KARY0KINES1S. 67 



me to be insufficient. No small amount of evidence is required to prove that nuclei, 

 which are so similar in all other respects, are so different in this one. I agree with 

 Boveri that research only can determine where (and I should add whether) this is 

 true. 



Returning now to Boveri's idea that in the artificial production of centrosomes 

 they are formed by a kind of regeneration : — Morgan's experiments make it ex- 

 tremely probable that numerous centrosomes may be formed independently of each 

 other in the cytoplasm. If these are formed of achromatic nuclear material it 

 is easy to understand that they appear wherever a sufficient accumulation of 

 this material is found in the cell body. But achromatic material separated from 

 the nucleus is not necessarily a centrosome with all of the morphological and physi- 

 ological features which that body exhibits, as is shown by the fact that such mate- 

 rial is distributed through the cell at every mitosis. Either there must be an escape 

 of some centrosomal Substance or structure, or the condition of the cell must be 

 such as to bring about centrosome formation from ordinary achromatic material ; 

 the latter is I believe Hertwig's view (see quotation on p. 64) ; the former is held by 

 Boveri who considers that the centrosome may be regenerated (repaired) from the 

 achromatic substance of centro-nuclei only. 



This very suggestive hypothesis of Boveri's makes it possible to harmonize the 

 well established fact of the persistence of the centrosome as a cell organ with that 

 other apparently contradictory fact that centrosomes may be produced experimen- 

 tally in the cell ; and this it does by practically adding another phase to the series 

 of changes through which the centrosome may pass, viz. : the phase of the centro- 

 nucleus. 



But the question of the persistence and morphological significance of the cen- 

 trosome does not hang on the fate of the uncertain hypothesis that nuclei belong to 

 two classes, pure nuclei and centro-nuclei. The one point of importance is that 

 centrosomes are not coagulation products, nor the mere expression of cell stresses, 

 nor sporadic or spontaneous structures, which may appear and disappear here, 

 there or anywhere, depending upon the physiological condition of the cell ; but that 

 all kinds of centrosomes, whether normal or artificial, are formed of a specific kind 

 of protoplasm which is genetically related to the achromatic substance of the 

 nucleus, from which, under certain conditions, they may be formed anew, that thev 

 have a characteristic structure and metamorphosis, that they possess the power of 

 independent growth and division, and that they are therefore cell organs, which are 

 at least relatively, even if not absolutely, persistent structures of high morphologi- 

 cal significance. 



(/) Homology of the Centrosome. — We may now again consider the parallel- 

 ism between the nucleus and the centrosome pointed out in a previous section (p. 

 55). This parallelism is seen not only in the alternate growth and diminution of 

 both, but also in corresponding changes in staining reactions and in similarities in 

 their modes of self propagation, both nucleus and centrosome being continued from 

 generation to generation by means of small granules (chromosomes, central corpus- 

 cles) which have the power of independent growth and division. 



