Interpretation of the forms of organisms and the forms of their parts, must depend mainly on the conclusions previously drawn respecting their phylogeny; and the drawing of such conclusions must be guided by recognition of the various factors of Evolution, as well as by recognition of certain extremely general results of Evolution and certain concomitants of Evolution. A primary one among these is that no existing species can exhibit more than approximately the ancestral structure of any other existing species. As all ancestors have disappeared, so, in a greater or less degree, the traits, specific, generic, or ordinal, which distinguished the earlier of them have disappeared. Setting out with the familiar symbol, a tree, let us regard its peripheral twigs as representing extant species; let us assume that the interior of the tree is filled up with some supporting substance, leaving only the ends of the living twigs projecting; and let us suppose the trunk, main branches, secondary branches, tertiary branches, &c., have decayed away. Then if we take these decayed parts to stand for the divergent and re-divergent lines of evolution which are represented by fossils in the Earth’s crust, it will be manifest, first, that no one of the living superficial twigs (or species) exhibits the ancestral organization whence any other of the living superficial twigs (or species) has been developed; it will be manifest, second, that the generic structure inherited by any existing species must be a structure out of which came sundry allied species--the fork, as it were, at which adjacent twigs diverged; and third, that the ancestor of an order must, in like manner, be sought at some point deeper down in the symbolic tree--a place of divergence of the sub-branches representing allied genera. Similarly with the ancestral types of classes, still deeper down in the tree or further back in time. So that phylogeny becomes more and more speculative as its questions become more and more radical. And the difficulty is made greater by the deficiency of palæontological evidence. One obvious corollary is that an ancestral type from which sundry allied types now existing diverged, was, speaking generally, simpler than these; since the divergent types became different by the superposing of modifications, adding to their complexities. There is a further reason for inferring that the least specialized member of any group is more like the remote ancestor than any of the others; for every adaptation stands in the way of subsequent re-adaptations: it presents a greater amount of structure to be undone. To get some idea of the ancestral type where no extant member of the group is manifestly simpler than the rest, the method must be to take all its extant members and, after letting their differences mutually cancel, observe what remains common to them all.