THE ANCESTRY OF THE VERTEBRATES 517 



ippear. In vertebrates, however, the reverse is true; the 

 nervous system is laid down, the brain is differentiated, the 

 notochord is formed, even the special sense-organs ap- 

 pear, and still the alimentary canal remains a sealed cavity, 

 without communication with the exterior. At last a mouth 

 appears, placed very far ventrally, in line with the gill-slits, 

 and in certain fishes appears first as two lateral openings which 

 eventually become confluent. All this seems a complete cor- 

 roboration of the fact arrived at independently through adult 

 anatomy that the vertebrate mouth has resulted from the con- 

 fluence of a pair of gill-slits anterior to those now functioning, 

 and still equipped \vith gill-arches which serve as jaws. There 

 comes, then, the inevitable conclusion that previous to this con- 

 version, and while the mandibular slits were still functioning as 

 gill-slits, the ancestral forms must have had another mouth, 

 traces of which are to be looked for in the earlier embryo. 

 Such a primary mouth is actually found indicated in precisely 

 the place where it would be looked for in the annelid, taking the 

 reversal of the body into account, and this indication appears 

 in the widely open " fourth ventricle " of the nerve cord. 



In annelids, as in all Articulata, the mouth is upon the ven- 

 tral side, and, since the alimentary canal is dorsal to the nerv- 

 ous system, this position is reached by means of an oesophagus, 

 which turns downwards almost at right angles to the remainder 

 of the canal, and runs between the two nerve connectives that 

 connect the first and second pairs of ganglia. The first gangli- 

 onic pair thus becomes the supra-cesophageal, the second, 

 infra-acsophageal, and the connections between them form a 

 circiun-ccsophageal ring through which the oesophagus passes. 

 These relationships will be clearly seen by reversing the accom- 

 panying figure (Fig. 140), which will thus give the conditions 

 as seen in annelids. In all true vertebrates the actual external 

 opening of this early mouth has disappeared, but it may be 

 identical with the neuropore in the embryo of Amphioxus, 

 which forms in this place a direct communication between the 

 lumen of the neural tube and the exterior and is otherwise un- 

 accounted for. Aside from the indications of the early mouth 



