PRODUCTIVENESS OF SELF-FERTILIZED CORN 11 
plants of a 7-4-1- line were mailed to Knoxville from Washington, 
D. C, and one additional cross with 10-3- was obtained in this way. 
Failure of lines to blossom at the same time and damage by storm, 
insects, and diseases prevented obtaining as many crosses as desired. 
Enough combinations were obtained, however, to determine the rela- 
tive value of the different lines in so far as major differences are 
concerned. 
METHOD OF COMPARISON 
The crosses were grown at Knoxville in 1923. One crossed ear 
represented each combination between individual lines, and the pro- 
ductiveness of this cross was compared directly with that of the 
variety by the hill-checking method (5). One seed of the cross and 
two seeds of the check were planted in each hill, the two kinds being 
about 6 to 8 inches apart and the check seed always being toward the 
southeast end of the row. A row was 120 hills long, and as many 
hills in the row were planted in the way described as there were seeds 
on the crossed ear, the rest of the row being planted with check seed. 
The plat was thinned to a stand of one crossed plant and one check 
plant in each hill, facts permitting, when the plants were 8 to 10 
inches high. The rows were gone over just before harvest, and all 
hills not containing one crossed plant and one check plant were 
eliminated. The remaining perfect hills then were harvested in 10-hill 
sections, the product of the plants from crossed seed and check seed 
in each section being harvested and weighed separately. The relative 
productiveness of the different crosses is compared through average 
superiority or inferiority to the check. 
The ears from the last 18 to 20 plants of the cross in each row were 
stored until dry. They were then reweighed and shelled, and the 
percentages of air-dry shelled grain were determined. The shrink- 
age and shelling percentages of the check were determined from six 
similar samples. The field weights were computed to terms of air-dry 
shelled grain on the basis of these data. The drying and shelling 
samples of some of the crosses, together with the product of a cor- 
responding number of check plants, are shown in Plates VII and VIII. 
These illustrate the excellent quality of the ears produced by some 
of the crosses. 
EXPERIMENTAL DATA 
A summary of the data is given in Table 6, the crosses being ar- 
ranged in the descending order of their relative productiveness. 
Column 1 shows the field-row number and is of interest in indicating 
any possible effect that location in the plat may have had on relative 
productiveness. There is some tendency for data from rows that 
occurred near each other in the field to occur near each other in the 
table, but examination of the pedigrees in column 2 shows that this 
probably is due largely to similar combinations having been grouped 
in the field to some extent. 
The total number of perfect hills given in column 3 indicates also 
the number of replications used in determining the probable errors. 
The yield of any one to four plants beyond the last multiple of 10 
was added to the yield of the last 10-plant replicate and the sum 
divided by the number of plants involved and multiplied by 10 to 
bring the yield to a 10-plant basis. If there were five or more such 
odd plants their yield was divided by the number, and the quotient, 
multiplied by 10, was treated as another replicate. 
