Same Vegetable, Same Pesticide. Seven Years Later, a Different Answer.

Somewhere between November 2015 and August 2016, a sample of New Zealand silverbeet went to a lab as part of MPI's routine pesticide survey.

It came back with 14 mg/kg of methamidophos, an organophosphate insecticide. The legal limit for silverbeet was 0.5.

MPI's report on that survey, published in June 2017, didn't dress it up. It called the sample "a feasible food safety risk". A one-off serve, for people who eat a lot of silverbeet, would have gone over the acute reference dose, the amount considered safe to eat in a single day. MPI added that this was still below levels reported to cause toxic effects.

Then it acted. MPI traced the sample to the grower, visited the farm, and retested silverbeet, cabbages and Brussels sprouts from it. None had methamidophos. The report also said that, from then on, finding methamidophos on a crop it wasn't approved for would be an offence.

That's what a food safety system is supposed to look like.

Seven years later

In the 2022-23 survey, it happened again. Same vegetable, same chemical.

This time the silverbeet had 3.6 mg/kg of methamidophos. The legal limit had dropped to 0.01 mg/kg by then, so the sample was 360 times over. The same survey found it on a celery sample and a choy sum sample. Methamidophos wasn't authorised for any of the three.

To be fair about the numbers: the 2016 sample was nearly four times more concentrated. The 360 times figure is as big as it is partly because the limit came down in July 2017.

MPI published its report on the 2022-23 survey in September 2024. It said there were "no food safety risks" from any of the 33 results over the limit.

MPI's own risk assessment said something else. Greenpeace obtained it under the Official Information Act, and RNZ reported what it contains. The residue came to 106% of the acute reference dose for adults and 232% for children. The assessment concluded that "a human health food safety risk for children cannot be excluded". It worked out that 64 grams of that silverbeet would be dangerous for a 23 kg child.

By the time MPI contacted the grower, the silverbeet had already been harvested and sold. There was no recall.

So in 2016, a sample over the safe daily limit was called a risk in public. In 2024, a sample over the safe daily limit was called a risk in the file and no risk in public.

MPI's explanation

New Zealand Food Safety's deputy director-general, Vincent Arbuckle, told RNZ the exposure would have been a one-off. He said washing silverbeet can cut residues by about a quarter, and that methamidophos breaks down above 80°C. On that basis, NZFS decided there was no immediate or ongoing risk.

Take a quarter off 232% and a child is still at about 174%.

The cooking argument also assumes it was cooked.

And the washing argument raises its own question, which is the next post.

What happened to the growers

In 2017, MPI said a detection like this would be an offence.

In 2022-23 there were three of them. According to the report, each grower was told the use wasn't allowed, had the label rules explained, and was given "further education".

The same report flags six more over-limit results from growers who'd been caught in earlier surveys. MPI's response to those was an on-site visit, a look in the chemical shed, and education.

In 2025, RNZ reported that NZFS hadn't issued a single fine, prosecution or recall for illegal residues in at least five years.

From 1 July 2024, methamidophos could no longer be imported or manufactured in New Zealand. The report saying the 2022-23 results posed no risk came out two months after that.

0.02% or 9%?

MPI leads with one number. Of 147,063 individual test results in 2022-23, 33 were over the limit. That's 0.02%.

The same report has another number in it. Those 33 results came from 27 samples, out of 300. That's 9%.

Both are true. Only one of them is about the food you'd actually buy. For New Zealand-grown samples alone, it was 22 out of 226, almost one in ten.

The latest survey, for 2024-25, is better: six domestic samples out of 251 went over, with ten results between them. They were lettuce, spinach, watercress and beans. That report also says there were no food safety risks.

It's also worth knowing what kind of survey this is. MPI's programme isn't built to catch food before you eat it. It tells you, a year or two later, what you ate.

Where organic fits

This is the whole reason a lot of people buy organic. Not because organic food is guaranteed residue-free. It isn't. One of the 2024-25 watercress samples got its residue from spray drift, and drift doesn't check certificates.

The difference is the starting point. MPI's own explainer says organic production doesn't use most human-made pesticides or most synthetic chemicals. An organic grower can't reach for an organophosphate for a pest on the silverbeet, on-label or off it.

And what you eat shows up quickly. When researchers switched families onto organic food, the pesticide markers in their urine dropped within days.

If you've been wondering why trust in the food system keeps falling, here's one small answer. Two reports, seven years apart, about the same vegetable and the same chemical, both over the safe limit.

The first one told you.

Before you go

If this was useful, you can sign up free to get new posts by email, or leave a one-off tip to help keep the research going.