On-Farm Testing

Originally written February 1, 2006 | Last updated February 23, 2014

Farmers today have an increasing number of tools for managing crops. New developments in precision farming technologies, biotechnology, and advancements in pesticides, equipment, and other ag inputs are converging and arriving at the farm-gate at an unprecedented rate. Sifting through the overwhelming milieu of technologies to find the tools that really work is a challenge for farmers and the consultants and agronomists that serve and support production agriculture.

New treatments may not necessarily be superior in local areas, even though their average performance over a wider region is superior. Conducting a performance test on your farm or in cooperation with neighbors using similar management practices can help you select the best treatments for your operation.

Often farmers use technologies with little or no evaluation prior to use. Industry heavily invests in technology research and development, thus, ramp-up is fast and products are often marketed and distributed quickly in an attempt to recover investments in the early phases of technology adoption. Often farmers, usually at great expense, must learn and re-learn management of these technologies as new and improved versions are released.

The objective of an on-farm trial is to predict how different management options will perform compared to each other under your environment and cropping system. The process of testing a hypothesis and using the information gained in a cooperative, systematic manner has been highly successful in providing viable options for making production decisions on the farm. The process of incremental change and gradual improvements has evolved into a system of research, development and production never imagined just decades ago.

In general, there are two major categories of on-farm research trials. The first is replicated trials that try to account for field variability with repeated comparisons. Examples include trials conducted by universities and by public and private plant breeders. The other type is non-replicated demonstrations such as yield contests, on-farm yield claims, demonstration trials and farmer observation and experience.

Identifying the source of variability is a primary objective in any on-farm trial. The use of statistical methods including replication and mean comparisons improves the reliability and confidence of results. An overriding strength of on-farm evaluations is the credibility of the results in the eyes of the end user, the farmer by showing how the practice responds within his production system. Often the power of these trials can be enhanced with simple modifications such as replication within locations and across multiple sites with coordinated effort.

The advent of effective tools for collecting data related to crop production such as weigh wagons, on farm scales and yield monitors have removed many of the traditional barriers of on-farm trials. The next phase in the development of agriculture is necessary coordination of multi-site trials that will require collaborative specialists for data collection and analysis.

What is on-farm research?

What's special about on-farm trials? From a standpoint of actually doing trials, there is nothing special. The nature of on-farm trials limits the number of comparisons, doesn't allow some treatments that require special equipment, and requires timeliness/priority that at times can be an issue. It is, though, real field conditions. The main advantage for on-farm trials is that it makes it possible to test over a large number of sites, thus it can more easily move research from description to prediction. Knowing what we want to say when the work has been completed should be a critical precondition for undertaking on-farm, applied research. If you can't even guess what such a statement might look like, you might want to hold off. A certain percentage of all on-farm research projects are wasted effort or even harmful in that they suggest the use of harmful inputs, suggest the use of useless inputs, and fail to predict (accurately) the benefit of a useful input.

Terminology

Check (Control) Plots: Your current practice is represented in the check or control plot. It does not receive the new technology being tested; rather it represents your current management style where your tillage practice, applied fertilizer, variety and/or applied fungicide is used in the usual manner. The check plot and the treated plot differ only in the specific treatment comparison being made. Aside from this treatment, plots are managed exactly the same to avoid biasing results.

In some trials, the new technology incorporates several practices. Avoid these if possible. For example, consider a trial that compares a farmer's current planting operation with another planting operation using different tillage, fertilization and row spacing systems. A fair comparison can only be made between the two complete systems, not any given part of either system. This kind of trial is difficult to interpret because of all the confounding interactions that may occur among the parts.

