852 Fraser and Snell .— The Vegetative 
(the end of the central spindle) lying within each group of chromosomes at 
the pole. When the nuclear membrane is formed water would pass into 
this mass from the outer cytoplasm, which is ex hypothesi less osmotically 
active, and the nuclear area would enlarge. If with the nuclear membrane 
the chromosome complex is carried outwards and therefore distended, the 
necessary conditions are obtained for the fission of the chromosomes in 
their more fluid interior under the pull of the lateral attachments. 
Abnormal Conditions. 
Some study has been made in recent years of the effect of abnormal 
conditions on nuclei, and Vicia has often been employed for this purpose. 
Nemec(ll) in an investigation of the effect of chloral hydrate and other 
reagents lately described a stage similar to that shown in Fig. 19 as repre¬ 
senting the separation of the gemini in a heterotype division consequent on 
the nuclear fusions he records. This stage is a very common one in our 
material, both in the roots and in other vegetative organs grown under 
natural conditions, and we are constrained to regard it as a normal phase of 
karyokinesis. That this is the case is perhaps worth recording in view 
of Kemp’s failure to find meiotic stages even in material subjected to 
abnormal conditions (8). 
We noticed, also, considerable variation in the size of the nuclei even in 
neighbouring cells (cf. PI. LXX, Figs. 9 and 10), but nuclei of irregular form, 
tripolar spindles, and other evidences of abnormality were entirely absent, 
as was indeed to be expected. 
Segmented Chromosomes. 
In a considerable number of cases some of the chromosomes on 
the spindle were seen to be made up of two or occasionally more distinct 
segments. It is perhaps possible to imagine that the segment rather than the 
chromosome or chromomere represents a discrete (hereditary) unit, and it 
might be suggested that the arrangement of these units in the chromosomes 
is indifferent and may vary. Such a speculation may throw light upon the 
fact that the number of independent Mendelian characters is in certain 
organisms greater than the haploid number of chromosomes, and moreover, 
if it were of at all general application, it would account for the often described 
variation in the chromosome number. Thus, if the average number of 
segments in certain chromosomes be two, their occasional independence or 
the union of three or four together would increase or diminish the apparent 
number of chromosomes. Again, the more or less permanent association of 
two segments would produce an appropriate physical basis for the Mendelian 
phenomenon of coupling. 
