1909] SCHAFFNER— REDUCTION DIVISION IN AGAVE 203 



properly. It begins to twist into loops and the chromatin granules 

 now appear double {figs. 14, 14a). Although the spirem is much 

 thicker at this stage than it was earlier, synizesis is occasionally present, 

 the spirem filling one-half of the nuclear cavity, as shown in fig. 75. 

 The double granules are at length prominent, although the spirem 

 does not split (figs. 16, 16a). Finally the whole spirem is thrown into 

 definite loops of various shapes and sizes. It is difficult to represent 

 the perspective of these loops in a drawing. One can trace out the 

 position and depth by focusing up and down, but in the camera 

 projection they appear nearly in a plane (figs. 17-20). There is no 

 question but that the spirem is continuous, since one can often follow 

 the thread through more than half of the loops without losing the 

 connection, and in uncut sections no free ends are present. In tan- 

 gential sections or half -sections one can frequently also follow through 

 three or four loops before coming to a free end (fig. 10) . Practically 

 also, it seems impossible that such twists ajid loops could be formed 

 unless the spirem were continuous. In fig. 18a a number of twisted 

 loops are shown. Some of the loops are produced by a single twist, 

 which results in ring-shaped chromosomes (fig. 20) . There are three 

 of these ring-chromosomes in the nucleus and they are developed 

 e b y Slae - The three main types of loops are shown in figs. 20a, 

 20b, and 20c. The loops are not formed, as in Lilium, with a central 

 toot, but more openly. In this stage synizesis was also present in 

 some of the material (fig. 21). 



After the loops are developed, they are pressed and curved against 

 1 e nuclear wall, the whole central part of the cavity becoming very 



>a me. At the same time they break apart to form the individual 

 * uction or bivalent chromosomes (figs. 22, 23). It was exceedingly 

 [ cu " to determine the number of chromosomes on account of the 

 "regular shapes of those bodies in some nuclei, as appears in figs. 

 2 4> 2 5; but it was finally determined that the number is twelve 



gS ' 26 ~3 2 > 37, 38) . In fig. 26 only fragments of the twelve chromo- 

 es are shown, a large part of the nucleus being cut away. 



wh n ^^ WherC the number of chromosomes is said to vary some- 



Va a ' . 8 re ater or less number may not be of any significance, if the 



Ration appears in vegetative division. Two or more chromosomes 



g * ecome united through a failure of transverse segmentation, 