Replication: Replication is used to determine whether the difference between plots is due to chance variation or treatment variation. Chance variation is caused by differences in weather, soil and other factors. These factors change significantly in space (field to field) and time (year to year). Through replication in both space and time, average treatment effect values can be obtained. Replication in space means that several plots of each treatment are grown in the field (field replication) or that single plots of each treatment are placed in several fields across the farm (farm replication). Replication in time is repeating the trial over several years. Comparisons between average values are more accurate than those between single plots. Replicating your check and treatment plots at least three or more times will give you much greater confidence in your results and final conclusions about a new practice if it is made only after being evaluated over several years and/or at several locations.

Randomization: Randomization prevents bias of any one treatment in any way (intended or unintended). To randomize a trial, randomly assign replicated check plots and treatments (Fig. 1) by drawing numbers out of a hat or flipping a coin as you assign treatments to plots.

How can on-farm research be conducted so that it means something?

The key to an on-farm test is that it must make repeated and unbiased side-by-side comparisons of the practices in question. These repeated comparisons are called replications. With replication, we can use simple statistical formulas to decide if one practice or "treatment" differs enough from the other to be sure the difference was not due to chance.

A well conducted performance test on your farm or in cooperation with neighbors using similar management practices is an important part of deciding upon new agronomic practices for efficient and profitable crop production. Key ingredients of an on-farm testing program include:

1. Establish goals and objectives.
2. Which treatments will be evaluated and what is the check (control) plot.
3. Select a site.
4. Lay out plots on the selected site.
5. Decide who, what where, when and how measurements will be collected.
6. Determine how data will be analyzed.
7. Extending the results.

Pre-season planning

On-farm trials must have a goal and specific objectives. Goals are statements of the overall theme of your experiment. A goal, for example, may be to reduce soil erosion on your farm. Objectives are statements of the problem you wish to evaluate in your project. These are the ideas you want to test or questions you want to answer. Objectives are measurable and relate to your overall goals. The objectives will determine what is measured and the type of data you will collect during the project. For example, you might hypothesize that a strip-tillage system will leave more surface residue that the tillage system you now use. The trial you establish would involve a comparison between the two systems. One objective will be to determine if the reside levels are different between the two tillage systems. Residue levels would be one type of data collected as part of the trial.

Selection of treatments to be tested: Keep treatments simple. Limit the number of treatments to no more than four (10 for a hybrid/variety trial), including one well-known treatment as a check (control) plot. As the number of treatments increase so does the complexity of the on-farm trial. Choose treatments that you expect to be significantly different. With experience you will gain confidence in your on-farm testing abilities and you can move on to testing treatments involving minor impact, difficult to test production practices. It is very important that production inputs remain constant, except for the tested treatments.

In field planning

Field variation can mask treatment differences, so chose as uniform a field site as possible. Organize the field plot layout so that that all treatments have an equal opportunity to perform. Consider previous crop history (fertilizer rates, herbicides, tillage, etc.), drainage, soil texture, soil depth, topography, pest infestations, and bordering influences such as trees, runoff from neighboring fields, lack of fencing from animals, and other factors. Avoid placing trials in runoff areas, near fence lines or in field corners as these areas are often subject to multiple or irregular applications of fertilizer and herbicides. The characteristics of a uniform field site depend on the type of test being conducted. Pay particular attention to things that strongly influence your treatments. Use county soil survey maps of the fields being considered to help you select the site.

Consider site access when selecting a plot location. Is the site easily accessible for mid-season treatment applications and data collection? If early or differential harvest is likely (such as with an alternate crop), can you get at the site with harvest equipment without destroying other crops? Will you hold a tour of your site? If so, is there ready access for visitors and their vehicles?

Selecting treatments. When selecting treatments, limit the number of treatments to no more than 10 and include two or three well-known check treatments. Select hybrids of similar maturity.

Replication and Randomization: Choose a uniform area in the field. Each practice should be repeated (replicated) two or more times in the trial. Two different layouts or designs are commonly used. Completely random designs are used on uniform sites. In this design, all treatments have an equal chance of being assigned to any given plot including identical treatments side by side. Use a randomized complete block design when it is not possible to obtain a uniform test site. Arrange the plots so that each treatment is used once in each replication. Treatments should be arranged in a different order in each replication (randomized). See Figure 1. Each replication contains a complete set of treatments. Each replication is placed in a uniform area. Using such an arrangement allows all treatments to have equal potential to perform within the replication. The effects of replications can be removed or blocked out when analyzing the data.

