372 



CHAS. W. HARGITT. 



In Figs. 3-5, as well as in those already cited, are shown 

 interesting and more or less common forms of variant canals. 

 Similar conditions are also to be distinguished in several of the 

 other figures given. For convenience we may discuss these 

 features under the following heads: (i) Bifurcations, (2) loops, 

 (3) spurs, (4) anastomoses. 



It may be altogether probable that these conditions graduate 

 somewhat insensibly into each other, so that they may really be 

 but varying expressions of one and the same fundamental process 

 But as suggested, such a classification will at least serve as. a con- 

 venience in description. 



I. Bifurcations. Several illustrations of this feature have 

 already been cited. Others are shown in Figs. 46. For the 

 most part these apparently arise as a result of division of the 

 peripherally developing canal, though the centripetally directed 

 type is not lacking, as is seen in Fig. 4. It may be said in passing, 

 that the figures given are for the most part typical ones, each 

 representing usually several similar cases. In development the 



FIG. 4. Showing centripetal and 

 centrifugal canals with bifurcations. 



FIG. 5- Showing union of per- 

 ipheral portions of two canals. 



canals usually arise at the gastric pouch and grow toward the 

 periphery. Agassiz (of>. cit.~) has said upon this point, "Additional 

 chymiferous tubes in the /Equoridse are developed from the 

 digestive cavity, as shown by Kolliker, and not from the vertical 

 tube, as in the case of the branching tubes of IVillia." This I 

 have found to be the case in a study of the very young medusa. On 



