The Nucleus and Heredity. 
97 
the centrosomes of the asters of 
the quadripolar spindle. When 
the chromosomes split we shall 
have the arrangement shown in 
Fig. B. Two groups of chromo¬ 
somes will meet in eachblastomere 
and we shall finally have the four 
hlastomeres with the number of 
chromosomes shown in Fig. C. 
Now the chromosomes are distri¬ 
buted on the spindle by chance, 
so that although each blastomere 
may have the number (eighteen 
at least) necessary for ordinary 
development, yet the qualitative 
nature of the various chromo- 
26 
26 
V 
A 
• <—6 
. <—10 10 —> • 
Fig. B. 
32 18 
36 22 
12 
12 
A 
V 
Fig. C. 
somes meeting in a blastomere is 
a matter of accident. 
Boveri points out that the different behaviour of the different 
hlastomeres cannot be explained on centrosome or protoplasmic 
differences, for the cytoplasm is distributed equally between the 
four cells ; the difference must then depend on differences between 
the nuclei of the hlastomeres. The abnormalities cannot be 
explained on the ground of the number of the chromosomes being 
too large or too small, for an examination of the size of the nuclei 
shows that malformations are found in embryos with neither 
abnormally large nor abnormally small nuclei. The abnormalities 
must thus depend on qualitative differences between the nuclei. 
The qualitative difference seems only to be explained by the general 
qualitative difference of the chromosomes of a single nucleus. In 
ordinary daughter-nuclei there is no qualitative difference, because 
each has a half of each of the different chromosomes of the mother- 
nucleus. In the particular type of division in question there is an 
irregular and chance distribution of the chromosomes, leading to 
qualitative differences between the four nuclei. 
The interpretation then to be placed upon these experiments 
of Godlewski, Herbst and Boveri is that the nucleus is the controller 
of development as a whole, but the egg-protoplasm (probably owing 
to its activation by the female nucleus in earlier stages) is responsible 
in many cases for the first stages of embryonic development, which 
are thus purely maternal in character. There appears to be nothing 
