i 9 7 
certain Varietal and Hybrid Ferns. 
assumed by iron filings in a magnetic field would differentiate such filings, 
if present, in a mixture of iron and other dust, provided that the mixture 
were brought into a suitable magnetic field. 
We do not think a question such as this can be really answered by 
an appeal to stains. The latter act differently according to the physical, 
as well as the chemical, state of a substance, and it may be that the 
well-marked staining reactions of ‘ kinoplasm ’ ought to be regarded as 
an expression of physical differentiation (which may be transient) rather 
than as a proof of chemical difference. On the other hand, it is possible 
that a chemical change may also be involved in the operation of those very 
physical conditions that are at the same time responsible for the differentia¬ 
tion itself. The aggregation of the nuclein-charged linin beneath the points 
of origin of the spindle cones indicates the probability that there exists a 
causal connexion between the two phenomena, and a suggestion that the 
immediate cause is electrical in nature appears to be supported by the ab¬ 
sence of any direct connexion between the chromatic aggregations and the 
fibres of the spindle cones, whether at this or at an earlier stage, such as 
any hypothesis of ‘ growth * would imply. 1 
As the course of the mitosis advances, the state of aggregation of 
the chromatic linin disappears, and finally the chromosomes, when fully 
formed, become uniformly distributed just within the nuclear wall at 
diakinesis. But the quadripolar character of the spindle persists through 
these changes. We have already given reasons for not regarding this fact 
as anomalous; perhaps it may be taken as an expression of a state of ‘ lag ’, 
such as has been regarded as probable by Hartog. 2 But it cannot be 
a merely fortuitous circumstance that as the brief stage of diakinesis ends 
with the dissolution of the nuclear wall (itself a coagulum or a precipitation, 
and probably to be regarded as due to the interaction or agglomeration of 
substances present in the colloidal masses of which the cytoplasm and the 
nucleus respectively are composed, and only lasting so long as certain 
definite physical conditions persist), the character of the spindle should con¬ 
comitantly change as well. The cytoplasm, gaining access to the interior of 
the nucleus, instantly brings about the differentiation of the portion .of the 
fibres which are attached to the chromosomes. At the same time the four 
cones rapidly coalesce into two opposite bundles, 3 thus forming at this late 
stage the bipolar spindle (PL XVI, Figs. 8, 9), which is, in the majority of 
plants, commonly reached at a much earlier stage. And just as the chromatin 
at the earlier stage appeared to determine the formation of the four or more 
poles, so now the chromosomes again appear to be the active agents 
in effecting the resolution of the quadripolar into the bipolar arrangement. 
1 We are of course aware that attempted explanations based on the assumption of diffusion 
currents have been advanced, but these appear to us to rest on no real foundation. 
2 Hartog, M., loc. cit. 3 Cf. Lawson, A. A., loc. cit. 
