No. I.] SPERMATOGENESIS OF BATRACHOSEPS. 65 



of fiber cones takes place only in the auxocytes. A new set 

 of fibers appear radiating from the poles of the spindles after 

 they have passed down through the ring-like nuclei of the 

 auxocyte. The object of these fibers is to pull the daughter- 

 cells apart. The latter part of the radiosomic process consists 

 in the reconstitution of the spheres and the reassumption of 

 its original position by the nucleus. With this the radiosomic 

 process can be considered finished in the auxocyte. In the 

 spermatocyte it commences in a different manner. Instead of 

 a central spindle being formed by the separation of the two 

 halves of the archosome, it is, at least in the majority of 

 instances, formed by the junction of two opposite spindle 

 cones. 



The fibers of these cones dissolve the nuclear membrane and 

 by approaching each other form the central spindle. This part 

 of the radiosomic process I have, from want of sufficient mate- 

 rial, been unable to study as carefully as that which takes 

 place in the auxocytes. I do not deny that a central spindle 

 may be formed in the same manner as in the auxocyte, but I 

 have failed to find any evidence of such a formation. There 

 is, however, undoubted evidence that a central spindle is 

 actually formed by two opposite fiber cones which, receding 

 from the cell wall, meet in such a way as to form the spindle. 

 The other parts of the process are the same as that which takes 

 place in the auxocytes. 



To the radiosomic process must also be referred the forma- 

 tion of the new cell walls which separate the two daughter-cells 

 and the pulling apart of the two new cells. 



The new cell wall between the two daughter-cells is formed 

 in the following way. Already during the metaphase (Fig. 53) 

 the plasmosphere has scattered, and many of its granules and 

 secretions have become located along the equator of the cell. 

 As the anaphase progresses, this accumulation of plasmospheric 

 granula becomes more prominent along the line of the future 

 cell wall. The plasmospheric granula are never distributed 

 evenly along the equator, but always in isolated rounded 

 groups (Figs. 54-56). In the figures referred to, the plasmo- 

 spheric fragments are stained deeper red than any other part 



