5 April 1996

A good sample needed to give a good analysis

Meaningful results from analyses of farm samples, be they of soil or forage, can only come through team effort. Andrew Blake reports from a leading laboratory

EVEN the most sophisticated equipment is helpless without a truly representative sample. That is the strong message from Mike Eustace, director of Berks-based Natural Resource Management.

Growers faith in scientists to analyse their offerings correctly took a knock recently when two companies apparently gave widely different nitrogen top-dressing advice for the same crop on identical soils (Arable, Mar 8).

But analysts can only work with what they are given, warns Mr Eustace. And pressure on laboratories to interpret results and give recommendations risks confusion unless each side knows exactly what is required. "What do we do with samples that arrive labelled simply analyse as before plus S – three to come?" he pleads.

NRM was set up in 1993 after ICIs Jealotts Hill Laboratory ceased such work. From an initial start of about 26,000 soil samples a year the firm quickly expanded to its current throughput of about 800 a day, says Mr Eustace. "We also do about two forage analyses for every three soil tests. I believe the biggest error lies in the sampling procedure."

Soils manager Steve Holmes echoes that view. "There can be big problems with sub-sampling. You cant expect a thimbleful of soil to provide a sensible picture of the nutrient status of a 50-acre field." NRM customers receive a booklet showing how and when to take samples and highlighting factors causing irregularities.

The firm bases its advice for phosphate, potash and magnesium on MAFF handbook RB209. It also publishes advisory notes on the main trace elements but not, so far, for nitrogen.

Operations director Christine Collins believes much of the recent criticism is unwarranted. Standard procedures under the Department of Healths Good Laboratory Practice scheme are regularly monitored by independent scientists.

"Our farmers records are treated better than those of human beings in a hospital," she claims. Every sample arriving at NRM is immediately given a unique identification number relaying its origin and indicating the type of report required. These range from brief summaries for farms and advisors to 20-page documents for some researchers.

"One of our problems is that we dont get much feedback from the results of growers actions," says Mrs Collins. This makes it harder to refine advice in the relatively novel but complex area of soil mineral-N testing, she explains.

Ring testing, an inter-laboratory procedure to ensure different firms arrive at the same results, is currently restricted to a sampling exchange scheme with ADAS, says Mr Eustace. But within the lab certified reference standards, for example of all main nutrients and trace elements, provide constant checks on equipment to avoid drift in the results.

"I am a great believer in ring tests," says Mr Eustace. But they are impossible when dealing with the ephemeral nature of available soil nitrogen. Results can vary depending on whether samples are frozen or merely kept cool. Unfreezing can allow further mineralisation, he explains. "ADAS freezes its samples, so its figures are probably higher than ours."

Ideally, for a truly accurate indication of soil N availability growers need both deep sampling (0-90cm) for the background picture as well as Kemira-style shallower (0-30cm) findings incorporating an incubation assessment, he suggests.

"We are only really in our full second year of soil mineral-N testing," says Mrs Collins. "It is a science that is bound to get better over the years. We are working at the frontiers of ignorance."

To get the best from current knowledge growers must provide representative samples, she says. But beyond that they must play the game by providing enough background information, say on previous cropping, variety, and yield and market expectations to permit appropriate advice to be given.

"Giving two reports for a single sample can be dangerous," comments Mrs Collins. It is important that growers define precisely what they require from the analysis at the outset, she says. &#42

&#8226 Team effort needed for practical benefit.

&#8226 Analysis only as good as sample allows.

&#8226 Soil mineral-N testing a complex area.

&#8226 Nitrogen recommendations merit caution.


Rock count…Soil analysis, and the resulting recommendations, can be useless unless growers take care with sampling and spell out precisely why they need the data, says NRMs Christine Collins.