Florida tested candy for arsenic. Here's what the data actually supports, and what it doesn't
Florida found arsenic in 28 of 46 candy products, but toxicologists question the testing method. Here's what's actually confirmed, brand by brand.
Emily Harper
In January 2026, Florida's Department of Health announced it had found arsenic in 28 of 46 candy products under its "Healthy Florida First" initiative. The numbers spread fast — 370 parts per billion in one gummy bear brand, "safe amount per year" figures for specific candies — and a lot of that spread with the wrong attribution. This isn't Consumer Reports or an academic lab. It's a state government program, and outside toxicologists have publicly questioned the method behind it. Here's what the data actually supports, brand by brand, and what it doesn't.
Where the numbers actually come from
The testing was run by Florida's Department of Health, announced by First Lady Casey DeSantis and Surgeon General Dr. Joseph Ladapo, not an independent research lab or consumer group. The state says it tested 46 candy products from 10 companies, analyzed by what its press materials describe only as "an independent certified laboratory" — the lab itself is never named in any public document. Product-level results, including the specific parts-per-billion figures circulating online, are published on exposingfoodtoxins.com, a site connected to the state program rather than a university or nonprofit.
A group of toxicologists has since raised a specific, technical objection: the state used EPA Method 6010D, a method built for testing soil and groundwater, not food. Florida also hasn't published its sample sizes, its limit of detection, or the risk-assessment math behind the "safe number of pieces per year" figures attached to each product. As one toxicologist, Dr. Alex LeBeau, put it: "the results are likely biased because of an inappropriate analytical technique, making interpretation difficult." That doesn't mean the arsenic readings are fabricated — it means the "safe annual amount" claims built on top of them aren't something anyone outside the state program can currently check.
The numbers that are confirmed
Three products carry specific figures that match what's published on the state's own data site:
- Black Forest Gummy Bears — 370 ppb arsenic, with the state's own math putting the "safe" amount at roughly 16 pieces per year for a child, 40 for an adult. Worth noting: the organic version of the same product tested non-detect for arsenic.
- Smart Sweets Caramels — 240 ppb arsenic, roughly one 45-gram bag per year for a child by the state's figures. Smart Sweets' Sweet Fish came in at 180 ppb; its Red Twists showed no elevated risk under the same testing — not every product from the same brand was flagged.
- Tootsie — this is where the number that's been circulating gets mixed up. The 570 ppb figure and "about 8 pieces per year" claim belongs to Tootsie Fruit Chew Lime, not the plain original Tootsie Roll. Florida's own data lists plain Tootsie Roll at 380 ppb and Tootsie Roll Vanilla at 370 ppb — real numbers, just a different product than the one usually cited.
Two claims from the viral version of this story don't hold up. "Natural flavors can legally hide up to 3,000 undisclosed compounds" isn't accurate as usually stated — that figure traces to FEMA's GRAS list, the industry's cumulative pool of roughly 2,800–3,000 flavoring substances self-certified over six decades, not a count of what's hidden inside any single product. Separately, the Center for Science in the Public Interest has flagged that of roughly 10,000 chemicals used in or around food, over 3,000 have never been substantively reviewed by the FDA — a real and separate concern, just not the same claim. And a widely repeated line about Unreal Milk Chocolate Gems containing "cadmium and lead at 2x the EPA drinking water limit" doesn't check out as stated: that testing came from an independent tester's single lab report, run more than a year apart from Florida's program and never blended with it, and it doesn't support a clean "2x" figure for either metal.
What's actually well-documented: heavy metals in chocolate generally
If you want the strongest version of this story, it isn't the Florida program — it's chocolate testing that's been repeated by multiple groups for years. Consumer Reports tested 48 chocolate products in October 2023, three samples each, and found detectable lead and cadmium in every single one; 16 of the 48 exceeded California's health-based threshold for at least one metal. There's no federal limit for heavy metals in chocolate at all. In 2018, As You Sow settled with 32 confectionery companies under California's Prop 65, setting tiered lead and cadmium thresholds by cacao percentage — evidence the industry itself has known about this for a while.
The reason isn't a bad recipe. Cadmium gets into cacao through soil uptake — the cacao tree naturally concentrates whatever cadmium is in the soil it's grown in, and Latin American growing regions tend to carry more of it. Lead is tied more to processing and drying, when beans are dried near roads or handled with contaminated equipment. Neither has anything to do with whether a candy is "clean label," organic, or made with simple ingredients — which is exactly why an all-natural or premium candy brand isn't automatically safer on this axis.
A related claim worth correcting: Annie's and "tested clean until General Mills bought them"
This one circulates alongside the candy story and deserves its own fact-check, because the real timeline runs backwards from how it's usually told. The actual event was a 2017 investigation commissioned by the Ecology Center and the Coalition for Safer Food Processing and Packaging, which tested cheese powder from 10 boxed mac-and-cheese products, including two Annie's Homegrown varieties. Phthalates showed up in all 10 — roughly four times higher in the cheese powder than in blocks of natural cheese — with no meaningful difference between organic and conventional products. General Mills acquired Annie's in September 2014, three full years before that testing happened. There's no published testing showing Annie's was "clean" before the acquisition and contaminated after; the contamination traced to shared industry processing equipment (plastic tubing, PVC gaskets), an industry-wide issue, not something specific to one brand's ownership change. To General Mills and Annie's credit, in 2021 they became the first food company to publicly commit to eliminating ortho-phthalates from their processing equipment and packaging, and a class action over the original findings was filed that same year.
What to actually take from this
Strip out the parts that don't hold up, and there's still a real, defensible story here: Florida's own testing did find arsenic in a majority of the candy products it checked, at levels worth knowing about even without a fully public methodology behind the "safe amount" claims. And the much better documented pattern — heavy metals in chocolate and cocoa specifically — shows up regardless of how "clean" the ingredient list looks, because the contamination comes from soil and processing lines, not the recipe. A candy with five recognizable ingredients isn't automatically safer on lead, cadmium or arsenic than one with fifty.
Where Tallow fits
The instinct behind "check before you buy" is right — the actual work is figuring out which numbers behind a viral claim are real, which product they apply to, and what's still missing. Tallow scans a product and shows you the lab data that actually exists for it, sourced and dated, instead of a wave of "independent lab" language standing in for a citation.
Read the ingredient list. Then go find out if anyone's actually tested it.
This article distinguishes between Florida Department of Health testing (arsenic figures) and independent single-tester or academic testing (chocolate, cheese powder) — they are separate efforts from different sources and should not be read as one unified "lab testing wave." Several claims circulating about this topic could not be independently verified and are explicitly not asserted here as fact, including the exact risk-assessment methodology behind Florida's "safe amount per year" figures and the identity of its testing laboratory. This is not medical advice. Primary sources: WUSF/Health News Florida, Food Safety Magazine — toxicologist critique, exposingfoodtoxins.com product data, Consumer Reports — lead and cadmium in dark chocolate, and Earthjustice — 2017 mac-and-cheese phthalate testing.