Every taken a statistics class? I recommend it, done right you can get good enough results with just a few samples. There is no need for 19,200,000 (6000 acres at 32000 plants per acre, reasonable numbers for a large corn grower) samples, you can't look at them all anyway. Turns out that farmer can get by with far less samples with the right analysis.

University and seed researchers sometimes do the plant by plant analysis you are taking about. They are looking at just a few hundred samples though and only getting slightly better data.