How to Get Results You Can Rely On
By Carol Miller, Editor, American Vegetable Grower
A big challenge facing growers when buying biological controls is a dearth of public trials and research. Years of attrition at university Extension offices — in personnel and funding — have taken a toll on how much research takes place.
To an extent, that has pushed testing new products onto growers. The only problem? Growers aren’t trained researchers. And that training is needed to help them gain reliable results when so many factors can impact how any given field or plot’s crops behave, from weather, to sprays drifting from adjacent areas, to variations in soil health.
So, we turned to one of the top vegetable pathologists in the country, Anthony Keinath, Professor of Plant Pathology and Research and Extension Vegetable Pathologist with Clemson University. He has tested both conventional and biological solutions for years, and hard-earned experience has taught him a thing or two.
“I am a scientist. And so I am very data driven. I think I've become more that way over the years, just from having published a lot. And you need to have numbers to back up your conclusions. I've sort of taken this approach toward on-farm research as well,” Keinath says.
After giving some thought to how growers can succeed on their own, Keinath came up with four keys to success.
“Let’s talk about them in order of importance,” Keinath says.
1. Always Have a Control
“A lot of growers want to test a new biopesticide, and they’ll go out and out and spray their 80-acre field,” Keinath says.
On the surface, it makes sense. If the field flourishes, then you might conclude that the product works.
But without an untreated section to compare the results to, it’s just as likely that other factors had a bigger impact. Especially when you are testing for vegetable disease treatments.
“That probably helped a lot. You don’t know whether it was a change of the environment that was not favorable for the disease so it just naturally slowed down — or did the biopesticide protect new growth?”
Keinath gave an example of a tomato grower who once thanked him profusely for curing his late blight problems. Later on, Keinath’s technician privately told him the temperature shot up the week following spraying.
“That probably helped a lot,” Keinath said. “You don’t know whether it was a change of the environment that was not favorable for the disease so it just naturally slowed down — or did the biopesticide protect new growth?”
An untreated plot could have answered that question for you.
Let’s say you decide to conduct a side-by-side comparison with treated and non-treated crops. You have a few options.
The simplest is treating half the field and leaving the other half untreated. But many growers are wary of having an untreated section of any significant size, since that can impact yield and income.
So, most growers prefer to treat half the field with their usual controls (you’ll hear researchers call this the “growers’ standard”) and the other half with the test product or products.
“That is another absolutely equally valid comparison,” Keinath says.
Some operations set up a product test on multiple large fields. But growers come in all sizes, and smaller farms are much more likely to have an array of vegetable crops. That doesn’t leave much space for testing a product on a single crop.
So how small can you go for each treatment and still get good results?
“That depends on the size of the plant,” Keinath says. “Warm season crops — cucumbers, watermelons, tomatoes — five plants would be the absolute minimum.”
For direct seeded, more densely planted crops, then three feet is the smallest you should go, he says. But he prefers you try closer to five to 10 feet for a small test.
A small plot has the advantage that you can do a comparison of a completely un-treated control plot without harming your yield.
“You're probably not going to be collecting observations over the entire treated area, anyway. So as long as you're collecting ratings/counts from the same sized areas, it doesn't really matter how big or small the actual treated or non-treated areas are,” Keinath says.
These small plots are not without risk, however. You might not have enough non-treated area for the diseases to develop.
“Soil-borne pathogens like Fusarium and Phytophthora are notoriously patchy,” Keinath says.
If your test plot happens to be in an area where the disease doesn’t develop, it can skew the results. But growers can work around that, he says.
If you make the effort to set up the trial in a disease hotspot, it’ll give your test products a true workout and give you data to collect.
2. Replicate Tests if You Can
“You will get better data and be able to make a more firm conclusion if you have replicated plots,” Keinath says.
Take the simplest test Keinath mentioned earlier — a field divided in half, with the control crops on one side and test product on the other. It’s a great deal more effective if you quarter the field instead, he says.
In fact, the more replications, the better.
“You will get better data and be able to make a more firm conclusion if you have replicated plots.”
One thing to keep in mind is that the test plots need to balance with the control plots.
“If you’re testing one new biopesticide, you should divide your field into eight sections. Four get the biopesticide, and four get the control treatment,” Keither says.
3. Randomize
To simplify things, it might be tempting to stick with dividing a field in half, with one side considered four plots, just adjacent to one another. In fact, Keinath goes one further and discourages you from alternating plots one after another.
That’s a pattern and can overlook uneven conditions in a field. Take soil types. If you have the test plots one after another, you may all be on a sand hill, or all of the controls are in silty soil with the tests in rocky areas.
“Randomization is to help minimize effects of outside factors on your comparison,” Keinath says. “You either want to exclude them or make sure that they’re all the same so that the differences between treated and control … do not have any differences in weather, rainfall, soil, moisture, holding capacity, pH, etc., affecting the results.”
With that in mind, place paired plots (with one control and one test adjacent to each other) in different parts of the field — one on a sand hill and another pair in muck soil, perhaps. That helps make the results from each set of pairs an apples-to-apples comparison.
If you are testing more than one product at a time, the same principle applies.
4. Include Repetition
Repetition is an easy concept to grasp. You simply conduct the same study at a different time or in a different place from the original study.
“Some of the larger corporate farms may be able to run their bio pesticide trials on literally on different farms. Even though it's done at the same time, it would still count as a repetition,” Keinath says.
For growers, repeating a study at a different time and growing season is better, Keinath says.
“Spring and fall are what I can typically do here in Charleston. Or different years,” he says.
While it may seem like wasted effort and time, it matters.
“Nothing is going to work exactly the same every time you apply it out in the field,” Keinath says.
