Nuclear Division in Funkia. 
381 
earlier prophases and synapsis, while, for the later stages, sections 
from 5 — 12 u were chiefly employed. Preparations thick enough 
to contain uncut nuclei of all stages were also examined. 
II. Reduction Division. 
a) Seriation of the stages. The large size of the nuclei in 
Funkia makes it an especially favourable plant in which to study 
nuclear division. Before deciding on the seriation of the ditferent 
stages obtained, it was found best to draw numerous figures and to 
arrange them provisionally; this arrangement was then confirmed or 
corrected by a re-examination of the preparations. It was observed 
that the pollen mother cells of any single anther loculus were ge- 
nerally in slightly diiferent stages of development, so that the nuclei 
at either or both ends differed slightly from those in the middle of 
the loculus; the assistance afforded by this fact can hardly be over- 
estimated. For example, it is found that in one loculus, most of 
the nuclei of which contain double spiremes, a few nuclei at one, 
or more often both, ends are in synapsis; in other cases the double 
spireme stage is present only in a few nuclei situated at the ends 
of the loculus, while in the centre of the same loculus the nuclei 
are in the condition of single spiremes. Since nuclei which are in 
synapsis and those which contain the single spireme are never found 
in the same loculus there seems to be no doubt that the double 
spireme is intermediate between the other two stages. The same 
method of determining the correct seriation has been followed in 
every case. 
b) The Heterotype division. I. Early prophases. The 
nuclear reticulum in the resting stage is composed of a number of 
knots connected by filaments. In sections treated with Flemming’s 
triple stain both these constituents stain in a similar manner, but in 
preparations faintly stained with Heidenhain, it can be seen that 
numerous dark granules are present, embedded in a colourless sub- 
stratum, the combination of the two forming the knots, while a few 
of these granules are also scattered in the filaments or threads con- 
necting the knots 1 ). It would seem to be almost impossible accu- 
rately to count the knots, but there can be no doubt that here, un- 
like those cases described by Overton 2 ) and others, their number 
*) Cf. Allens description in L. canadense (II) P. J. 1906. 
2 ) Overton. P. J. 1906. 
