364 LEWIS HEN1{Y (SOUdH. 



development the radial fibres are seen to enter the cell and to 

 1)6 continuous with the fibrillar plasma. I consider these 

 modified epithelial cells of the dorsal canals to be nephro- 

 cytes ; they do not appear to have attracted the attention of 

 other observers as yet. As no flame-cells arise from the dorsal 

 canals they must represent the terminal nephridial element of 

 the dorsal canals. 



In the ventral canals a similar change takes place. The 

 epithelial cells also sink into the surrounding parenchyma 

 (fig. 30), remaining connected with the membraiie by radial 

 fibrillae. They do not, however, become directly modified into 

 gland-cells, but indirectly give rise to the fiame-cellsand their 

 capillaries. Figs. 32, 33, and o6 show the developnient of the 

 flame-cells. The first stage is represented by fig. 32 ; here 

 we see a single epithelial cell sinking into tlie parenchyma, 

 away from the membrane of the ventral canal, but remaining 

 attached to the canal. The next change apparently takes 

 place very rapidly, but without doubt it consists in the multi- 

 plication of the cell, a row of five nuclei, as shown on fig. 33, 

 being produced. These five nuclei belong to four developing 

 flame-cells and their caj^llary cell. A group of young fiame- 

 cells is to be seen t)n the same figure. The development of 

 the flame-cells in Avitellina thus goes very much on the 

 lines described hy Biigge (1002), only here we have four 

 flame-cells in a group instead of only three, the capillary cell 

 in each case originating- in the same way. As Bugge, I have 

 only found developing flame-cells quite close to the ventral 

 canal, never at any distance from it. Examination of Tsenia 

 serrata gives the same results, only iu this species the flame- 

 cells are in groups of three, plus one capillary cell (see 

 fig. 36). 



I cannot agree with Young (1908) that the flame-cells 

 develop from parenchynui cells, and that " the capillary cell 

 is at least a parenchyma cell of separate origin from the 

 flame-cells, and that the capillary is formed at first as a 

 passage in parenchyma strands, to become modified later into 

 a definite tube with a specially modified wall." On the other 



