The Columella Aiiris in Amphibia. 617 



ceratohyal did not ''demonstrate that the cartilage in question may 

 not be a rudimentary element of the hyoid arch, since each element 

 of the cartilaginous arches arises from an independent center of chon- 

 drification and secondary fusions of cartilaginous elements do not 

 necessarily show original association." 



The cord of cells which extends to the under surface of the squa- 

 mosum and which must be regarded as a part of the columella blas- 

 tema, is joined by an extension from the perichondral cells of the 

 ceratohyal, so that, at an early stage (19 mm.) the two structures 

 are in fact connected (Kingsbury '03, p. 318, and Fig. 2a.). In 

 Ambystoma there is a line of cells from the ceratohyal to the colum- 

 ella in the 13-14 mm. specimen. In Plethodon (Text Fig. 6) the 

 juxtaposition is still closer and a cellular continuity exists. In 

 Cryptobranchus the ceratohyal is not joined to the columella by cells 

 in the just hatched larva; just what the condition is at an earlier 

 stage we cannot say from lack of material at a suitable stage. 



This early connection by cells does not seem a chance association, 

 but rather to indicate that columella and ceratohyal chondrify out 

 of a common blastema. The Ligamentum hyo-columellare previously 

 described as existing in several forms seems to be a secondary devel- 

 opment, though it may express a primitive relationship. The dia- 

 gram introduced to illustrate the ligaments (Fig. 21) may also serve 

 to indicate the skeletal connections and relations of the columella 

 as a probable hyomandibulare. 



The position of the columella above and behind the nervus facialis 

 satisfies the requirements of the hyomandibular homology. It is 

 but necessary to recall its position in elasmobranchs and Polypterus 

 (Ruge '96, Pollard '92) to recognize this fact. No detailed study 

 has been made of the relation of this nerve in fishes, however; nor 

 has the significance of the rather constant position of the columella 

 between the vena petroso-lateralis and the arteria carotis interna been 

 investigated. 



The connections and development of the columella do, we believe, 

 strengthen its homology with the Hyomandibulare of fishes, in sup- 

 port of which so much has been written. 



