8 March 1996

Labs defend technique – but blame sampling

CRITICISMS of soil mineral testing have drawn a swift response from organisations offering such services.

All reckon their laboratory procedures are accurate. Instead they blame poor sampling and interpretation of results. Mark Gillingham, consultant to Usborne Fertilisers Soil Services admits "there is a much bigger spread of recommendations" than usual.

"For the best results you have to test at the full 1m depth, and take 10-12 samples for each measurement. Shallower tests, or other short cuts, like letting the farmer do his own sampling, can lead to some weird and wonderful results."

Samples must also be chilled correctly to stop nitrogen volatilising, he adds.

Dr Ian Richards of Levington Agriculture, which does work for Hydro, says sophisticated interpretation is vital. "Its never possible to give a perfect recommendation for a field." The long-term aim is to strive for a system that within, say, 80% of cases gets within 10-15% of the optimum, he says.

Company 1 may have advised higher rates because it was chasing all-out yield, while the latter was "playing safe" for a malting sample, Dr Richards notes.

Paul Sweeney of Kemira maintains his companys 30cm samples are adequate, since two tests are made during the season. Drought last year could have caused the variable results, he thinks.

For ADAS, Peter Dampney says soil N tests are more accurate than "crude" conventional methods based on indices. Soil N tests are accurate to within 15kg/ha (12 units/acre), but need realistic yield assessments and full account of weather after sampling if they are to result in accurate rates. &#42