If farmers are confused about which soil sampling procedure is better - by soil type or by grids (such as 2.5 acres or less) - it's understandable. Soil scientists and crop consultants are divided on the issue.

Crop consultants Phil Cochran and Dave Mowers are good examples. Cochran, of Cochran Agronomics, Paris, IL, and Mowers, of Mowers Soil Testing Plus, Toulon, IL, are friends. They agree on many aspects of crop production, but not on soil testing.

Mowers claims grid sampling yields more information at less cost than does sampling by soil type. And when grids are used with GPS, samples can be gathered from the same spots each time.

But Cochran believes sampling by soil type is more practical and beneficial.

"Soil sampling by soil type and topography is the commonsense approach," says Cochran. "That's because we base our sampling areas on the lay of the land. We segregate a field according to its inherent productivity.

"We also look at the history of the individual field, including previous cropping, manure application, drainage areas, diseases, compacted spots and yield monitor information, if available," says Cochran. "And we factor in the management practices and resources of the farmer. We consider all these things when deciding where to pull soil samples and, later, when making fertility recommendations."

He calls this "smart sampling" or sampling by management zone. "Sampling in this manner generates the same or more information at less cost per acre than does grid sampling."

Cochran also says that, based on actual case histories, there generally is little, if any, correlation between yield and fertility level as determined by grid sampling. But yield and soil type correlate significantly.

"Grid sampling often lacks the human judgment element," Cochran maintains. "A computer determines where each sample is to be collected by simply calculating the grid size it has been told to use. There is no consideration of conditions that could dramatically affect soil chemistry."

What's more, Cochran notes, the follow-up fertility recommendations often are made by a computer, based on rigid, straight-line numerical values developed by the state land grant university.

"For example, let's say the recommended value for phosphorus is 50 lbs/acre. If the soil test is 49, phosphate is needed. If it's 51, no phosphate is needed. No human interpretation is involved."

Cochran concedes, however, that intensive soil sampling in 2.5-acre or less grids does have a place.

"That's when a farm is sampled for the first time," he says. "Grids make it easier to identify man-made problems, such as those caused by tillage and cropping practices, that soil-type sampling often miss. But that's the only time grids should be used if the objective is to gain relevant information for management decisions."

Dave Mowers, in contrast, favors sampling by 2.5-acre increments, or grids, as a standard procedure.

"First of all, it establishes a basis for variable-rate fertilizer application," says Mowers. "Many fertilizer dealers now have variable-rate applicators, so this technology often is available. But even if a farmer chooses not to use variable-rate P and K application, the basic knowledge of nutrient variability is of value for making management decisions. A primary example is for correcting pH levels. That can be done by flagging a field."

There also are other reasons that intensive grid sampling makes sense, says Mowers. "This is especially true if there are multiple soil types on a farm, if a farm is worked across soil types, if there is high variability in soil fertility, if soil testing is a long-term process done at multiyear intervals to inventory nutrients accumulated in the soil, and if a farmer wants to measure the response of soil fertility to fertilizer applications.

"One of the greatest values of soil sampling by grid is that, with the use of GPS, we can go back to the same spots each time we collect samples," Mowers points out. "That allows us to make precise comparisons."

Mowers reports that, in his experience, fertility varies significantly within a soil type 95% of the time. This is due to factors such as manure application, variable yields, past fertilizer rates and application methods (broadcast or banded), and cropping history. Grid sampling is necessary to accurately measure this variation, he says.

"One of the biggest reasons for grid sampling is to get the most value for $ spent. Cost per sample is less than sampling by soil type. Typically, testing by soil type costs $4.50/acre. One sample is taken for every 10 acres, which is $45/sample.

"With grid sampling, the typical cost is $7.50/acre. But a sample is taken every 2.5 acres. That means a cost of only $18.75/sample, and we believe strongly that every one of those samples is needed. It's four times more information at 58.5% less cost."

To achieve further economy, Mowers suggests grid sampling the poorest-yielding fields first.