Pollen Development in Lactuca . 383 
adjacent one. Cytomyxis was also observed to take place during the 
spireme stages, but not so frequently as in synizesis, probably because the 
mother-cells are more separated at later stages. One case of such transfer 
of nuclear material was seen at the telophase of the heterotypic division, 
where it is evidently to be regarded as an abnormality. It seems probable 
that the phenomenon is not peculiar to anyone stage of pollen development, 
but is merely dependent on the nucleus becoming eccentric and taking up 
a position where there happens to be an opening at the point of contact 
with an adjacent cell- wall. That it is in some way connected with a 
pathological condition of the loculus seems to be indicated by the fact that 
the phenomenon is of much more frequent occurrence in loculi which are in 
other ways abnormal than in otherwise perfectly normal ones. 
This process, by which a portion of the chromatin from one pollen 
mother-cell nucleus passes through an opening in the cell-wall into the 
cytoplasm of an adjacent mother-cell, was first observed by Koernicke 
(1902) in Crocus . The same process was found by Digby in Galtonia (1909), 
Primula (1912), and Crepis (1914). Gates (1911) described the process in 
Oenothera and called it cytomyxis. Nakao (1911) has seen it in rye. 
Gregory (1905) observed it in a race of Lathyrus odoratus which had sterile 
anthers, but mistook it for an abnormal division of the pollen mother-cells 
by constriction. Rosenberg observed it in Crepis (1909 a) and also in 
Drosera longifolia (1909 b, Fig. 11 , p. 22 ). Fraser (1914) found it in Vida. 
Faba. West and Lechmere (1915) described it as occurring on a large 
scale in Lilium candidum , although none of the appearances of pollen 
sterility were found. In this case it must either be normal or induced by 
the treatment. Sakamura (1916 and 1920) figures it in Vicia Faba , and 
found it much more frequent in anthers which had undergone chloralization 
than in those which had not. 
The most frequent opinion expressed is that the process is an abnormal 
one, either pathological or induced by the fixation or treatment, and that 
it will be followed by break-down of the cells in which it occurs. There are, 
however, difficulties with this view, and in some cases it is followed by 
return of the nucleus to a central position and absorption of the extruded 
. chromatin in the cytoplasm of the invaded cell. 
Development of the Tapetum. 
Some features of the early tapetal cells have already been mentioned 
(pp. 366, 371 ). The tapetum is a well-defined layer even at the very 
early archesporial stages, and is remarkable even then for the elonga- 
tion of the cells in the direction of the long axis of the anther, that 
is, at right angles to the usual direction of elongation. The two mitotic 
divisions of the nuclei in the tapetal cells are usually completed about the 
end of the synaptic period in the pollen mother-cells. But many of the 
C c 2 
