38o 
CLIFFORD H. FARR 
in general the drawings of Guignard are in harmony with my own 
observations recorded below. There are nevertheless certain features 
in Guignard 's work that require further consideration. One concerns 
the occurrence of a cell plate in the meiotic divisions of Magnolia, 
another the hour-glass spindle which is shown in his figure 21. In 
appearance the latter resembles somewhat an oblique section through 
a cell after the homoeotypic nuclear division is finished, showing two 
non-sister nuclei and the spindles between them which traverse the 
heterotypic equator. Since the paper by Guignard, two more pub- 
lications upon the cytology of Magnolia have appeared. But Andrews 
(i) and Maneval (13) respectively contribute no additional points on 
these problems. 
A recent paper (6) by the writer purported to establish quadripar- 
tition of the pollen mother cells of certain Dicotyledons by a process of 
furrowing rather than by the typical method of cell plate formation. 
Since the time of Strasburger's monumental work (15) in 1875, the 
division of all cells of the higher plants had been supposed to be by 
cell plates. A search through the literature, however, revealed a 
number of drawings which unwittingly on the part of their authors 
point to another interpretation. Furthermore, a study of six genera 
representing the Compositae, Primulaceae, Solanaceae, and Tropaeo- 
laceae respectively, the last being the one in which Strasburger himself 
studied the pollen mother cells in this regard, convinced the writer 
that cell plates are not formed during the division of these cells. 
Nicotiana was the form most carefully investigated, and it was estab- 
lished that no cell division normally occurs between the first and second 
nuclear divisions, and that cytokinesis is accomplished by furrowing 
after the homoeotypic karyokinesis. At this time the four nuclei are 
tetrahedrally arranged in the cell and a spindle connects each pair of 
them. A furrow is formed along the equator of each of these six spindles. 
There are thus four points on the plasma membrane at each of which 
three furrows meet. At these points the depression of the plasma 
membrane toward the center of the tetranucleate cell is greatest. 
These four projections finally become united at the center of the cell, 
which thus becomes transformed into a four-lobed structure. The 
isthmuses connecting these lobes gradually become narrower until the 
division is complete, each lobe becoming one of the microspores. 
During this process the mother-cell wall swells and at all times fills 
the furrow, so that a layer of it lies between the microspores as soon 
