202 ' BOTANICAL GAZETTE [march 



wall or flattened out in the "sickle stage/' is in itself sufficient proof 

 that synizesis is an artifact. But as stated, the chromatin can also 

 be faintly recognized in the living nuclei, and it should be still more 

 evident if in a contracted ball, since the cavity in typical synizesis is 

 entirely empty of threads, flakes, or granules. The granular material 

 in the nucleus often radiates outward from the nucleolus, and some- 

 times it is prominently distributed over the surface of the nuclear 

 membrane. 



With salt solution and also with 95 per cent, alcohol, the cells 

 contracted considerably and soon became indistinct, so that it was 

 difficult to make out any details. The nuclei were displaced to some 

 extent. The weaker chrom-acetic acid solution, used for the paraffin 

 material, caused the whole mass of sporocytes to contract violently, 

 but not much displacement of the nuclear contents was noticeable. 

 This was probably because the cells were lying rather free and could 

 contract readily from all sides, or the fluid may not have acted long 

 enough. However, it is probable that the synizesis occurs rather 

 suddenly. 



An attempt was made to stain the fresh material, both before and 

 after treatment with killing fluids; but this proved unsatisfactory 

 the stained material showing no more detail then the living cells. 



A study of the living microsporocytes of Agave virginica indicate 

 that synizesis, as seen in the usual paraffin sections, is an artifact. 

 When the chromatin is comparatively free in the nuclear cavity and 

 is expanding, we find the most decided synizesis. Meanwhile, as will 

 appear further on, synizesis is not confined to this stage, but occurs 

 to a greater or less extent until the chromosomes are fully developed. 

 It is 'largely on account of the erroneous idea that synizesis occurs at 

 but one stage of division that a number of inaccurate interpretations 

 have been advanced, through which the whole subject of reduction 

 has been confused. 



Development of the chromatin loops.— The spirem begins to thicken 

 while the chromatin granules are still in a single row (figs. U> Jia >' 

 At this stage synizesis is still frequent, the spirem usually being 

 crowded to one side, but occasionally lying entirely around the nuclear 

 wall (figs. 12, i 3 ). The spirem now becomes very distinct, so that 1 

 is often possible to trace out great lengths of the thread by focusing 



• 



