ing with the egg, the young entoconcha passes through a series of stages, a, b, c, d, e, f, g, h, which are like stages in the development of an ordinary gasteropod, and therefore like stages in the life of its gasteropod ancestor, A; hut, after reaching a certain point, it takes the back track shown by the unbroken line, and, gradually losing the structural complexity which has been acquired, becomes an adult, D, which has reproductive organs, but is, in other respects, as unspecialized as an ordinary gasteropod at one of its earliest embryonic stages, b.
It is obvious that paleontology can give us little help in tracing out such a life-history as this, and we turn to the remaining source of evidence, embryology, to examine how far the facts furnished by this department of life-science can afford a basis for phylogenetic generalizations.
The case which we have just examined shows that the embryology of two related forms may be essentially the same, since both of them have inherited the greater part of their life-history from a common parent, and it would seem at first sight as if all that we need, to enable us to trace out the relationship of all living animals, is a complete acquaintance with the embryology and metamorphosis of each one of them. A comparison of all the stages in the life of one species with all the stages in the life of another species of the same genus ought to show essential identity; and a comparison of the stages of development of the species of one genus with those of the species of a related genus ought to show how far their history has been the same: the common features in the embryology of two allied families should show how far the history of the species in one of them has been the same as that of the species in the other, and so on, each wider and wider comparison showing broader and broader relationships, until the features which are common to the embryos of all animals unite them into one great group.
As this may be clearer in a more abstract shape, I will try to state it in the form of a diagram (see Fig. 4).
Suppose that, in studying the development of four species, 1, 2, 3, 4, we find that 1 passes through a series of stages, a, b, c, d, e, f, g, h; that 2 presents the series a, b, c, d, e, f, i, j, k; while 3 passes through the stages a, b, c, l, m, n, o, p, q; and 4 through the stages a, b, c, l, m, r, s, t, u. A comparison of these four life-histories would indicate that their common relationships are such as are represented by the four branched tree shown in Fig. 4. We have already seen that it is perfectly possible that n or c or e may not have been an adult animal, but simply a stage in the development of an unknown adult, x; so there would not be much chance of finding m or e or c as a fossil, and the embryonic record would not show us what the common ancestor of 1 and 2 or the more remote ancestor of all four species actually was, but would simply show that they are related in this way; but it would
