing parts, but time-displacements of embryonic parts and organs would be necessary, as has been very clearly proved by the excellent recent investigations, which we owe in particular to Oppel, Mehnert, and Keibel. A shunting forward or backward of the individual organs takes place--conditioned apparently by the decreasing or increasing importance of the organ in the finished state; for in the course of the phylogeny everything may vary, and not only may a new, somewhat modified, and often more complex stage be added on at the end of the ontogeny, but each one of the preceding stages may vary independently, whenever this is required by a change in its relations to the other stages or organs. Adaptation is effected at every stage and for every part by the process of selection, for all parts of the same rank are ceaselessly struggling with one another, from the lowest vital units, the biophors, up to the highest, the persons. If we reflect that, in the course of the phylogeny of every series of species, a number of organs always become superfluous and begin to disappear in consequence, we can understand what great changes must take place gradually as such a series of phyletic stages is compressed into the ontogeny, for all organs which are no longer used are gradually shifted further and further back in the ontogeny till ultimately they disappear from it altogether. But, while the primary constituents of these 'vestiges' play their part in ontogeny for a