298 Evolution of the Vertebral Column of Fishes. [June 21, 



each colony or cluster of cells developing into a separate piece of car- 

 tilage. The basidorsal does not fuse with its interdorsal, because 

 both are the offspring of two different sklerotomes, nor can the basi- 

 dorsal fuse with its own offspring, namely, with the interventral, 

 because both became, and remain, separated by the chorda and its 

 sheath ; they are connected only by the indifferent connective tissue 

 of the membrana reunions, but not by cartilage-forming cells. 



Concerning the formation of centra or bodies of the vertebras, we 

 distinguish : 



I. Chorda-centra, i.e., centra cut out of the full of the chordal 

 sheath, which itself has been strengthened by invasion of cartilagin- 

 ous cells from the skeletogenous layer. This migration of cartilage 

 into the chordal sheath had already been hinted at by Kcelliker more 

 than thirty years ago ; it has recently been proved by Klaatsch, and 

 has been corroborated by us. Chorda-centra are possessed by all 

 Elasmobranchs, potentially by Dipnoi and Holocephali. 



II. Arch-centra, i.e., centra formed by the skeletogenous mass 

 which remains entirely on the outside of the chordal sheath, which 

 latter takes no share in their formation : osseous Ganoids and 

 Teleostei. 



Chorda-centra and arch-centra represent two different modes of 

 development, each starting from an acentrous condition. This can 

 be expressed as follows : 



Chordal sheath remaining Chordal sheath strengthened by invasion of 



entirely chordagenous. skeletogenous cells, therefore with 



possibility of chorda-centra. 

 Cyclostomata, 

 Cartilaginous Ganoids. Dipnoi and Holocephali. 



Formation of Centra. 



Osseous Ganoids, Teleostei. Elasmobranchs. 



ARCH-CENTRA. CHORDA-CENTRA. 



The formation of chorda-centra being independent of the arcualia 

 explains how and why the number of " centra " does not necessarily 

 agree either with that of the arcualia or with that of the trunk- 

 segments, e.g., Hexaiichus and tail of most other Elasmobranchs. 



These leading differences and their modifications have been traced 

 in Petromyzon, Acipenser, Amia, Lepidosteus, Protopterus, Chimasra, 

 and in numerous Elasmobranchs. 



In Amia calva, of which the adult and a young specimen of 57 mm. 

 were examined, the postcentrum, i.e., the posterior, archless disk of a 

 complete tail-vertebra, was found to be formed by the interdorsalia 

 and interventralia of the same sklerotome, while the precentrum, i.e. 

 the arch-bearing disk or anterior half is formed by the basidorsals of 



