Why A Laboratory Can Report "Not Detected" When It's There

Nearly every multi-residue pesticide panel in the world rests on an extraction called QuEChERS — Quick, Easy, Cheap, Effective, Rugged, Safe. Sample is extracted into acetonitrile, salts are added to force the water and solvent apart, target compounds partition into the organic layer, and a dispersive cleanup strips co-extracted matrix before analysis.

It is an excellent method, and it is excellent because most pesticides are moderately fat-loving. They prefer the organic layer.

The trouble starts with compounds that don't.

Daminozide is a small, water-loving molecule with a partition coefficient of minus 1.53 and a free carboxylic acid group. At the salting-out step it stays resolutely in the aqueous phase.

The cleanup compounds the problem. The standard sorbent, primary–secondary amine, is a weak anion exchanger — and daminozide is anionic in the extract. What survives the first step risks being retained in the second.

This is not a matter of opinion. The European Union Reference Laboratory formally classifies daminozide as non-QuEChERS-amenable and publishes an entirely separate reference method for it: extraction in acidified methanol, a specialist column — either mixed-mode or hydrophilic interaction chromatography — and an isotopically labelled internal standard that the laboratory describes as non-negotiable. The EU's front-line screening method uses an amide column, not the reversed-phase chemistry a standard panel runs.

Chlormequat presents the same practical problem via different chemistry. It is a permanently charged quaternary ammonium cation. It has no meaningful vapour pressure, it decomposes rather than evaporating, and it is not recoverable by QuEChERS either. It too requires a dedicated method for polar compounds.

The practical consequence is the same for both: a laboratory can run a scrupulous, accredited, three-hundred-compound panel, report "not detected", and be entirely correct about what its method found — while saying nothing at all about whether the compound was present.

When the correct method *is* used, these remain the hardest analytes in the set. A 2021 study of a sixty-six compound panel reported 74 per cent recovery for daminozide — the lowest of every compound in its highest-priority category.

The historical position was far worse. The international residue evaluation from 1989 records that methods based on chemically converting daminozide for measurement showed "recoveries of only about 25 per cent… likely to be a feature of all methods based on this general approach". An entire generation of monitoring data rests on a technique that recovered a quarter of what was there.

The compounds turn up.

A 2025 study from Canada's National Research Council validated a ninety-six pesticide method and applied it to samples seized from unlicensed retailers. It found daminozide at 4,915 parts per billion in one sample — roughly fifty times the maximum level the method was required to detect — and growth regulators in half the illicit samples tested.

A separate 2023 Canadian study using a three-hundred-and-twenty-seven compound panel found the contrast starkly: six per cent of regulated samples carried any residue, and none carried a growth regulator, against ninety-two per cent positivity in unregulated samples.

And a peer-reviewed 2025 survey of 11,479 growers across eighteen countries documented something more troubling for anyone acting in good faith: daminozide and paclobutrazol have been found undeclared in commercial hydroponic fertiliser products. A grower following the label may have no idea what they are applying.

Put together: these compounds are demonstrably present in unregulated supply chains; they are invisible to the default analytical method; they are absent from most mandatory testing panels; and they have been found hidden in products that do not name them.

"Not detected" is a statement about a method. The useful questions are which method, which compounds were on the panel, and what the limit of quantification was. A certificate that doesn't say is not telling you what most people think it is telling you.