32 THE STRUCTURE OF CELLS 



though it is not certain as to whether the details of their develop- 

 ment are similar to that of the spermatogenetic tetrads as described 

 by Brauer. 



Meanwhile, other changes have been proceeding, both within 

 and without the nucleus. The nucleolus commonly can be seen to 

 lose substance during the growth of the chromosomes, as is testified 

 by its vacuolated appearance. Often it fragments into smaller 

 particles during the prophases. But it has been definitely ascer- 

 tained, in many cases, that some part of this body is cast out into 

 the cytoplasm where it degenerates, and it is not improbable that 

 this will prove to be of very general occurrence. The possible 

 significance of this event ought not to be overlooked in view of its 

 striking occurrence in the lower forms of life, for it is in them that 

 the clue to the meaning of the complex changes observed in the 

 higher animals and plants must probably be sought. 



In the cytoplasm, also, remarkable changes connected with the 

 centrosome and spindle mechanism have been proceeding. The 

 latter reaches very different degrees of completeness in different 

 organisms, and, as has been said on a previous page, even in the 

 different cells of the same organism. In the simplest case the two 

 centrosomes, when present, diverge and form a spindle not dis- 

 similar from that already described for somatic nuclear division. 

 In other instances (e.g. in Salamander) the two centrosomes diverge 

 tangentially to the nucleus, and the spindle is formed between 

 them, and, in the first place, without immediate reference to the 

 nucleus. Later on, however, from the poles of the central spindle 

 thus differentiated not only do the radiating fibrils reach into the 

 protoplasm, and even to the periphery of the cell, but there are 

 also others that extend to the nucleus and become attached to the 

 chromosomes. The latter are thus, as it were, roped up and pulled 

 on to the periphery of the first formed spindle (Fig. 15, D, E). 

 Almost every gradation occurs between the extreme forms here 

 sketched, and the matter seems to be essentially one of more or 

 less complete division of labour between the constituent parts of 

 the achromatic spindle regarded as a whole. In the Salamander, 

 and those other cases in which it appears in a more or less complete 

 form, the central spindle seems to function as a sort of support to 

 keep the two poles apart, and to serve as a sort of railroad along 

 which the daughter chromosomes can be pulled along by the con- 

 tracting peripheral or mantle fibrils that attach them directly to the 

 poles. 



In the divisions of the higher plants, as has already been ob- 

 served, no centrosomes have been identified with certainty, and the 

 spindle at first starts into existence quite irregularly. It speedily, 

 however, becomes for the most part bipolar, although not unfre- 

 quently isolated fragments of extruded nucleolar substance exert a 



