154 



Helen KeUey : 



A little later, 51-52 hours, the two canals have become smaller 

 thoiio-h still unequal in size, as seen in text fig. III., 2, 3 and 4. 

 The larger of the two {cj) extends distinctly further back than 

 the other, diminishing in size as it does so. At the extreme 

 anterior end of this subdivision the septum between these two 

 canal (cj and c,) becomes very thin, and just at this part the 

 larger canal (cf. plate XLII., fig. 3, and text fig. III., 5) shows a 

 distinct though very thin septum, resulting in an incipient 

 division into three canals. 



> -» *' <■ 1 



Fm. A: 



The complete division posteriorly of the central canal into 

 three canals, as met with at 52-56 hours, is illustrated in text 

 fig. IV. The inequality in transverse diameter and also in the 

 longitudinal extension of the three canals is clearly seen. In 

 most of the specimens examined three canals were distinctly 

 visible : in several there were four, and in one instance an 

 indication of a fifth canal was present. 



The maxinmm extension of this subdivision is on an average 

 0.126 mm. in specimens of 52-5'6 hours' incubation. As might be 

 expected, the lining of the several subsidiary canals is ependyma 

 of precisely the same structure as the lining of the central canal 

 proper. 



The cuiious shape of the internal septa subdividing the 

 central canal is clearly shown in sagittal section in Plate XLII., 

 Figures 4 and 5, taken approximately at 56 hours' incubation, 

 at which stage three canals are present. It will be seen, further, 

 that the presence of these canals and the development of their 

 divisional walls alter considerably the external configuration 

 of the cord at the posterior end, causing it to become veiy much 

 swollen, and <)ften being constricted off from the general central 

 canal. 



I have not been able to determine exactly what happens to the 

 spaces or the manner in which they close up, but this must 



