1. The data obtained from our experiment on the
frequency of mutations at specif }c loci with two
dosages of x -rays, of 1000 r and 4000 r,
respectively, and involving the examination of
91726 F₁ females in all, are reported.
2. Included in our study were gross rearrangements
of three types - (a) translocations of the fourth
chromosome giving a cubitus "position effect", (b)
rearrangements involving at least one break in some
heterochromatic region and one in a very limited
part of the "active" region, detectable by their
influence on somatic variegation (in this case, by
causing mottled alleles of white), (c) large
deletions of the X-chromosome, giving exceptional
F₁ males. Although all these of these types of
changes have been reported by certain previous
workers to occur in frequencies directly proportional
to the dosage, like gene mutations, and unlike what
has been reported by our colleagues for the commoner
types of gross rearrangements, our data show clearly
that the frequency of the gross rearrangements of
these special classes varies at a rate that is
definitely higher than the first power and lower
than the second power of the dosage used, and that
is most probably about the 1.5 power. Both the
breakage 'and contact hypotheses are discussed, and
it is pointed out that the power 1.5 is that to be
expected on the breakage hypothesis, according to
which the breakages involved in gross rearrangement
occur first, as a result of separate individual
ionizations, and union of the broken ends occurs
later. Moreover, there is no theoretical reason
to expect the three types of gross rearrangements
here studied to follow a different rule from other
gross rearrangements in this respect.
3. In the sane collection of flies in which we
found the above exponential relationship for gross
structural changes, the frequency of minute
rearrangements - both of those (yellow in sc⁸)
involving heterochromatic regions, and those
(Notches and reverse Bar) involving euchromatic
regions-proved to be linearly proportional to the
dosage.
4. It was also found that the frequency of
mutations involving recessive visible effects
allelic to known mutations was linearly proportional
to the dosage. As these mutations were found to
include a high proportion of lethals, which must
predominantly have been minute deletions, it nay
be inferred that the latter minute rearrangements
also, like gene mutations, have a linearly
proportional frequency-dosage relation.
5. It is concluded that in the production of the
minute rearrangements by radiation, breakage at
two or more nearby points in the chromatin results
from a spreading of the effect of one and the same
ionization. It is conceivable (but not necessarily
true) that in the production of gene mutations
essentially the same process takes place, but on
a still minuter scale, involving breakage and
rearrangement within what have been considered single genes. Other methods will be needed to
attack this latter question, however.
6. The data show clearly that the loci of
different "visible" mutations differ in regard to
the frequency with which surviving mutants,
deviating in regard to visible characters, are
produced from their normal alleles.