Plot size: Optimum plot size for on-farm tests may vary greatly with the size of uniform land area, number of treatments and size of equipment. Adjust plot lengths so that each treatment is within a reasonably uniform area or so that each uniformly covers the field variation as discussed above. Plots should be similar in size. Avoid using field edges in plots. Field edges should be left as borders. Plot width is determined by the width of equipment used to apply treatments (e.g., planter, sprayer, etc.) and/or harvest plots. The width of the established treatment should be larger than the harvest width. This way there will be a uniform harvest width and errors in harvesting will not affect side by side treatments. Typical treatment plots are between 1/10 and 1/2 acre.

Management: Each plot should be managed exactly the same and as close as possible to the conditions which normally occur on your farm.

Data collection and record keeping

It is important to plan ahead and identify what should be measured, and when and how to take measurements. What you will measure depends on the project's objectives. If the purpose is to increase yield, then a measure of yield is required. If the objective of a new practice is to increase soil moisture, then soil water tests are needed. If the purpose is to increase net farm profit, then you must analyze costs and returns (including yield).

If you need help in deciding what to measure and how to measure it, consult your county extension agent. Without appropriate data and a method to measure treatment differences, your trial will have little value or could lead to inaccurate conclusions. Detailed written records are often required to interpret data. Written documentation preserves the details of your on-farm trial so you can share information with others. Remember, the more you plan and document, the greater confidence you will have in your results.

The following information includes the baseline data needed to document and interpret a valid, unbiased test.

Trial description: State the goals, objectives, treatments and experimental design of the trial.

Field history: Record obvious variations within the test site and the previous cropping history including crop rotation, tillage practices, previous crop and variety, and previous fertilizer and pesticides applied. Make a diagram showing the layout of the field trial.

Soil test and fertility: Sample soil using university guidelines and send samples to a laboratory for analysis. Apply appropriate fertilizer uniformly to the entire trial area.

Field notes and observations: Keep accurate and up-to-date records. Walk the area every few weeks during the season. Note obvious strengths and weaknesses of the treatments plus any general problems with the test. Record any conditions that might influence stand establishment including variety, seeding, planting date, soil temperature, type of planter, seeding depth, row spacing, and residue levels. Take notes of other field operations, such as the type of equipment, depth of tillage operations and materials applied.

Weather: Record significant weather events such as high winds, frost and hail. Precipitation is more variable than temperature between sites. If practical, place a rain gauge at the test site with a little oil in it to prevent water from evaporating and after each storm, record rainfall and empty the rain gauge.

Insects, Weeds and Disease: Record the presence and density of pest (diseases, insects and weeds) and the extent or severity of crop damage. Note differences between treatments, if any.

Crop Growth and Development: During the growing season record the crop stage at the time of treatment and management operation applications, such as pesticide spraying. When abnormal conditions occur, such as drought, note the differences in plant growth response among treatments. It is just as important to record no differences among treatments at a certain growth stage as it is to record obvious differences.

Measurements: Depending on the test, take stand counts of the crop, ratings of weed control, disease or insects. Weigh the yield of each plot, take a moisture sample, and adjust yields to the same moisture content. Yield estimates are needed to make production and economic comparisons between treatments. Measure the size of the harvest area using a measuring tape or before or immediately after you harvest each plot. These distances are then multiplied by the width of the combine head to arrive at the harvested area and yield per acre.

Harvest the middle area of each treatment plot so that border effects do not bias the results. Yields can be measured with a local truck scale, a weigh wagon, or a properly calibrated yield monitor. Harvest equipment must be completely empty and clean before each treatment is harvested. Save a sample from each treatment to determine moisture content at harvest and any other quality factors that may be important such as test weight and protein content. If moisture contents differ between the treatments, you must be corrected to constant moisture.

