MODIFIED EAR TO ROW BREEDING FOR CORN
Grow out a large population that can openly pollinate. Harvest ears from the best plants (as above) and save the seeds from each ear separately (at least 200 ears). This is your base population. Now plant out 20 seeds from each ear in the following season in replicated plots (separate "ear" rows). The order of the ear rows is randomized differently in each field:
In the third field, the ear rows to be tested are divided up into groups of four rows. In between each of these groups is a row of plants that are planted from a bulk of the seed from all of the ears. In this field we are evaluating the yield and plant type of each row, just like in the first two fields, but we are also detasseling the tested rows and allowing them to be pollinated from the bulk seed rows. This makes the third field a "crossing block" too. When the ear rows are picked in each field we estimate yield for each row. In the third field we also mark or paint ears from especially desirable plants in each row and save those ears after we weigh all the ears from each of the ear rows. After calculating an average yield across the three fields, we then choose the marked ears from the 10-30 best performing ear rows overall. These provide the ears that go into the next season of selection. The rest of the seed can be planted out for seed production or fed off. This is repeated annually and can achieve rates of gain in yield that are 2-3 times that of mass selection, and can improve standing ability if you select for rows that yield and stand well.