POLLEN MOTHER-CELLS OF LILIACE^E. 439 



the desired developmental state. We open a very young flower- 

 bud, take out an anther with the forceps, place it in a drop of acetic 

 methyl-green, or acetic gentian-violet, lay a cover-glass upon it, 

 and press upon it with some flat object till the anther-cells are 

 flattened and their contents evacuated. The evacuated contents 

 are immediately fixed by the acetic acid and stained by the methyl- 

 green or gentian-violet, and we can see at once whether we have 

 nuclei resting or in a state of division. If the pollen mother-cells 

 are already divided into four daughter-cells, or the young pollen- 

 grains are already separated from one another, we must go back 

 to younger flower- buds. We can easily recognise, by the thick, 

 colourless walls which the pollen mother-cells have, whether we 

 are looking at them or at young pollen-grains. We go back to 

 successively younger buds until we can see in the nuclei of the, as 

 yet, thin- walled and connected mother-cells a fine threadlike coil, 

 and a flat nucleolus lying against the wall of the nucleus. At this 

 stage of the development the coiled thread contracts under the 

 influence of the reagent, withdraws (Fig. 154, a) from the nuclear 

 wall (this latter remaining uncoloured), and we can determine that 

 this wall is a membrane formed of surrounding cytoplasm. 



The nuclear thread then becomes continually shorter and 

 thicker, and its coils untwist. A differentiation of the thread into 

 successive disks of varying thickness becomes noticeable, where- 

 upon the thread undergoes longitudinal cleavage, and at the same 

 time, by cross-division, falls into a definite number of segments. 

 These segments represent longitudinally cleft chromosomes, of 

 which the number in a pollen mother-cell appears to be only 

 half as great as in the vegetative cells of the same species 

 (reduction of the chromosomes). Each pair of sister chromo- 

 somes lies together lengthwise, and a more or less complete fusion 

 takes place between them. These juxtaposed pairs of chromo- 

 somes are chiefly distributed upon the nuclear membrane (6), 

 the nucleolus is dissolved, and simultaneously there follows the 

 formation of the nuclear spindle out of delicate threads (spindle- 

 fibres), which partly join on to the chromosomes, partly proceed 

 from one pole of the spindle-figure to the other. The chromo- 

 some pairs become arranged into a nuclear plate (or nuclear disk) 

 in the equator of the spindle (c). If we happen to have a polar 

 view of the nuclear plate (d), the number of twin chromosomes 

 forming it can be readily counted in this case, apart from occa- 

 sional exceptions, twelve. By the contraction of those spindle- 