Data analysis and Statistics

Data analysis largely depends on how the project was designed and conducted. Simple statistical software packages are available. Microsoft Excel has a very good analysis of variance procedure.

Economics: Use a partial budget analysis where

Grower return = (Yield*Price) - [Yield * (Handling+ Hauling+ Storage+ Drying+ Trucking)]

• Price = Weighted Price per Bushel = 50% November 15 Average Cash price + 25% March CBOT Futures price ($0.15 basis) + 25% July CBOT Futures price ($0.10 basis). November 15 Average Cash price derived from Wisconsin Ag Statistics; CBOT Futures prices derived from closing price on first business day in December.
• Handling costs = $0.02 per bushel
• Hauling costs = $0.04 per bushel
• Storage costs = $0.02 per bushel for 30 days
• Drying costs = $0.02 per bushel per point of moisture
• Trucking costs = $0.11 per bushel for 100 miles

Analysis: Use averages over replicates to compare treatments. A well-conducted test will have small differences among plots of the same treatment and some large difference between treatment averages. Consider all-important traits and not just yield.

Variations in yield and other measurements because of variations in soil and other growing conditions lower the precision of the results. Statistical analysis makes it possible to determine, with known probabilities of error, whether a difference is real or whether it might have occurred by chance.

Means are often separated using a number labeled "LSD" which stands for least significant difference. LSD's at an appropriate level of confidence (usually 10%) are used. Where the difference between two selected treatments within a column is equal to or greater than the LSD value at the bottom of the column, you can be sure that in nine out of ten chances that there is a real difference between the two treatment averages. If the difference is less than the LSD value, the difference may still be real, but the experiment has produced no evidence of real differences.

Statistics are only a tool to help prevent us from deceiving ourselves and others. Growing conditions in any particular year can have large effects on certain practices. Two years of replicated data are a minimum for supporting most practices. Statistics, as commonly used, often describe better than they predict. But, stats used over a lot of site-years can provide a measure of the usefulness of a prediction based on data. And yet, statistical statements always involve probability, and this is not always easy to "apply" when it comes to inputs. Statistics do NOT substitute for the large amount of data (site-years) that good on-farm testing always requires.

Time: Data from one field in one year may be misleading. About two to three years of your own tests in conjunction with other reliable information should be adequate to select treatments to be practiced on larger acreage. One year's data may be adequate to discard poor treatments from the test. Replace discarded treatments with new treatments in any tests conducted next year.

Adjustments in the number site-years should be considered if expected variability due to soil type is high (then you need more soil types), if expected variability due to years (weather) is high (then you need more years), and/or if variability is expected to be high over both soils and years, (then you will need a lot of sites and years). If variability is expected to be high over varieties/hybrids, you have a problem.

Summary of results

To interpret results from your on-farm trial, carefully summarize management history, data collected, and observations made. Summary forms are provided in the back of this guide for this purpose. The summarized results should address your goals and objectives. If your objective was to reduce costs, equal or even lower yields may be an acceptable result as long as costs are reduced and the net return has improved.

Take the time to share the results with your neighbors and county extension agents. This flow of information and experience is necessary for the progress of agricultural production and management.

On-farm testing is not a quick cure for anything, but it should greatly accelerate innovation and adoption of new practices by providing reliable, quantitative answers that apply directly to a producers situation. Treatments frequently differ in performance and these differences may vary with management practices, weather patterns, soil conditions, and other environmental and management practices. Replicated trials which take into account field variability are the most reliable and need to be part of informed selection of new practices for profitable crop production.

