4. Genetic diversity of 12-row lines. Data indicating genetic heterogeneity of certain 12-row inbred lines of maize have long been available, but have not been reported heretofore in this "unpublished publication". A brief summary of some of these data follow.

Numerous 12-row lines have been obtained from various sources. Some (A, B, G, b, c) from crosses of 8-row flints with 16-row dents and others (III, IV, VI, VII) by selection from varieties of dent, flint, and popcorn. No 12-row type produces only 12-row ears. There are always some 10-row and 14-row ears and occasionally an 8-row or a 16-row ear. To determine whether a 12-row line is homozygous it is necessary to grow progenies from selfed ears of the more extreme variants. To get such selfed ears it is necessary to hand-pollinate many plants. This has been accomplished for the 12-row types involved in this account. A single example of the results obtained is given here.

Line b had in F₉ a distribution ranging from 8 to 16 rows with frequencies of 4-18-49-14-1. In F₁₀, progenies from selfed ears of diverse row numbers were produced as follows:

The other lines gave similar results. It was concluded, therefore, that all were approximately homozygous. When any two of the nine 12-row lines were crossed, except only b × c, it was easily possible to establish lines of different row number. For example, the cross b × IV exhibited row-number ranges from 10 to 16 in F, and 8 to 18 in F₂ with these frequencies, respectively, 1-43-8-1 and 2-12-37-21-7-2. In F₃ the following frequency distributions were observed:

Given different genes for row number in the several 12-row types, it should be possible, by multiple crossing followed by selection, to assemble the row-number genes of the several 12-row lines into a single line of high row number. In the accompanying table are shown the frequency distributions of all of the ears produced by seven inbred lines during several generations when selected for twelve rows and similar data from certain single, double, and multiple crosses of these lines when selected for high row number.

The seven inbred lines had frequency distributions ranging mostly from 8 to 16 rows with strong modes at 12 rows and means very near 12. During the five to eight generations shown in the table and among the total of more than six thousand ears, not a single ear had more than 16 rows. After repeated intermittent crossing followed by selfing with selection for high row number, lines have finally been established with modes at 24 rows and means near 23. Two of these lines have not produced an ear with so few as 16 rows.

Frequency distributions of number of kernel rows of inbred lines
and their single, double, and multiple crosses

I am now ready to admit that number of kernel rows in maize is a much more complex quantitative character than I assumed it to be when I began a study of its inheritance.

R. A. Emerson