304 THE AMERICAN NATURALIST [Vol. XL 



large dorsal process. Even in the third caudal, these processes 

 (Figs. 6 and 8, D., P. A.) have the appearance of being mere 

 articular surfaces on the sides of the spinous process. The 

 anterior articular processes (Figs. 6 and 8, A.) also diminish in 

 size towards the posterior until, on the last few vertebrse, they, 

 like the posterior articular processes, practically disappear, and 

 the vertebrae in this region are united by their centra only. 



The haemal arch, which is characteristic of the caudal vertebrae, 

 is well developed on the third vertebra (Fig. 8, //. A.), in which 

 the two sides of the arch are prolonged ventrally into a long spine 

 (7/.). As may be seen in Fig. 8 the canal formed by the haimal 

 arch is larger than the spinal canal; this is true of all the caudals 

 except the first, which is without a haemal arch. 



The hinder caudal vertebrae, besides being laterally compressed, 

 show an actual concavity in each side of their centra. 



The two cartilaginous rudiments at the end of the tail, that 

 are described by Osawa, I have not been able to determine, 

 though I am not willing to say that they do not c\i-.t in the Amer- 

 ican as well as in the Japanese form. There is, however, except 

 in the number of vertebrae, a strong resemblance between the 

 vertebral columns of the two animals. 



