UROSTYLE OP THE ANUROUS AMPHIBIA. 241 



was a well-marked furrow. This furrow, which was deepest 

 anteriorly, began at a point just in front of the place where the 

 neural canal appeared to reach the dorsal surface (text-fig. l,a,a). 



The passage of a bristle definitely established the further fact 

 that the neural canal in the specimen studied did not end blindly, 

 but, near the end of the first third of the length of the bone, 

 opened out into this groove or furrow upon the dorsal surface. 

 Thence it was continued nearly to the end of the middle third of 

 the urostyle as the open groove referred to (text-fig. l,f.f.t.). 

 In the fresh condition this groove is occupied (as I afterwards 

 determined by means of serial sections) by the extremity of the 

 filum terminale, which thus has a position precisely similar to 

 that which it occupies in the young R. iemporaria. The open 

 groove, extending approximately along the middle third of the 

 dorsal surface of the urostylar crest, is thus a posterior uncovered 

 extension of the neural canal. 



An examination of a large number of urostyles (34) of specimens 

 of li. tigrina showed that this terminal opening of the neural 

 canal was an absolutely constant feature, occurring even in 

 specimens showing marked abnormality of the vertebral column. 



A point of difference noted, however, was that in some speci- 

 mens the dorsal furrow was of much greater extent than in 

 others. In all it ended at approximately the same point, but in 

 some the furrow began nearly at the anterior extremity of the 

 bone (text-fig. 1, b, d, d'), In some instances the dorsi-ventral 

 diameter of the neural canal was unusually large, and the canal 

 then extended to the dorsal margin of the urostyle at a point much 

 nearer to its anterior end (text-fig. 1, b). In other examples the 

 neural canal, though not of larger calibre than usual, passed 

 dorsalwards much more abruptly. In these latter the dorsally 

 situated (exposed) groove for the filum terminale was, relatively, 

 much longer. A condition, varying in a manner precisely oppo- 

 site to this, also occurs in which the filum terminale passes 

 backwards in a direction much more nearly approaching the 

 horizontal. In this case the uncovered extremity of the neural 

 canal is, relatively, extremely short (text-fig. l,c). In those 

 examples, however, in which the neural terminal filament runs 

 for a considerable distance in this groove upon the upper margin 

 of the urostyle, it is not always uncovered for the whole of that 

 distance. Not infrequently a delicate layer of bone roofs in a 

 portion of the groove anteriorly. In a single specimen there was 

 a small dorsal aperture leading into the neural canal and situated 

 far anteriorly, separated from the open stretch of groove by an 

 intervening bony roof (text-fig. 1, d, d'). 



Such variation in the extent of the dorsal open groove may 

 perhaps indicate that the number of fused vertebrae in this region, 

 in which the neural arch is incomplete, is quite variable. Behind 

 the furrow the urostyle must, presumably, be regarded as com- 

 posed of fused centra only. 



On account of the severity of the drought prevailing at the 



