6 ARKIV FÖR ZOOLOGI. BAND 10. NIO 8. 



the vertebrae of Spheroides does not seem to support the 

 latter theory. If, as Schaunisland considers to be probable, 

 the skeleton-pieces in the ligament are identical with rudi- 

 mentary cranial arches displaced dorsally, then the dorsal 

 arches of the vertebrae of Spheroides must be formed only 

 by the original caudal ones. As will be shown below, how- 

 ever, there is reason to beheve that each of the dorsal arches 

 of Spheriodes must be formed by the coalescence of two 

 original arches, one cranial and one caudal. What is then 

 the nature of these skeleton-pieces? It seems to me most 

 probable that they are distinct skeleton-portions, developed 

 in that part of the primary »membranous» vertebrae which 

 forms the wall between the neural canal and the so-called 

 dorsal spinal canal. The primary »membranous» vertebrae 

 consists of skeletogeneous tissue, and skeleton elements can 

 be developed in all its portions. We know that the ossifi- 

 cation takes place from the points that are called dorsal and 

 ventral arches. It begins with the formation of cartilage 

 pieces, which låter on ossify. From the original arches the 

 ossification continues in the perichordal tissue and in the 

 »membranous» neural arch so that at last a more or less 

 completely bony vertebra is formed, in which the originally 

 distinct portions have completely coalesced. I think that the 

 skeleton-pieces whose nature we are now discussing are of a 

 distinct origin just as the arches. They coalesce, however, 

 with the dorsal arches. The question, however, is not settled 

 by this statement. Scheel maintains that the primitive ver- 

 tebra has one dorsal arch on each side and that the neural 

 canal of such a vertebra is closed by a horizontal roof con- 

 necting the both arches. This roof lies ventrally to the dor- 

 sal spinal canal. The fusion of the dorsal arches dorsally to 

 this canal is a secondary character, which, as he writes, p. 

 23, »vielleicht durch die grössere Entwickelung der dorsalen 

 Muskulatur bedingt worden ist.» This theory seems to me 

 rather plausible. The skeleton-pieces in question are perhaps 

 to be homologized ivith the so-called »Deckplaiten öder Schluss- 

 stilcke» of Chimaera. 



Thilo has shown that the number of vertebrae with a spina 

 bifida varies a good deal. The greatest number of such verte- 

 brae or in other words the greatest length of the spina bi- 

 fida, is found in such forms as are able to inflate the body 



