What the aisles hand you

The wall of wrapped produce

223,000 particles per gram, the highest median count in the fruit samples

Oliveri Conti's team ran fruit and vegetables from Italian shops through electron microscopy, looking for particles under 10 micrometres. The highest median was 223,000 particles per gram in fruit and 97,800 in vegetables, with apples highest and lettuce lowest. The authors put the particles down to uptake by the plant rather than to the film around it.

What changes it. Nothing on this shelf separates a wrapped apple from a loose one. These counts are about what the plant took up, not what it was packed in.

Sources: Oliveri Conti 2020

The bag you tear off the roll

0.46 to 250 microplastics shed per centimetre of packaging opened

Sobhani and colleagues opened bags, caps, tape and containers by hand, with scissors and with a knife, and counted what the opening itself shed. Across materials the range was 0.46 to 250 microplastics per centimetre opened. Stiffness, thickness and density drove the difference more than the choice of polymer did.

What changes it. The variable this study found was the material, not the technique. Thicker, stiffer packaging shed more per centimetre opened.

Sources: Sobhani 2020

The bottled water shelf

325 particles per litre, averaged across 259 bottles of 11 brands

Mason's team tested 259 bottles of 11 brands bought in nine countries, tagging particles with Nile Red. Ninety three percent carried some microplastic: 10.4 particles per litre above 100 micrometres, and 325 per litre once smaller particles were counted. Polypropylene was the most common polymer identified, and that is what bottle caps are made from.

What changes it. The study's own reading is that some of the contamination comes from the packaging and the bottling line rather than the source water.

Sources: Mason 2018

The same bottle, counted again

About 240,000 particles per litre once the method could see below a micrometre

Qian and colleagues imaged bottled water particle by particle with stimulated Raman scattering microscopy and counted about 240,000 per litre, roughly 90 percent of them nanoplastics. These bottles were not dirtier than the ones Mason's team tested six years earlier. The instrument could see smaller.

What changes it. When a published count jumps by three orders of magnitude, the detection limit usually moved. On this shelf that is exactly what happened.

Sources: Qian 2024 · Mason 2018

The fizzy drinks fridge

234 to 450 particles per litre from polypropylene bottles, the higher figure after four freeze-thaw cycles

Chen's team filled plastic bottles with carbonated water and varied the pressure, the temperature and the sugar. Polypropylene bottles released 234 particles per litre, more than polyethylene or PET, and four freeze-thaw cycles pushed that to 450. Shop-bought carbonated drinks measured 260 to 281 particles per litre in the same study.

What changes it. Carbonation, cold and repeated freezing each raised the count here. A bottle that has been frozen and thawed is not the same bottle it was.

Sources: Chen 2023

What nobody has measured

Whether the film around a cucumber puts anything into the cucumber. The published produce counts come from the flesh of the fruit. The comparison that would settle it, the same crop measured wrapped and unwrapped, has not been run.

The board behind the deli counter

0.014 to 2.6 particles per gram of chicken and fish bought over the counter

Habib's team bought chicken and fish from markets, food stores and supermarkets in the Middle East and counted particles in the meat itself. Chicken ran 0.03 to 1.19 particles per gram and fish 0.014 to 2.6. They traced the source to the polyethylene cutting boards the food was cut on, and meat cut from the bone carried more than fillets prepared on other surfaces.

What changes it. Rinsing before cooking lowered the count in this study without taking all of it off. What the shop cuts on is not on the label.

Sources: Habib 2022

The fish counter

1 in 4 fish sampled from Californian markets carried man-made debris

Rochman's team bought fish and shellfish at markets in California and in Makassar, Indonesia, and examined the digestive tracts. Debris turned up in 25 percent of individual fish in the USA, 28 percent in Indonesia, and 33 percent of the shellfish. Most of the American debris was textile fibre; all of the Indonesian debris was plastic.

What changes it. The debris was counted in the digestive tract. A gutted fillet loses that tract, and shellfish eaten whole does not.

Sources: Rochman 2015

The milk fridge

3 to 11 particles per litre across 23 samples of branded milk

Kutralam-Muniasamy's team tested 23 milk samples from eight brands in Mexico and found microplastic in every one, from 3 to 11 particles per litre with an average of 6.5. The polymers were polyethersulfone and polysulfone, the membrane materials used in dairy processing rather than anything in the carton.

What changes it. These are the lowest counts reported for any liquid food, and the study points at the processing line rather than the packaging.

