the adult, replaced by almost smooth ridges. Now, Semon in his beautiful studies on the development of Neoceratodus[1] has shown that the teeth of this fish at one stage in ontogeny, are represented by rows of denticles even more discrete than the denticles in the Devonic Dipterus; then the denticles gradually merge at their bases, the separate cusps, however, still showing—a stage comparable with the Carboniferous Ctenodus; then they merge still more and assume the ridge-like form seen in the adult Neoceratodus.
Another example: In many sharks the alimentary canal is longer in the embryo than in the adult, the anal opening being situated near the posterior end of the trunk. From such cases one is inclined to believe that in the ancestral sharks this must have been the condition in the adult form; that is to say, the anal opening probably was near the posterior termination of the trunk. We may therefore ask: are there any early fossil sharks which show such a condition? Recently Professor Dean has described[2] a remarkable specimen of Cladoselache from the Upper Devonic of Ohio which seems to indicate such a condition. In this specimen remnants of both kidneys are preserved. They extend in the posterior half of the fish and by their direction indicate that they were drawn together, toward their external opening, not far from the posterior termination of the trunk. This shows that the anal opening in this ancestral shark was very much as in the early shark embryo to-day.
In conclusion perhaps I may venture to make one other point in regard to this question. A vast amount of skepticism concerning the doctrine of recapitulation is to be found in the literature of to-day; and if we study the reasons for this skepticism we find that it is in some measure justified. It is clearly established that among vertebrates as well as among invertebrates there are many examples of structures appearing during embryonic growth which are identical with structures found in the adult of some remote ancestor. But when we reflect on the amount of adaptation which any embryo has undergone in its long evolutional history; when we remember how palingenetic characters are on every hand overlaid by cenogenetic ones; who will say that recapitulation is a principle of general application, or that it is safe to draw conclusions from all embryos concerning their long extinct ancestors? Who will believe that a bony fish which runs through its embryonic development in a few days repeats its ancestral history, when we see at every stage of its ontogeny how it has been adaptively modified for this and for that special need? Only when series of related forms have certain onto
