20 ARTHUR BOLLES LEE 
‘Pfitzner’s granules’, ‘chromomeres’, and the like, are, as 
already explained, nothing but ill-seen and faultily interpreted 
images of bulges and twists of the axis of the chromosomes 
(figs. 8 to 23 and many others of this paper should make this 
sufficiently clear). It therefore only remains to be seen whether 
any other mode of division can be made out. 
To settle this point, the first step must be to make out at 
what stage chromosomes can first be seen to be double. Accord- 
ing to Flemming (“ Neue Beitraige zur Kenntniss der Zelle”’, 
i, in ‘ Arch. mikr. Anat.’, xxxvu, 1891, pp. 787, 744, and 745) the 
supposed splitting takes place in the spireme stage. And 
this is apparently the view still taken by the great majority 
of cytologists. 
IT am not aware that any observer has asserted a division of 
chromosomes during the interphase. A longitudinal splitting 
at the telophase has been asserted by several writers, 
and with especial insistence by Dehorne. ‘This writer even 
maimtaims (in his “* Recherches sur la division de la cellule ’’, in 
‘ Arch. f. Zellforschung’, vi, 1911, p. 613) that it may take place 
as far back asthe anaphase. This is indubitably erroneous. 
For beyond all doubt at this stage the chromosomes show no 
hint of duplicity. But as regards the telophase I find 
that—in some cases at least—at that stage the chromosomes 
are certamly double—in a sense; and I acknowledge the 
essential correctness of Dehorne’s clever figs. 7, 9, 10, 11, 12, 
and 18 (his fig. 6, which corresponds to my fig. 48, I think has 
been imperfectly understood by him). But I find no trace of 
any evidence that this duplicity is brought about by a 
longitudinal splitting. 
A division of the chromosomes at the telophase has also been 
maintained by K.C. Sehneider. In his‘ Lehrbuch der ver- 
eleichenden Histologie’, 1902, pp. 10, 118, 848, and 989, he 
states it as a probable inference. He suggests that at this stage 
the chromosomes segment transversely at the polar 
bends ; and that the two moieties thus formed grow past one 
another so as to become parallelly approximated throughout 
their lengths. I have duly investigated this point, and find no 