Lessons from experience

1. If possible, include the cooperator's current practice
2.  You can't over document a trial: maps, plot layout, field markers, distances, etc. A field at planting looks a lot different five months later.
3. Communicate with the cooperator early and often - planting and harvest are especially important
4. Plan multiple locations......you'll probably lose a couple
5. Be nice to people with scales and weigh wagons. You need them.
6. Have an efficient data collection system.
7. Visit the trial during the growing season. Measure distances, take stand counts, look for environmental and manmade problems. Document location of potential problems
8. Send a letter of thanks and the results to the cooperator
9. Purchase, borrow, make, or steal equipment when you can wheel, 300 ft. tape, generator, large shop vacuum, PVC stand count stick, PDA, moisture tester, sprayer, etc.
10.  THE COOPERATOR can make or break a project. It helps if they're interested in the result. Be willing to accommodate the cooperator's schedule.
11. Commitment to the project......a poor trial is worse than no trial

Evaluating Corn Hybrid Demonstration Plots

This is the time of year when many farmers visit and evaluate hybrid demonstration plots planted by seed companies and county Extension personnel, among others. When checking out these plots, it's important to keep in mind their relative value and limitations. Demonstration plots may be useful in providing information on certain hybrid traits, especially those that are usually not reported in state corn performance summaries. The following are some hybrid characteristics to consider while checking out hybrid demo plots.

PLANT/EAR HEIGHT. Corn reaches it maximum plant height soon after tasseling occurs. Remember that although a big tall hybrid may have a lot of eye appeal, it may also be more prone to stalk lodging in the fall. Unless your interest is primarily silage production, increasing plant height should not be a major concern. Generally later maturity hybrids are taller than earlier maturity hybrids. Big ears placed head high on a plant translate to a high center of gravity, predisposing a plant to potential lodging. The negative effects of stalk rot on stalk lodging in the fall may be worsened by high ear placement. Plots that have been subjected to early season (V7 or earlier) defoliation caused by hail or frost often have lower than normal ear height.

STALK SIZE. Generally speaking, a thicker stalk is preferable to a thinner one in terms of overall stalk strength and resistance to stalk lodging. As you inspect a test plot, you will see distinct differences among hybrids for stalk diameter. However, also check that the hybrids are planted at similar populations. As population increases stalk diameter generally decreases. Keep in mind that uneven emergence and development may make such comparisons difficult because late emerging plants are "spindlier".

DISEASES. During the grain fill period, leaf diseases can cause serious yield reductions and predispose corn to stalk rot and lodging problems at maturity. Ear rots can also impact yield and grain quality. The onset of leaf death shortly after pollination can be devastating to potential yield, since maximum photosynthetic leaf surface is needed to optimize grain yield. Hybrids can vary considerably in their ability to resist infection by these diseases. Demonstration plots provide an excellent opportunity to compare differences among hybrids to disease problems that have only occurred on a localized basis. Look for differences in resistance to northern corn leaf blight, gray leaf spot, and diplodia ear rot.

Check to see if foliar fungicides have been applied and what crop rotation has been followed. Typically you'll encounter more severe foliar disease problems in no-till, continuous corn.

STALK ROTS. Hybrids will likely differ widely when faced with strong stalk rot pressure. Begin checking plants in late August or about 6 weeks after pollination by pinching lower stalk internodes with your thumb and forefinger. Stalks that collapse easily are a sure indicator of stalk rot. Remember that hybrids with thicker stalks may be in plots having thin stands.

LODGING. Perhaps as important as stalk rot resistance is the stalk strength characteristics of a hybrid. Sometimes, superior stalk strength will limit the adverse effects of stalk rot. If your variety plot is affected by stalk rot in late August and early September, evaluate stalk lodging of the different hybrids. Most agronomists characterize plants with stalks broken below the ear as "stalk lodged" plants. In contrast, corn stalks leaning 30 degrees or more from the center are generally described as "root lodged" plants; broken stalks are usually not involved. Root lodging can occur as early as the mid-to- late vegetative stages (as it did this year) and as late as harvest maturity. Both stalk and root lodging can be affected by hybrid susceptibility, environmental stress (drought), insect and disease injury.

