Count ears, count kernels, divide by kernel size
The estimate is built from three field counts. Measure 1/1000th of an acre, count the harvestable ears in it, count kernels on a few representative ears, then divide by how many kernels make a bushel.
Ears in 1/1000th acre times 1000 gives ears per acre. Kernels per ear is rows around the cob times kernels down a row. The 90,000 is the kernels-per-bushel factor, which you slide up or down for a stressed or bumper year.
Row length that equals 1/1000th of an acre
| Row spacing | Length to walk |
|---|---|
| 30 in | 17 ft 5 in |
| 36 in | 14 ft 6 in |
| 38 in | 13 ft 9 in |
| 20 in | 26 ft 1 in |
| 15 in | 34 ft 10 in |
Count the ears in one length of a single row at your row width. That count times 1000 is your ears per acre.
Kernel size is the factor that swings the answer
A typical ear at 16 rows × 32 kernels is 512 kernels. Only the divisor changes with the season, and it moves the estimate a lot.
- Good year (80,000). Full, heavy kernels. Same counts give a higher bushel estimate.
- Average (90,000). The default when grain fill looks normal.
- Stressed (100,000+). Drought or disease shrinks kernels, so more of them fit in a bushel and the estimate drops.
Sample at least five spots across the field and average. Skips, drought pockets and end rows will bias a single count.
Common questions
What is the yield component method?
It estimates corn grain yield before harvest by counting the ears in a small measured area, counting the kernels on a sample of those ears, and dividing by a kernel-size factor. It works from the milk stage (R3) onward, so you can forecast weeks before combining.
What is the fudge factor of 90,000?
That is the assumed number of kernels in a 56-pound bushel of grain. Big, well-filled kernels in a good year run nearer 75,000 to 80,000; small kernels in a stressed year run 95,000 to 105,000. Lowering the factor raises the yield estimate, so pick it to match the season.
How accurate is a pre-harvest yield estimate?
Within about 20 bushels an acre if you sample well. The main errors are guessing kernel size before grain fill is complete and taking too few samples in a variable field. Count at least five spots and average them.
