256 s. WATASE 



will be observed that the hindmost process (e) is single in the former 

 and double in the latter (e, e'). 



Figs. 12 and 13 are two corresponding bones, the fourth hy- 

 pural bone, from a single.; and a double form of the tail respectively. 

 The former is single whereas the latter consists of two diverging 

 limbs. Its apical portion lies wedged in between the two pieces of 

 urostyle. 



In the tail fin with its ventral edge furrowed by a shallow 

 groove, (Figs. 8 & 9), the heemal spines from the penultimate and 

 antipenultimate centra proceed from their origin in a distinctly paired 

 state in order to support the paired portion of the tail. In other 

 cases in which the groove is deeper, those skeletal supports corres- 

 ponding in position to the groove occur in pairs. 



Figs. 19 — 23 represent diagramatic cross-sections of the several 

 forms of the tail-fin. 



Fig. 19 shows the arrangement of bones in a vertical and normal 

 tail. On comparing this figure with others it will readily be seen 

 that the difference in) the depth of the ventral groove affects the bony 

 structures in corresponding degrees. 



The paired state of hypural bones in gold-fishes is of no small 

 interest in connection with the view that the hypural elements are 

 homologous with the more anteriorly situated subvertebral appen- 

 dages, and that the styliform urostyle is a coalesced representative 

 of a certain number of vertebral bodies. On comparing the sub- 

 vertebral appendages in various'; regions of the body, we observe that, in 

 the trunk, tlie appendages start in pairs (transversal pro'cesses) and 

 end in pairs; that, in the caudal region, they start likewise in pairs 

 (haemal arch) which however soon coalesce and terminate in a median 

 process (hœmal spines); and that in the caudal extremity, the ap- 

 pendages start from the beginning in a coalesced state as median 