Root lodging may be associated with western corn rootworm injury. However, much root lodging occurs as the result of other factors, i.e. when a hybrid susceptible to root lodging is hit by a severe windstorm. A hybrid may be particularly sensitive to root lodging yet very resistant to stalk lodging. A cornfield may exhibit extensive root lodging in July but show little or no evidence of root lodging at harvest maturity in September (except for a slight "goose necking" at the base of the plant).

TRANSGENIC TRAITS: Because damage from European corn borer (ECB) and western corn rootworm (RW) can be very localized, strip plot demonstrations may be one of the best ways to assess the advantages of ECB Bt and RW Bt corns. The potential benefit of the ECB Bt trait is likely to be most evident in plots planted very early or very late; the potential benefit of the RW Bt trait is likely to be most evident in plots planted following corn or in a field where the first year western corn rootworm variant is present.

HUSK COVERAGE/EAR ANGLE. Hybrids will vary for completeness of husk coverage on the ear as well as tightness of the husk leaves around the ear. Ears protrude from the husk leaves are susceptible to insect and bird feeding. Husks that remain tight around the ear delay field drydown of the grain. Hybrids with upright ears are often associated with short shanks that may be more prone to ear and kernel rots than those ears that point down after maturity. Differences in ear "orientation" among hybrids can be strongly influenced by growing season and plant density. Also, under certain environmental conditions, some hybrids are more prone to drop ears, a major problem if harvesting is delayed.

The following are some additional points to consider during your plot evaluations:

1. Field variability alone can easily account for differences of 10 to 50 bushels per acre. Be extremely wary of strip plots that are not replicated, or only have check or tester hybrids inserted between every 5 to 10 hybrids. The best test plots are replicated (with all hybrids replicated at least three times).
2. Don't put much stock in results from ONE LOCATION AND ONE YEAR, even if the trial is well run and reliable. This is especially important in years with tremendous variability in growing conditions (e.g. planting dates) across the state. Don't overemphasize results from ONE TYPE OF TRIAL. Use data and observations from university trials, local demonstration plots, and then your own on-farm trials to look for consistent trends.
3. Initial appearances can be deceiving, especially visual assessments! Use field days to make careful observations and ask questions, but reserve decisions concerning hybrid selection until you've seen performance results.
4. Walk into plots and check plant populations. Hybrids with large ears or two ears/plant may have thin stands.
5. Break ears in two to check relative kernel development of different hybrids. Use kernel milk line development to compare relative maturity of hybrids if hybrids have not yet reached black layer. Hybrids that look most healthy and green may be more immature than others. Don't confuse good late season plant health (stay green) with late maturity.
6. Differences in standability will not show up until later in the season and/or until after a windstorm. Pinch or split the lower stalk to see whether the stalk pith is beginning to rot.
7. Visual observations of kernel set, ear-tip fill, ear length, number of kernel rows and kernel depth, etc. may provide some approximate basis for comparisons among hybrids but may not indicate much about actual yield potential. Hybrid differences are common for tip kernel abortion ("tip dieback" or "tip-back") and "zipper ears" (missing kernel rows). Even if corn ear tips are not filled completely, due to poor pollination or kernel abortion, yield potential may not be affected significantly, if at all, because the numbers of kernels per row may still be above normal.
8. Find out if the seed treatments (seed applied fungicides and insecticides) applied varied among hybrids planted, e.g. were the hybrids treated with the same seed applied insecticide at the same rate? Differences in treatments may affect final stand and injury caused by insects and diseases.

What are the implications? How far can the research take you?

On-farm testing is not a quick cure for anything, but it should greatly accelerate innovation and adoption of new practices by providing reliable, quantitative answers that apply directly to a producer's situation. Treatments frequently differ in performance and these differences may vary with management practices, weather patterns, soil conditions, and other environmental and management practices. Replicated trials that take into account field variability are more reliable than non-replicated trials and improve the confidence of implementing of new practices for profitable crop production.