72 bulletin: museum of comparative zoology. 



proceeded a considerable distance in their migration, while several more 

 are already definitely formed. In Figures 95 and 96 all the archoplasm 

 except a very small portion, which remains permanently in this region, 

 has migrated along the axial filament. 



Recalling the phenomena connected with the origin of the axial fila- 

 ment and the early stages in its growth, I think but one conclusion can 

 be drawn regarding the significance of the migration of these archoplas- 

 mic bodies. They contribute to the further growth of the axial fibre. 

 At the stage represented in Figure 9G these small deeply stained granu- 

 lar masses may be seen at various points along the course of the filament 

 and always close to it. In later stages they are also found, but they are 

 then much less definite in outline and less deeply stained. Apparently 

 their substance is gradually broken down during this period of the 

 elongation of the cell and the continued growth of the axial filament in 

 length and thickness. 



However, there are other phenomena which might be construed as 

 indicating a different origin of the filament. In Figures 95 and 9G are 

 shown at various places in the cytoplasm fine short granular or monilated 

 fibres which in many ways resemble the beginnings of the axial filament. 

 Like it they are stained black by the haematoxylin-Congo-red method. 

 As a genera] thing they lie in the reticulum of the cytoplasm and often 

 appear merely as thickenings of it. From this it might be argued that 

 the axial filament itself is derived directly from the cytoplasm by the 

 mere condensation and fusion of parts of the reticulum. However, 

 there are frequently structures of this kind which have no connection 

 with the reticulum whatever, but lie free in vacuoles of hyaloplasm. 

 Figure 101 shows a striking example of this condition, but similar con- 

 ditions are also to be observed in Figures 95 and 96. 



From these phenomena and those already desci-ibed in connection 

 with the earlier stages in the formation of the filament, it is very im- 

 probable, I think, that the axial filament arises directly from the cyto- 

 plasmic reticulum, although it is possible that the network may undergo 

 a chemical change similar to that experienced by the archoplasm' and 

 thus may afford some of the material for the formation of the axial 

 filament. 



The typical arrangement of the centrosomes shortly after the beginning 

 of the axial filament — one on each side of this structure and one form- 

 ing an enlargement at its proximal end — continues to exist until the 

 elongation of the cell body is well under way (Figs. 93, 94). After this 

 stage the two lateral centrosomes are invisible for a considerable time ; 



