14 BULLETIN 1396, U. S. DEPARTMENT OF AGRICULTURE 
seems probable that the prepotency of this strain is the explanation 
of the high performance of the F, and F, in 1924. 
Figure 7 shows the descent of the progenies in the selfed experi- 
ment up to and including the plantings in 1924. The progenies are 
numbered each year in the order of their standing with respect to ear 
length. The portion of the diagram within the dotted line shows 
the descendants of progeny No. 1, 1921. 
DISCUSSION 
In the light of the modern knowledge of inheritance, theories of 
maize improvement fall rather naturally into three groups, depending 
on whether departures from the maximum vigor are assumed to be 
due to (1) mutation, (2) the segregation of simple Mendelian char- 
acters, or (3) the segregation of multiple-factor characters. 
MUTATIONS 
Attempts to apply the findings of genetics to the improvement of 
maize have been directed almost exclusively to the elimination of 
existing variations. Indeed, if mutations are of sufficiently frequent 
occurrence to be the important factor, existing knowledge of the laws 
of genetics can be of little aid to breeding. The proper procedure 
would be that which reduces the occurrence of mutations to a mini- 
mum, and at present practically nothing is known regarding the 
causes of mutation. 
The selfing of maize varieties brought to light such large numbers 
of, existing recessive characters that the appearance of new varia- 
tions by mutation has been disregarded to a large extent. If dele- 
terious mutations are taking place with any great frequency the im- 
provement of varieties by the combination of inbred strains may 
afford relief so temporary as to be economically impracticable. It 
should be kept in mind also that mutations may be more frequent 
in homozygous than in heterozygous strains. There is no direct 
evidence on this point, and if the findings of Emerson (2) with re- 
spect to somatic mutations of the variegated character in the peri- 
carp are any index of gametic mutations, no increase in the rate of 
mutations need be expected to follow selfing. 
SEGREGATION OF SIMPLE MENDELIAN CHARACTERS 
The hope that maize may be improved by intensive inbreeding is 
based very largely on two facts: (1) The continuous appearance 
of unproductive off-type plants even in carefully selected crossbred 
stocks, and the behavior of these variations as Mendelian characters; 
and (2) heterosis, or the vigor of F, hybrids. 
The generally accepted view of heterosis that followed the work 
of Bruce (1), Keeble and Pellew (4),.and Jones (3), is that the first 
and second facts mentioned are essentially the same phenomenon, 
the only difference being one of magnitude in the effect produced by 
. the genetic factors involved. Admitting thisfundamental similarity, 
however, the indicated breeding methods would differ profoundly, 
depending on whether the object was the elimination of a compara- 
tively few recognizable recessive characters or a multitude of reces- 
sive characters too small to be detected individually. 
