205 
certain Varietal and Hybrid Ferns. 
A peculiar feature, already observed in P. mdg. var. elegantissimum , and 
once also in the typical P. vulgare , is to be seen in the nuclei which go through 
division without the disappearance of the nuclear wall (PI. XVIII, Figs. 33, 
34). Such examples, as already pointed out, at first sight suggest amitosis, 
but the fact that the chromosomes are quite distinguishable serves to indi¬ 
cate the need for caution in drawing this conclusion respecting them. As 
a matter of fact, they are more properly to be described as imperfect 
mitoses—imperfect inasmuch as the nuclear wall does not become destroyed. 
The whole nucleus is pulled out like a sac, and this then is distended at its 
two ends. The ends gradually receive a moiety each of the chromosomes, 
which are sometimes much smaller and shorter than in the normal mitoses, 
or, if the process has set in earlier, they may be in the long thread stage. 
The two end-sacs are for a time connected by a tubular bridge which 
gradually becomes more and more attenuated and finally is empty of 
chromosomes, these having passed to the distended ends. In the cyto¬ 
plasm which lies between the latter the interzonal spindle fibres are developed. 
These are often somewhat irregular, but they are invariably present, and 
when the tubular bridge is centrally and symmetrically placed in the cell, 
it is seen to be equally surrounded by the fibres. When, however, as is 
often the case, the tubular connexion is excentrically placed, it may happen 
that the interzonal fibres are absent or sparsely developed on the outer side. 
The whole process seems to us to afford a confirmation to the view that the 
fibrous system of lines is due to an indirect influence of the nuclear contents 
on the cytoplasm and to be dependent on the induction of a state of stress 
as the result of the chemical changes going on in connexion with the 
nuclein transformation that is taking place. At any rate there is here 
no question of any kind of open communication between the nuclear con¬ 
tents and the cytoplasm such as accompany an ordinary mitosis. A cell- 
plate is formed across the interzonal fibres, and sometimes this seems as 
if it were about to nip the tubular connexion into two parts. 
The homotype mitosis (Fig. 35) calls for no special comment beyond 
the statement that this curious form of division may also occur in connexion 
with it; indeed it seems to do so generally when it has occurred in the 
preceding (heterotype) mitosis. 
During the reconstruction of the normal daughter-nuclei the cytoplasm 
is seen to contain numerous granules which are coloured by dyes that are 
capable of staining chromatin. These are very well known to all who are 
familiar with the details of karyokinesis, and there is some doubt as to their 
origin. We have been able to trace at any rate some of them in these 
plants quite definitely to a nuclear source. They are given off (Fig. 35) 
by the vesiculating chromosomes as buds which pass out from the nucleus 
into the cytoplasm (cf. PL XVII, Fig. 12), in somewhat the same way as 
occurs at the commencement of the division (PI. XVI, Fig. i) e Whether 
