Why certified organic bread still tests positive for glyphosate

In February 2026 the Florida Department of Health tested eight loaves of supermarket bread for glyphosate and found it in six of them. Most of the coverage stopped at that sentence.

The more interesting detail is further down the table. The two lowest results in the entire test were the two certified organic loaves. They were also not zero.

That gap — between "lowest in the test" and "zero" — is the whole story, because it is the difference between what the USDA Organic seal actually promises and what almost everyone assumes it promises.

What Florida published

The results come from the state's Healthy Florida First initiative, announced on February 6, 2026 and posted in full on the state's testing results page.

Both Dave's loaves are built on organic flour. White Bread Done Right opens its ingredient list with eight organic flours, the brand's own food safety FAQ describes its products as USDA Organic certified and Non-GMO Project Verified, and Flowers Foods, which owns the brand, calls it the number one organic bread brand in the country.

Two things jump out.

Organic cut the number by roughly fifteen times. That is a real, large effect, and it is the single biggest variable in the table.

Organic did not take it to zero. Ten to twelve parts per billion is small, but it is a detection, on a loaf whose entire value proposition is that no synthetic herbicide was used to grow it.

One footnote worth noticing: five of the eight loaves tested belong to the same company. Flowers Foods owns Nature's Own, Wonder and Dave's Killer Bread — the top of that table and the bottom of it. We wrote about that pattern in more detail in six "health" brands that got caught, and who actually owns them.

Organic bans the spray. It does not ban the residue.

This is where the label gets misread, and it is not really the shopper's fault.

Under 7 CFR 205.105, a product cannot be sold as organic if it was produced with synthetic substances, except for the narrow list of exceptions the USDA maintains. Glyphosate is synthetic and is not on that list, so it cannot be applied to an organic crop. Under 7 CFR 205.202, the land itself must have had no prohibited substance applied for three years before harvest, and the farm must maintain "distinct, defined boundaries and buffer zones" to prevent contact with a prohibited substance applied to adjoining land.

Every one of those rules governs what the farmer does.

The rule that governs what a lab finds is a different one. 7 CFR 205.671 says a product must not be sold or labeled as organic when residue testing detects a prohibited substance "at levels that are greater than 5 percent of the Environmental Protection Agency's tolerance for the specific residue detected."

So the organic standard has a number in it. Here is what that number works out to for bread.

EPA's glyphosate tolerances live in 40 CFR 180.364. There is no wheat-specific line; wheat falls under "Grain, cereal, group 15," which carries a tolerance of 30 ppm — 30,000 ppb. Five percent of that is 1,500 ppb.

Dave's results are roughly 130 times under the organic threshold and roughly 2,500 times under the federal tolerance. The company says exactly this in its own food safety FAQ: it "absolutely does not add or use glyphosate," it attributes trace results to drift from nearby farms and shared infrastructure, and it points out that USDA allows organic products up to 5 percent of the EPA tolerance.

That is an accurate description of the rules. It is also, read carefully, an admission that the organic seal is a process standard. It certifies what may be applied. It does not certify what may be detected — except at a ceiling of 1,500 ppb, which is roughly eight times higher than the worst loaf in Florida's entire test and about 130 times higher than either organic one.

Where the residue comes from when nobody sprayed it

There are two different stories in that table, and they should not be blurred together.

The conventional loaves. Glyphosate has been used since the 1980s as a pre-harvest desiccant — sprayed on a standing crop to kill and dry it evenly so the combine can come through sooner. Ensia's reporting puts the typical application at three to seven days before harvest, concentrated in the northern Great Plains and the Canadian prairies, where wet autumns make uniform drying hard. Because the spray goes on days before the grain is collected, the residue does not have time to dissipate. The UK's Soil Association — an organic-sector body, so read it with that in mind — reports pre-harvest use accounting for around 30 percent of glyphosate application on UK wheat and barley, and glyphosate turning up in 41 percent of tested UK-grown cereal samples. That mechanism explains results in the 130–191 ppb range cleanly.

The organic loaves. Desiccation is not available as an explanation, because it is prohibited. What is left is drift from neighboring conventional fields, shared water, and shared handling — trucks, elevators, mills and bakery lines that also run conventional grain. Dave's names drift and shared infrastructure specifically. Florida did not trace the route for these particular loaves, and neither can we; nobody published a contamination investigation. What we can say is that the federal rules were written in anticipation of exactly this, which is why 205.671 has a threshold in it at all rather than a flat prohibition on detection.

USDA has run this test on its own seal

This is not a hypothetical. In 2010 and 2011 the National Organic Program pulled 571 domestic and imported fruit and vegetable samples carrying the USDA organic seal off retail shelves and screened each one for roughly 200 pesticides, publishing the results in a 2012 pilot study.

• 57 percent had no detectable residues at all.
• 39 percent had residues below 5 percent of the EPA tolerance — allowed, still organic.
• 4 percent (21 samples) came in above 5 percent of tolerance, in violation of the organic regulations.

