PATTEN— COOPERATION AS A FACTOR IN EVOLUTION. 623 



unless a new opening into the alimentary canal were already available 

 elsewhere. Such an opening is available in what I have called the 

 cephalic navel, a temporary embryonic opening into the enteron, 

 lying on the hremal surface of the embryo in a region corresponding 

 to that where the vertebrate mouth is located. 



What has probably taken place, then, is this : in the vertebrates the 

 old invertebrate oesophagus (Fig. 5, A) has been gradually choked 

 up by a vigorously growing nervous system and its external opening 

 competely enclosed within the brain chamber. A new entrance to 

 the alimentary canal was then established through an old channel, 

 of unknown significance, in a more convenient place. The remnants 

 of this old, shut-in mouth and oesophagus are still conspicuous 

 features (otherwise inexplicable) in the brain of all vertebrates, 

 /'. c, the infundibulum, the saccus vasculosus, and the large opening 

 in the roof of the fourth ventricle, now closed by a thin membrane^ 

 the choroid plexus (Fig. 5, 5). 



The position of the jaws is also changed by the same cause, for 

 the great size of the embryonic forebrain, at an early period, lifts the 

 head of the embryo off the surface of the egg, and forces the jaws, 

 or oral arches, apart, toward the haemal surface, where they con- 

 verge around the new oral opening (Fig. 4, 31-34). At least three 

 pairs of arches are involved in this movement, and the important 

 steps in the process may still be observed in many vertebrates. 



The transfer of all three pairs of oral arches to the haemal sur- 

 face may be readily observed, although heretofore overlooked, in 

 frog embryos (Fig. 6). Their ultimate union, along an elongated 

 median depression, gives rise to the fronto-nasal process, the maxil- 

 lary and mandiblar arches, and gives us the key to the morphology 

 of the facial region in all the higher vertebrates. 



Similar conditions may be seen in the adult stages of a very 

 primitive living vertebrate, Myxine (Fig. y,E), and also in the 

 Ostracoderms, which form the connecting link between the giant 

 sea scorpions and the true fishes. In Bothriolepis (Fig. 7, A') both 

 the maxillary and mandibular arches are provided with bony plates 

 located on the ha?mal side of the body and which work against each 

 other in a transverse direction, thus furnishing an instructive tran- 



