L. DONCASTER AND J. GRAY. 
*)02 
able from one of the thicker rods, and it may be present 
in one cell of a 2-cell stage, and apparently absent in the 
other. 
Miliaris ? x Esculentus and Miliaris ? x 
A c n t n s ^ . 
These two crosses, the converse of the last two described, 
may be taken together. As has been mentioned above, the 
miliaris eggs in 1912 were very unsatisfactory, only a small 
proportion yielding larvm even when fertilised by their own 
sperm. In our material of the hybrid eggs, in both cases an 
exceedingly small proportion showed evidence of fertilisation. 
In the cross with esculentus (J about 2 per cent, are begin- 
ning to develop; in that with acutus the percentage is 
considerably lower. Most of the developing eggs in both 
cases are in the 2-cell stage, but nearly all stages of both first 
and second segmentation divisions have been found, and at 
least one clear case of the conjugation of the egg and sperm 
nuclei. 
The figures in both crosses are distinguished from those of 
the converse crosses by the greater irregularity of the chrotno- 
somes on the spindle. 'I'he extent of this irregularity varies; 
some figures are almost normal, in others the chromosomes 
are muc-h scattered, but cases are rarely found in which an 
accurate count of chromosomes is possible. In one unusually 
regular first anaphase of miliaris ? x acutus J* cut 
transversely to the spindle, 35 may be counted at one end 
and about 36 at the other; in this case, however, four 
chromosomes are scattered entirely outside the main group 
around the pole. In late anaphases and telophases of both 
crosses it may often, but not always, be seen that one, two, or 
probably sometimes moi’e, chromosomes lag to such an extent 
that they are not included in the daughter-nuclei (figs. 29, 30). 
In some cases they appear not to divide; in others the halves 
separate so late that they are left behind in the cytoplasm. 
.Sometimes this causes the daughter-nuclei to have a tail-like 
