An on-farm trial is a strip that gets one change, planted beside a strip of your current practice that gets none. At an example price of $20 an acre, a foliar product costs $20,000 a season on 1,000 acres of corn. A strip read from memory tends to confirm what you already believed, and the product stays on the list. With a page written before planting, you can read the strip and find that the answer is no. Write that page, the trial sheet, in the hour before the planter goes out.
What goes on a trial sheet?
A trial sheet is one side of paper with five lines, filled in below for a foliar product on corn. None of the five lines chooses what to test. You already know what you want to try this year: a variety, a product, a machine setting. What rarely gets written is the question the trial should answer.
On a 1,000-acre corn farm in Ohio, the first line reads: does the foliar product raise yield enough to pay for itself? Ohio State University Extension frames on-farm studies the same way. Its factsheet on on-farm research starts with a concise, testable question, such as the impact of a foliar treatment compared to no treatment.
SARE, the US Department of Agriculture’s sustainable agriculture research and education program, adds a check. Before you commit the ground, make sure the answer is something you can measure. If you cannot weigh or count the answer, expect an argument at harvest instead of a result.
That $20,000 is why the question comes first. The trial exists to decide whether to spend it again next season.
What do you compare the change against?
Against your current practice, left unchanged, in the same season and on like ground. The University of Florida extension guide calls this strip the control and describes it as normally the current practice or the common cultivar. Last year’s field is not a control, because last year had different weather. Comparing this season’s strip with last year’s field measures the weather as much as the change.
Ohio State’s factsheet adds a second rule: only the variable being tested should change. On the Ohio farm, the second line reads: corn with no foliar product, same hybrid, same tillage, same planting date, in strips beside the treated ones.
Many farmers split the field in half, the Ohio State C.O.R.N. newsletter notes, and the split does not account for soil type changes, low spots or hills. The Mississippi State guide calls the split the most basic form of trial. It warns that the natural variation may not divide evenly between the halves, so any conclusions may not be valid. The same guide notes that strips make it more likely that variation is represented equally in every treatment.
Write the comparison on the sheet in words, with the same hybrid and the same tillage named. The unchanged strip is the comparison itself.
How many pairs, and where do they go?
A pair is one strip with the change and the unchanged strip beside it. Laying the pair down more than once is called replication. One pair can land on better ground, and the change then gets credit the soil earned. The Florida guide is clear that a large plot does not make up for a lack of replication.
The sources give slightly different counts. The C.O.R.N. newsletter requires at least three replications and encourages four. The Ohio State factsheet recommends five, with three to four as the minimum. The Mississippi State guide finds four to six complete blocks sufficient for most trials, and the Florida guide recommends four blocks. Four pairs sit inside all four ranges, so write four on the sheet.
Where they go matters as much. The Mississippi State guide advises putting the repeats in separate fields where possible. It also advises using prior yield maps, soil surveys and your knowledge of the field to find variation before you assign the strips. On the Ohio farm, the fourth line reads: two pairs in field A and two pairs in field B, each pair on like ground.
Draw the map on the sheet itself. The Mississippi State guide suggests labeling each strip and counting rows for each treatment, so strips are not lost when the crop is green and tall.
What do you measure at harvest, and can you trust a small gap?
Fix the measure before you know the answer. The Mississippi State guide says to decide beforehand what data answer the research question and let the results rest on those data. The measure should be something you can weigh. The C.O.R.N. newsletter lists the tools: a weigh wagon, a grain cart with scales or a calibrated yield monitor.
Judging a strip by crop health, appearance or greenness is much harder to evaluate objectively. A judgment of looks is also easy to bend toward what you hoped to see. On the Ohio farm, the third line reads: bushels per acre from the weigh wagon, one weight for each strip.
A small gap needs care. The Ohio State factsheet includes a sample report from its PLOTS app, an example and not a farm. In the report, a seeding-rate test found no significant differences in yield between 26,000 and 30,000 seeds per acre, even though one rate yielded five more bushels per acre. A few extra bushels can come from field variability and not from the treatment.
Four pairs show whether the gap repeats in every pair or appears in one. How to run the statistics is outside this article. The sheet needs the measure and the number of pairs.
Two more rules from the Florida guide go on the sheet. Keep a detailed log of procedures and management through the trial, and make no changes to the protocol once the strips are in. A trial adjusted mid-season can be made to agree with almost anything.
Who reads the result, and what do you write first?
Name one person. The Florida guide is written for extension agents who run trials with farmers. It names a failure that is not agronomic: it is very common for the agent to expect the farmer to check the trial, while the farmer waits for the agent to visit. The same gap can open between a farmer and a hired manager. The strips get harvested with the rest of the field, and the question stays unanswered.
The fifth line holds one name, a date and a signature, written before planting. If you keep a record of each field’s yield year by year, write the trial’s answer there, under that field and year.
Who the reader is depends on how the trial reached your ground. When your agronomist proposed it, the reader is often the agronomist. What your agronomist decides covers which decisions are whose. When a supplier or a company brought the trial in, who reads the result and who keeps the data go in a written agreement signed before the trial goes in. That agreement is a separate document from this sheet.
Write the question first, because every other line depends on it. Then work down the list.
Writing down before planting what a trial should answer, and who will answer it, is part of running a farm.