Sources: Kutralam-Muniasamy 2020

The cup at the coffee bar

About 25,000 particles released into one 100 millilitre cup in 15 minutes

Ranjan, Joseph and Goel filled disposable paper cups with water at 85 to 90 degrees and left them 15 minutes. Fluorescence microscopy counted about 25,000 micrometre-sized particles per 100 millilitre cup, shed by the polyethylene film that lines the paper. Fluoride, chloride and sulfate came off the same lining.

What changes it. A ceramic cup at the counter, or a steel flask carried from home, has no polymer lining at brewing temperature.

Sources: Ranjan 2021

The hot food counter

1 to 30 particles per container, the high end from boiling water in polystyrene

Caponigro's team filled takeaway containers with water at room temperature, at pH 4.5, and at 100 degrees. PET and polystyrene containers released between 1 and 30 particles each, the most at boiling, sized 13 to 32 micrometres. Polypropylene containers released none the method could detect under any condition.

What changes it. Temperature moved the count further than acidity did, and which polymer the container is made from mattered most of all.

Sources: Caponigro 2024

What nobody has measured

What the rest of the counter adds. One study has traced particles back to a shop's cutting board. The slicer, the scales, the tongs and the lined paper the food is handed over in have not been counted.

The rice

3.7 mg of plastic per 100 gram serve of unwashed store-bought rice

Dessì's team measured plastics in store-bought rice in Australia by pyrolysis gas chromatography, which weighs polymer instead of counting particles. An unwashed 100 gram serve carried about 3.7 milligrams and a washed one 2.8, while instant rice held four times as much. Polyethylene was quantifiable in 95 percent of samples.

What changes it. Rinsing lowered the measured mass. Paper packaging and plastic packaging made no statistical difference in this study.

Sources: Dessì 2021

The salt

0 to 1,674 particles per kilogram across 28 sea salt brands, against 0 to 148 in rock salt

Kim's team tested 39 salt brands produced at sites across six continents. Sea salts ran 0 to 1,674 particles per kilogram, rock salt 0 to 148 and lake salt 28 to 462. The sea salt counts tracked plastic pollution in the surrounding seawater. Karami's earlier survey of 17 brands, counting only above 149 micrometres, found 1 to 10 particles per kilogram.

What changes it. The gap between those two surveys is the size cutoff, not the salt. Within sea salts, where it was harvested moved the number most.

Sources: Kim 2018 · Karami 2017

The sugar

270 items per kilogram of white sugar, against 103 in brown

Patterson's team analysed commercial white and brown sugars and found 270 items per kilogram in white and 103 in brown, mostly black fibres under a millimetre across. Characterising the particles pointed at the refining steps rather than the cane or the packaging. That puts white sugar inside the range Kim's team measured for sea salt, and above every rock salt they tested.

What changes it. Less refining meant fewer items in this study, which is the opposite of what packaging alone would predict.

Sources: Patterson 2026

The tin of sardines

16 of 20 brands of canned sardines and sprats held no plastic particles at all

Karami's team opened 20 brands of canned sardines and sprats from 13 countries and looked for particles above 149 micrometres. Sixteen brands had none. The other four had between one and three particles each, mostly polypropylene and PET.

What changes it. This is one of the few food counts that comes back near zero, and the method could not see below 149 micrometres.

Sources: Karami 2018

The tea bags

1.20 billion nanoplastics per millilitre from a polypropylene bag, against 8.18 million from nylon-6

Banaei's team brewed three commercial tea bags and counted the nanoplastics in the leachate by particle tracking analysis. Polypropylene bags gave 1.20 billion particles per millilitre, cellulose 135 million and nylon-6 8.18 million. The material the bag is made from moved the count by more than two orders of magnitude.

What changes it. A bag that looks like paper is often heat-sealed with polypropylene. Loose leaf in a steel infuser brews without a mesh of any kind.

Sources: Banaei 2024

The freezer case

97 to 377 particles per 500 millilitres from a heated co-extruded pouch

Zhu's team heated two common flexible pouches with water, acetic acid and ethanol standing in for food. A polyamide and polyethylene co-extruded pouch released 97 to 377 particles per 500 millilitres, a retort pouch 72 to 167, most of them 20 to 50 micrometres. Their scenario estimate for a diet leaning on these packs was 106 to 754 particles a week.

What changes it. Release tracked heat and package structure here. Tipping the contents into a dish first keeps the pouch out of the hot step.

Sources: Zhu 2026

What nobody has measured

The whole trolley. Every figure in this shop is one food or one container measured on its own. No published study we could find follows a week's shopping from the shelf to the plate and adds up what the counts come to together.