466 H. J. MULLER 
in Drosophila is about three days. The duration here involved 
is obviously a very important factor in determining the great 
excess of the double line over the single; the combined lengths 
of these parts of the double line are by themselves about equal 
to the total length of the single line in the diagram. In addi- 
tion to this, it is also likely that the four divisions which give 
rise to the odcyst, and which evidently occur just before the 
growth period, consume a comparable time, as represented in 
the diagram. But there is still another circumstance that would 
tend to result in an even greater preponderance of the double 
line than that figured. For it may be shown that, if the actual 
cell lineage is at all of the type which has been pictured, the 
disproportion between the double and the single lines must 
increase with an increasing number of eggs, so that the actual 
disproportion would probably be very considerably greater than 
that shown in the simplified diagram above. 
If, however, the celi lineage does not conform in its general 
pattern to that pictured in figure 1, it almost certainly does not 
differ from the latter more than do the two opposite extreme 
types of cell lineage indicated in figures 2 and 3. A scheme of 
the type shown in figure 2, if it were so constructed as to allow 
the six days above assumed for the multiplication and matura- 
tion periods, and also so as to give the same number of eggs as 
figure 1 in the same length of time, would result in a ratio of 
double to single line of not less than 2 to 1. The ratio on figure 
2 may vary from this minimum figure, however, all the way up 
to an infinitely large number to one; the exact number will 
depend upon which cell generations the germ cells ancestral to 
the various eggs remain longest at rest in. A scheme like that 
of figure 3, if constructed for the same multiplication and growth 
period and for the same number of eggs in the given time as 
figure 1 would result in a ratio of not less than 12 to 1. The 
actual cell lineage is evidently something between the two 
extreme types of 2 and 3 and it is therefore certain at least that 
the chance for the production of single mutants is considerably 
greater than that for the production of twin or multiple mutants. 
Accordingly, the actual finding that more cases of single mutants 