USDA's own read on the violations was split: some looked like conventional product that had been mislabeled as organic, and others looked like genuine organic product that "hadn't been adequately protected from prohibited substances."

The caveats are real and USDA states them plainly: it was produce, not bread, across only six commodities, published in 2012, and the agency warns that because certifying agents deliberately target the samples most likely to contain residues, you cannot extrapolate a national rate from this. Take it as a shape, not a statistic: a majority clean, a substantial minority carrying legal trace amounts, a small slice out of compliance.

Does buying organic actually lower your exposure?

Yes, and it has been measured directly rather than assumed.

Fagan et al., Environmental Research, 2020 (PubMed 32797996) followed four racially and geographically diverse US families — 16 people, 158 urine samples — for five days on a completely non-organic diet and five days on a completely organic one. Mean urinary glyphosate fell 70.93 percent and its metabolite AMPA fell 76.71 percent, reaching baseline within three days. It is a small study, and it is a pilot, but the effect size is not subtle and the design is a within-person crossover, which is the right way to ask this question.

Barański et al., British Journal of Nutrition, 2014 (PMC4141693), a meta-analysis of 343 peer-reviewed publications, found detectable pesticide residues in 46 percent of conventional crop samples versus 11 percent of organic ones — roughly a four-fold difference in frequency.

Put those next to Florida's table and the picture is consistent. Organic is a large and reproducible reduction in pesticide exposure. It is not a guarantee of zero, and it was never written to be one.

The part where we argue against our own headline

We would rather say this out loud than let a scary number do work it cannot support.

These are small amounts. Poynter's fact-check worked out that a 150-pound adult would have to eat on the order of 18,850 slices a day to reach the federal limit. Toxicologist Norbert Kaminski told them that "based on the weight of evidence, these are not particularly high or dangerous levels of glyphosate." No recall, enforcement action or withdrawal followed the Florida results.

Florida's report has real gaps. Food Safety Magazine noted that the release "lacks any details about the sampling, testing, and analysis methodology" — no lab named, no method, no limits of detection, which also means "non-detect" is an unquantified claim. The New Lede reported that Flowers Foods did not respond to a request for comment. State Surgeon General Joseph Ladapo's framing — chronic exposure linked to gut microbiome changes, liver inflammation and neurologic effects — was not accompanied by the exposure levels those findings came from.

The hazard science is genuinely contested. IARC classified glyphosate as Group 2A, "probably carcinogenic to humans," in March 2015, on limited human evidence, sufficient animal evidence and strong genotoxicity evidence. EPA's assessment reached the opposite conclusion, that glyphosate is not likely to be carcinogenic to humans at label-directed exposures. Both of those are real conclusions by real scientists reading overlapping evidence with different questions in mind — hazard versus risk at realistic doses.

So no, we do not think 11 ppb in a slice of organic toast is an emergency. We think two separate things are true at once: the dose here is small, and the fact that a seal marketed as "no synthetic pesticides" legally tolerates up to 1,500 ppb in grain is a labeling fact shoppers are entitled to know before they pay double for it.

How to shop this

1. Organic is still the biggest lever you have. Fifteen times lower in Florida's table, about 71 percent lower in urine after five days. Treat it as a large reduction, not a shield.
2. The variable is the wheat, not the bakery. The spread in that table tracks how the grain was farmed and dried, not who baked it. Conventional wheat from a desiccation region is the exposure; the loaf is just the delivery.
3. Do not read "Non-GMO Project Verified" as "no glyphosate." The Non-GMO Project's own product verification page describes a standard for genetically modified inputs, documentation and ingredient analysis. It does not describe pesticide-residue testing.
4. Ask brands for the actual number. A bakery that publishes a certificate of analysis with a named lab, a method and a limit of detection is telling you something no adjective on the bag can. Florida's report is a good example of how much a result loses when those three things are missing.
5. Keep it in proportion. If you eat a lot of conventional wheat, switching the bread is a cheap, high-leverage change. It is not a reason to panic about a sandwich you already ate.

What we would actually want

Not a ban. A number.

Every party in this story already knows something the shopper does not. The mill knows whether the wheat came from a desiccation region. The certifier knows what the residue tests said. The USDA already picked a threshold — 5 percent of tolerance — and simply never put it on the package. If the threshold is defensible, and we think it mostly is, then printing it costs nothing and ends the confusion permanently.

Until then, "organic" means the farmer followed the rules. It does not mean the lab found nothing, and the only way to know the difference is to go looking for numbers that nobody is required to show you.

Tallow reads the label, tracks published lab results and scores what is actually in the product, so you are not holding four federal regulations in your head in the bread aisle. Scan your bread in Tallow.

Figures in this post come from the Florida Department of Health's published Healthy Florida First results, federal regulations as currently codified, and the peer-reviewed studies linked above. Florida did not publish its laboratory, methodology or limits of detection, and we have flagged that where it matters. No product described here was recalled or found in violation of a federal limit. This is not medical advice.