Division of the Chromosomes in Lilium. Bij J. B. Farmer. 503 
of a number of those in which the longitudinal fission had extended 
throughout their entire length. Each chromosome now bends over 
on itself, so as to place its originally distant ends into close apposition. 
This and the following stages can easily be understood by bending 
an elliptical hoop of wire in the corresponding fashion. The 
bending takes place in such a way that the approximated ends 
are directed radially outwards to the periphery of the spindle, and 
thus each chromosome appears split almost, but not quite , to its 
central end, with the peripheral limbs diverging. The divergence of 
the limbs may be quite obliterated, if the bending on itself of the 
chromosome is very complete, as the slit then becomes indistin- 
guishable. I may remark that Guignard * seems to have seen this 
figure, without, however, having appreciated its real significance. 
At any rate it is a very common one, and no one can have looked 
at one of these nuclei at this stage without having noticed it. The 
general effect is that of a Y, with the limbs directed outwards and in 
the plane of the equator , at right angles to the direction of the spindle. 
It is obvious that these divergent limbs can hardly represent 
the longitudinal halves of a chromosome, since, in order to 
assume their proper position on the spindle, the chromosome as a 
whole would have, first to rotate on its axis through 90°, and next 
to twist round in the radial plane of the spindle through 180°, 
in order to bring the vertical bar of the Y outwards, and the 
limbs to lie on the spindle fibres. This difficulty seems to have 
escaped Guignard, who, in fig. 14, f indicates a split at right 
angles to that shown in fig. 13, though he seems to regard the two 
as identical. The vertical part of the Y consists then of two ap- 
proximated loops, which represent the middle portions respectively of 
each of the longer sides of the ellipse, and they now commence to 
divaricate along the spindle fibres, forming the well-known divergent 
bars of the dividing chromosome. When the process is nearly com- 
plete a “split” appears in each, which is identical with the one 
previously described by me.t 
The split does not extend to the extreme end of the diverging 
daughter-chromosome, but it is of course continuous through the 
humped projection at the middle, and represents simply the reopening 
of the original chromosome ring. Finally a split appears across the 
hump, and the separation of the two daughter-chromosomes is then 
complete ; each opens out to form a Y, and retreats to the poles, with 
its free ends still directed to the equator. Very often the line of 
separation across the hump appears before the ring has opened out 
again, or the two events may happen simultaneously. The appearance 
presented by the chromosome just before the final separation of its 
* “ Nouvelles Etudes sur la Fecondation,” Ann. Sci. Nat. Bot., 7 e se'r. xiv. (1891) 
pi. 10, fig. 13. Compare also pi 13, fig. 52. t Op. cit. 
+ “ Uber Kerntheilung in Lilium- Aniheren besonders in Bezug anf die Centro- 
somen-Frage,” Flora, 1895, Heft 1. 
2 l 2 
