quite different when the younger stages are compared, and at first sight no one would suspect that the Sergestes larva (Fig. 7) and the Lucifer larva (Fig. 8) are corresponding stages in the development of two animals as similar to each other as those shown in Figs. 5 and 6. Not only do we find animals whose young stages differ more than 
Fig. 8. the adults, but we also meet cases—and they are very numerous indeed—where the order of appearance of organs and features of the greatest taxonomic importance differs in the embryos of closely related forms.
To take a particular instance, it is plain that, since the features which all the two-gilled cephalopods have in common, and which are characteristic of the group as a whole, must have been inherited from the common ancestor of the whole group, they ought, unless the embryonic history of the different recent species has undergone secondary modifications, to appear in the same order in the embryos of all the existing forms; and, if they do not, it is clear that descriptive embryology alone can not furnish a key to systematic affinity.
As a matter of fact, each one of the three species of two-gilled cephalopods with whose embryology we are most familiar differs from, both the others in the order in which such significant organs as the arms, the shell, the eyes, the siphon, the gills, and the mouth make their appearance; and it must be obvious that, unless we have some means of analyzing these three life-histories, and determining which of them gives the true ancestral order, we can not make use of their embryology as a key to phylogeny. One who is not familiar with the whole field of life-science may fairly ask how it is possible to discover the relationships of animals from the study of their embryology if it is true that the early stages in the life of closely related species may differ so greatly, and if it is true that the order and manner in which structures make their appearance in the embryo are not
