560 



The next stage is illustrated in figs. 2—5. These figures are 

 from cross sections through an embryo about 8 mm in length, and 

 show the beginning of the thyroid outgrowth from the floor of the 

 alimentary canal. The upper boundary of the figures is formed by 

 the floor of the alimentary canal continued to the ectoderm in the 

 hyobranchial cleft. Although this cleft is less] inclined to the 

 transverse axis of the embryo than the hyomandibular , the dorsal 

 part of the cleft nevertheless falls in sections anterior to those that 

 pass through the ventral part, as shown by the difi'erence in the 

 relative height of the cleft in figs. 2 and 5, which are separated by 

 eleven sections. Fig. 5 passes through the ventral margin of the 

 right cleft. The actual difference in height is rather greater than the 

 figures indicate, since the ventral surface of the embryo also descends 

 in passing from the plane of fig. 2 to that of fig. 5. 



Between the stage first described and that now given, the lateral 

 extensions of the oral fusion , outlined in fig. 1, have appeared. In 

 passing outwards these lateral wings also extend upwards. The curved 

 anterior margin of the oral fusion follows the posterior surface of the 

 infundibulum , which the cranial flexure forces against the anterior 

 wall of the alimentary canal. 



The development of the mouth in Necturus seems to me to 

 furnish as striking an argument in favor of its branchial origin as 

 the paired oral invaginations found in the Teleostei. The hyo- 

 branchial clefts are primarily incKned to the transverse plane as 

 above described, the hyomandibular are still more inclined, and the 

 oral clefts appear to be absolutely fused in a horizontal plane in 

 consequence of the horizontal position in which the mandibular arches 

 first lie. Such a supposition readily explains the shape of the oral 

 fusion outlined in fig. 1. If, however, the mouth is merely a median 

 transverse opening to the alimentary canal, as it frequently appears 

 to be, it is difficult to understand why in Necturus the fusion 

 between ectoderm and endoderm should primarily extend to the 

 posterior margin of the space outlined. A ventral fusion of the 

 branchial clefts is not unique. The rudimentary hyomandibular pockets 

 fuse with one another in Necturus, below the hyoid outgrowth into 

 the oral cavity which forms the base of the tongue, and the hyo- 

 branchial clefts also fuse with one another ventrally forming a kind 

 of operculum of the overlying hyoid tissue. It is consequently far 

 from unreasonable to suppose that a similar fusion united a pair of 

 oral clefts in the median mouth, as Dohrn ') long ago suggested. 



1) A. DoHEN, Der Mund der Knocheufische. Mitteilungen aus 

 der Zool. Station zu Neapel, Bd. II, 1881. 



