46 ÏHOS. H. MONTGOMERY jr., 



bulk ; a view which seems improbable, since in the dense spirem stage 

 of the spermatocytes the microsomes appear to have about the same 

 volume as those of the preceding anaphase. Or we must assume that 

 the addition of nourishment to them superinduces the formation of 

 new microsomes. Such formation of new elements might possibly be 

 due to the direct transformation of nutritive substance into micro- 

 somes. Or it might be caused by addition of food particles to each 

 preexisting microsome, this addition of food (implying increase of 

 volume) indirectly causing division of the microsomes. Division of 

 the microsomes has been observed by Brauer ('93 b) in the spermato- 

 genesis of Ascaris. However this may be, may the process of division 

 of microsomes be referable to nutritive processes, either directly or 

 indirectly? The segregation of microsomes at different points of a 

 chromosome, however, could hardly be explained on this ground, since 

 the segregation takes place in the prophases of mitosis, when active 

 nutritive processes have probably ceased to greater or less extent. 



In the preceding rest stage we found the centrosomes, in the 

 few cases where they could be observed, lying in pairs in the cytoplasm 

 of the cells; and from this we concluded that their division had 

 been accomplished in the early anaphase. They are difficult to discover 

 in the early prophases (dense spirem) of the 1st spermatocytic mitosis, 

 owing to the absence of astral radiations at this stage. A long search 

 has been rewarded by the discovery of them in only a few cells at 

 this stage, from which I conclude that the process of their wandering 

 to opposite poles of the nucleus must take place very rapidly. In 

 two of these cases, representing dense spirem stages of the large 

 generation (Figs. 101, 102), the centrosomes are seen at some distance 

 from the nucleus, the axes connecting them approximately paratangential 

 to the latter. In each of these cases, which represent successive stages, 

 the two centrosomes are connected by a doubly- contoured line, which 

 is the optical representation not of a primary centrodesmosis, since 

 this line is very pale and does not stain like the substance of the 

 centrosomes, but probably of a delicate central spindle; it represents 

 probably only the outline of a spindle, possibly a central spindle sheath, 

 such as has been described by Lauterborn ('96) for Surirella and 

 other Diatoms. A central spindle at about the same stage is shown 

 also in Fig. 103, a cell of the small generation ; and also in Figs. 1 19, 

 121. But I am doubtful whether the two other figures (137, 138) 

 represent true central spindles : here each pair of centrosomes appears 

 to be joined by a single line, which would be difficult to explain, 



