ISTichols, A morphological study of Juniperus communis var. depressa. 207 
and in close contact, either parallel or twisted around each other 
corkscrew-fashion, but frequently they are in contact at only one 
or two points. 
Coincident witk these nuclear metamorphoses, changes have 
been taking place in the cells themselves. They haye lost their 
angular shape and compact arrangement and have become more or 
less separated from one another (fig. 11). This Separation is 
apparently brougkt about by the dissolntion of the middle lamella 
of the mother cell in the manner described by Strasburger (1882). 
Allen (1905) reports an interesting process in Connection with the 
Separation of the pollen mother cells in Lilium ccmadense. There 
the wall of the mother cell dissolves, leaving the cells protected 
only by - a plasma membrane, and subsequently an entirely new 
wall is developed about the cell. In Juniperus, immediately after 
the Separation of the mother cells, the wall appears very thin, bnt 
at all times places may be found where the cytoplasm has shrunk 
away from the surrounding membrane, which wonld hardly be true 
if the cells were surrounded merely by a differentiated layer of 
cytoplasm. 
Up to this time the cytoplasm has exhibited afairly uniform 
alveolar structure. As the nucleus emerges from synapsis, however, 
a fiber-like, radial arrangement of the cytoplasmic materials is 
faintly discernible (fig. 20), similar to that figured in the pollen 
mother cells of Lilium by Mottier (1897, fig. 3), and in Larix 
by Allen (1903, fig. 6). Shortly before diakinesis there is evident 
a concentration of the cytoplasm toward the nucleus, and by the 
time the Segmentation of the spirem has taken place the nucleus 
is enclosed by a dense granulär layer, while toward the periphery 
of the cell the cytoplasm has become thin and stains very lightly 
(fig. 21). This dense layer of cytoplasm around the nucleus doubt- 
less represents the Telted 7 structure described in Larix by Bela- 
jeff (1894), Strasburger (1895), and Allen (1903), but on account 
of the small size of the cells it is impossible to ascertain definitely 
its nature here. Occasional indications are seen, however, of 
a fibrous structure like that described by the above mentioned 
authors. » 
The nuclear membrane, which until now has been sharply 
outlined, disappears soon after the formation of the ‘feit’, and the 
dense cytoplasmic layer appears to press into the nuclear cavity. 
The origin of the spindle übers, however, cannot be clearly de- 
monstrated. Osterhout (1897) and Mottier (1897) find that during 
diakinesis, in the species which they studied, fibers of unmistakably 
nuclear origin become attached to the chromosomes. But a careful 
study of the chromosomes of Juniperus before the dissolution of 
the nuclear membrane has failed to reveal any such structure, 
and it is impossible to state whether the spindle fibers are derived 
entirely from the cytoplasm or whether they originate partly within 
the nucleus. 
The spindle when first formed is multipolar (fig. 24), but 
rapidly resolves itself into a multipolar diarch (fig. 25), and 
