392 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9, 
Splitting of the chromosomes into two parallel threads takes place, as 
described by Digby (191 9), or whether the events conform rather to ac- 
counts such as those given by Sharp for Vicia (1913) and Tradescantia 
(1920). It was found impossible, however, to get very definite results, 
as the nuclei at this stage were too small, and the chromosomes too tightly 
packed to allow of close observation. The following account, therefore, 
begins with the return of the nucleus to a resting condition after the last 
archesporial division. 
The cells of the archesporium pass through a fairly long-continued 
resting stage previous to entering upon the maturation prophase. They 
lie usually in one or two rows and fit tightly together with no spaces be- 
tween them or between the archesporium and the well-differentiated tape- 
tum. The nucleus is comparatively large (lo-ii /-t), very nearly spherical, 
and possesses a prominent nuclear membrane (PI. XXV, figs. 1,2). Within 
it may be found one or more nucleoli. Ordinarily but one of these attains 
to any size. It is spherical, and lies toward the center of the nucleus. 
Within it may often be seen one or more hyaline circular areas, resembling 
vacuoles. These may contain, or occasionally give place to, angular, 
crystal-like bodies. The other nucleoli when present are usually smaller 
and are stationed here and there throughout the nucleus. All of these 
bodies are in intimate contact with the linin network. 
Scattered evenly throughout the nucleus is a reticulum of very delicate 
threads. Some of the threads are very thin and free from granules. Others 
are dotted over more or less unevenly with chromatin particles. Wherever 
threads join, chromatin accumulations varying in size from tiny dots to 
fairly large masses are to be found. These may be irregular and shapeless, 
or rounded and indistinguishable, perhaps not different, from small nu- 
cleoli. In well fixed material the network closely envelops the nucleolus, 
to which it is connected in various places. There is, therefore, no space 
around the nucleolus. Wherever the fixation is faulty, however, a clear 
space is observed, due to shrinkage of the reticulum. During the resting 
stage there is little if any indication of parallelism of the threads. For 
the most part they are single and are uniformly distributed throughout 
the nucleus. 
The resting condition lasts for quite a time, long enough for the anther 
practically to double in length. Throughout this period all trace of the 
individual chromosomes is lost — they have been resolved completely into 
a network of threads. 
Although there is little or no sign of parallelism of threads during the 
resting stage, there begin to appear occasional instances of such an approxi- 
mation as the nucleus begins to draw near to the time when it will enter 
upon the heterotypic prophase (fig. 3). By no means all the threads 
become thus arranged. Most of them remain single. Moreover, the 
parallel threads do not come together very closely, and there is not much 
