468 H. J. MULLER 
individuals formed in the two kinds of cases would be alike, but 
the number of different cases of mutation would be much greater 
for the later mutations than for those occurring at the earlier 
stages. 
It may be observed that the chance for the observation of 
half or quarter mutants, although these are due to mutations in 
very early stages, is somewhat high as compared with the 
chance for multiple mutants. The chance for such mosaics 
may be estimated by determining the total length of the dotted 
lines of the offspring; in the diagram the ratio of the lengths of 
dotted to single lines is about 7:5. The reason that mosaics 
can be found so often is because the mosaic condition may be 
observed directly in the individuals (F:) in which the mutations 
arose, whereas mutations in the germ tract can only be detected 
if they occurred in a parent (P;). There is, however, a chance 
that the mutated somatic cell will not happen to come to lie in 
a part of the body that would be affected by the mutation it 
bears. For example, if the ‘new white’ cells had not happened 
to include any eye tissue, the mutant would not have been dis- 
covered. For this reason, a correction must be applied to the 
criterion of ‘dotted line length’ referred to above, depending on 
what proportion of the early somatic nuclei have descendants 
in parts of the body in which the mutation could be observed. 
Another important correction, of opposite effect, would consist 
in the addition to the dotted line of the early cleavages of all 
those later somatic lines within which the mutation would pro- 
duce a visible effect. These corrections will vary greatly for 
different kinds of characters. 
With the exception, then, of cases of mosaics and other 
somatic mutations, the general conclusion may be drawn that 
the number of cases of mutation observed should become pro- 
gressively greater the later the occurrence of the mutations in 
the germ-tract cycle, even though cells at all stages may be 
equally likely to mutate. 
Furthermore, the evidence showing that mutation may occur 
during various cell generations in the life-cycle makes it highly 
probable that older animals, or animals of species having a 
