Balls .— The Mechanism of Nticlear Division. 661 
The nucleolus is thus a mere swelling in the achromatic structures, 
whether in dividing or resting nuclei. 
Discussion of Observations on Cotton. 
The two criticisms to which all cytological work has to submit relate 
to the pre-existence of observed structures in the living cell, and to the 
accuracy of the observations made upon those structures. 
In respect to the first point we have the following facts in support 
of the reality of the thread-rings. The rings were seen 1 in metaphase 
from material which had been very badly fixed with acetic-absolute. The 
majority of the observations were made on material fixed with chrom- 
acetic, or with strong Flemming ; the fixative was injected by an air-pump, 
transference through alcohols, xylols, and paraffin by twelve-hour stages 
took ten days, and after sections of five to eight microns thick had been cut, 
they were stained with Heidenhain’s haematoxylin. In such material the 
fixation was excellent, even the cytoplasm being but little distorted. 2 The 
appearances were consistent, and similar at each stage, differing only in such 
details as the number and position of the loops of the thread-rings. The 
loops are far more conspicuous—individually—than the spindle fibres, and 
if the reality of the latter is granted, that of the loops and black dots must 
also be conceded. 
The ‘ artifact criticism ’ is of the greatest value in exercising a sceptical 
control over generalization upon insufficient evidence. Nevertheless, when, 
as in this case, an elaborate structure is developed stage by stage with the 
development of the cell which contains it, when the same structures can be 
recognized at every stage, and when nothing inconsistent with expectation 
makes its appearance, then the burden of proof rests on the artifact critic. 
The fact that these structures have not before been found in other 
plants is only partially negative evidence against them ; the dots at the 
ends of the spindle fibres have been seen by other observers. It must 
be borne in mind that the twenty chromosomes of cotton are extremely 
minute, that the nucleus is merely small, and that there is a large proportion 
of achromatic substance in comparison with the amount of chromatic 
substance. 
The writer is less optimistic with regard to the accuracy of the observa¬ 
tions. Still, he believes that subsequent corrections will only affect details, 
most of which have been indicated above. The absence of proper cyto¬ 
logical refinements, microphotographic apparatus, and especially of a 
mercury vapour lamp, have all been against great accuracy of vision. 
The Zeiss IV A stand was used, with Abbe condenser. The lens first 
employed was only the y^-inch achromatic oil immersion ; the main facts 
were ascertained with this. Subsequently the Khedivial Agricultural 
1 See also note on Cannon’s observations at the end of this paper. 2 See Figs, i and 2. 
