56 BULLETIN 1489, XT. S. DEPARTMENT OF AGRICULTURE 
cially, however, are those which are more productive than the best 
varieties available, and the proportion of such crosses has been rela- 
tively small. Crossing varieties of corn at random can not be recom- 
mended as a means of obtaining maximum yields. The superior com- 
binations must be determined by careful experiment over a period of 
years. Even then there usually is no assurance that this superiority 
will be maintained, as the parent varieties themselves are but mix- 
tures of hybrids and therefore grossly unstable. Varietal crossing 
may have a place in obtaining larger yields of corn when the object 
is to bring into expression in a single stock favorable dominant char- 
acters from each of two varieties. Thus, there is some evidence that 
the partial dominance of quick maturity in small-growing varieties 
and of larger size in sorts that require a longer season to mature may 
both be expressed in the crosses between these types. If so, such 
crosses might have an important place near the northern limits of 
corn production. Aside from such special cases, however, the use of 
F-l varietal crosses seems to offer little that is practical in the way of 
corn improvement. 
Selection within selfed lines is a genetically sound basis for corn 
improvement, and it is by methods involving selection within selfed 
lines that corn breeders of to-day are attempting to obtain materially 
larger acre yields of corn. Preliminary results of these attempts 
indicate that crosses or double crosses between selfed lines can be 
used to obtain consistent increases of 20 to 30 per cent or more above 
the acre yields of the best present varieties. 
The crossed seed from which the large yields are obtained must 
be grown each year, and farmers using such seed would have to 
obtain a new supply either by purchase or by growing it themselves 
in special seed plats. This involves difficulties in seed production 
and distribution that must be overcome if crossed seed is to be used 
on a large scale. Seedsmen will be able to supply part of the de- 
mand. If the purchase of seed became at all general, however, only 
a small part of the approximately 20,000,000 bushels of seed corn 
that is used in the United States annually could be supplied through 
existing commercial channels. Individual farmers in different sec- 
tions could specialize in the production of crossed seed for sale. It 
also seems probable that groups of farmers could combine for the 
purpose of producing enough crossed seed for their own use. These 
are only suggestions. The best methods will be worked out only 
when crosses of high productiveness are known and the demand for 
seed of them is created. There is little doubt that ways will be 
found to meet the demand when it exists. 
Before the better methods of producing and distributing crossed 
seed can be determined, high-yielding crosses must be found for 
various sections of the United States. Very few crosses as yet have 
been tested thoroughly enough to warrant their recommendation for 
commercial planting. Experimental corn breeding involving selec- 
tion within selfed lines is being conducted by the specialists of the 
Bureau of Plant Industry and of most of the State agricultural 
experiment stations. Not only are attempts being made to find lines 
which will produce high-yielding crosses, but different methods of 
utilizing selfed lines are being compared and studies are being made 
of the factors that cause lines to be good or poor in the selfed condi- 
