This justifies the inference that the primitive vertebrate ancestor had a pair of long fins running along the sides of the body, but bending slightly downward toward the rear so as to meet one another and continue as a single caudal fin behind the anal opening.  Such fins, like the feathers of an arrow, could be useful only to keep the animal “on an even keel” as it was forced through the water by the lateral sweeps of the tail.  They would have been useless for creeping.

But there is another piece of evidence that he was a free swimming form.  All vertebrates breathe by gills or lungs, and these are modified portions of the digestive system, of the walls of the oesophagus, from which even the lung is an embryonic outgrowth.  Now practically all invertebrates breathe through modified portions of the integument or outer surface of the body, and their gills are merely expansions of this.  In the annelid they are projections of the parapodia, in the mollusk expansions of the skin, where the foot or creeping sole joins the body.  Why did the vertebrate take a new and strange, and, at first sight, disadvantageous mode of breathing?  There must have been some good reason for this.  The most natural explanation would seem to be that he had no projections on his outer surface which could develop into gills, and farther, that he could not afford to have any.  Now projections on the lower portion of the sides of the body would be an advantage in creeping, but a hindrance in any such mode of swimming as we have described, or indeed in any mode of writhing through the water.

Furthermore, if he lived, not a creeping life on the bottom, but swimming in the water above, he would have to live almost entirely on microscopic animals and embryos; and these would be most easily captured by a current of water brought in at the mouth.  The whole branchial apparatus in its simplest forms would seem to be an apparatus for sifting out the microscopic particles of food and only later a purely respiratory apparatus.  Moreover, we have seen that the parapodia of annelids naturally point to the development of an external skeleton, for their muscles are already a part of the external body-wall and attached to the already existing horny cuticle.  The logical goal of their development was the insect.

Now I do not wish to conceal from you that many good zooelogists believe that the vertebrate is descended from annelids; but for this and other reasons such a descent appears to me very improbable.  It would seem far more natural to derive the vertebrate from some free swimming form like the schematic worm, whose largest nerve-cord lay on the dorsal surface because its branches ran to heavy muscles much used in swimming.  Later the other nerve-cords degenerated, for such a degeneration of nerve-cords is not at all impossible or improbable.  “No thoroughfare” is often written across paths previously followed by blood or nervous impulses, when other paths have been found more economical or effective.