132 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



steadily until the chromosomes remain only as small granules in a 

 line reticulated structure. In the case of the last premeiotic 

 division as described it may be clearly seen that the chromosomes 

 form the prochromosomes of the pollen mother-cells as Rosenberg 

 ('09), Digby ('10), and others have claimed. 



Reduction divisions. — The pollen mother-cells were studied in 

 D. rotundifolia, D. intermedia, and D. filijormis. In my material 

 of D. longifolia the pollen grains were already mature so that the 

 divisions forming the embryo sac alone were available for chromo- 

 some counts. 



D . filiformis is most favorable for the study of the pollen mother- 

 cells on account of the somewhat larger cells. The germ cell 

 nuclei are all provided with a single large nucleole and in it one 

 invariably finds one or more large vacuoles as shown by Martins 

 Mano ('05), Nichols ('01), and Digby ('08). These vacuoles 

 afford no evidence that chromatin is budding off from the nucleole 

 as suggested by Nichols, although in all species studied small 

 globular bodies may occasionally be seen lying in contact with 

 the nucleole as figured by Rosenberg ('09). The nucleole gen- 

 erally occupies a peripheral position in the nucleus and the prochro- 

 mosomes (Fig. 15) are scattered in a fine linin network. These 

 bodies are slightly elongated or rod-like granules which generally 

 lie in pairs in the periphery of the nucleus. The prochromosomes 

 are readily distinguishable from the other granular substances 

 which appear in the nucleus at this time by their deep gentian- 

 violet stain and their compact homogeneous consistency. By 

 close study there may be found all stages in the transformation 

 of the prochromosomes to form the leptoneme spireme. The 

 process is similar to that shown by Overton ('05) for Calycanthus 

 floridus. Two fine thread-like bands are formed, the leptoneme 

 spireme (Figs. 26, 27, 28) as figured by Berghs ('05) for D. rotundi- 

 folia and later by Rosenberg ('09). The mature leptoneme spireme 

 can be clearly seen to be made up of two thin, parallel bands of a 

 homogeneous structure. In some stages these threads seem to 

 have fused and a single continuous spireme band appears (Fig. 16). 

 Occasionally this fused band may have a beaded appearance, as 

 shown by Rosenberg, Overton, and Allen. In this condition 

 there is no evidence of its bivalent nature. It is quite apparent 

 from the above that the pairing of the prochromosomes is to be 

 side by side. 



