Carrots Haven't Lost Their Nutrients — What Changed Is the Zinc and Iron in Wheat
The popular claim that depleted soil is making vegetables less nutritious has no data behind it; what is actually documented is the dilution effect in high-yield dwarf wheat, where zinc has fallen by about half.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
The popular claim that a listener's letter brings up
The episode starts with a listener, Sarah, who read a line from Dr Karen Rajan in a BBC news piece: we know soil is not as nutrient-dense as it was decades ago, so a carrot in the 1950s was far more nutritious than a carrot in 2026. Host Tim Harford breaks that sentence into two testable propositions: whether fruits and vegetables really have lost nutrients, and whether soil really has been depleted. The answers to those two propositions are not the same, and the episode handles them separately.
— Tim HarfordNutrition databases were never built to compare across time
Robin Miles, a former senior scientist at the Canadian Food and Nutrition Bureau, points out that the conclusions about large declines in calcium and iron mostly come from nutrient composition databases. Those databases are designed for registered dietitians building menus, the sample sizes are usually very small, and the analytical methods have kept changing over the years. Using them for a scientific comparison across decades is not appropriate. Robin says the statistically more rigorous studies conclude that levels of the various nutrients in crops have largely not declined.
— Robin MilesThe real decline in a 180-year wheat experiment
The Broadbalk experiment at Rothamsted Research has grown wheat since 1843 and has kept samples of soil, crops and even fertiliser; it is the oldest experiment in the world. Steve McGrath says that for a long time nutrient concentrations did not change, until around the 1960s or 70s, when some nutrients began to fall sharply: zinc down 50%, iron, magnesium and copper down 20% to 30%. This is the hardest set of numbers in the whole piece, and the only part of the nutrient decline claim that holds up.
— Steve McGrathThe cause of the decline is yield, not soil
McGrath attributes the decline to the higher grain share brought by dwarf varieties: the dwarf wheat that Norman Borlaug promoted in the Green Revolution greatly raised the ratio of grain to straw, and as the harvest index rose, the concentrations of zinc, iron, magnesium and copper in the grain fell. In his words, more carbohydrate was packed in for higher yield, the nutrients did not keep up, and the concentration was diluted. That is the dilution effect. The mechanism is not fully understood and may also involve how nutrients are taken up by the grain from the short stem or the roots, but he is confident it was not caused by changes in soil nutrients.
— Steve McGrathSoil nutrients have actually risen
McGrath's team analysed soil nutrients on that experimental plot over 180 years and found concentrations had not fallen but had risen somewhat. He gives two reasons why soil nutrients could not have been lost on a large scale: first, there is almost no data on how soil changes over time, so the claim that nutrients all vanished from the soil has nothing to stand on; second, most of the micronutrients he studies cling tightly to soil particles and are not soluble, so they do not wash away with water, and he sees no mechanism for a large decline.
— Steve McGrathSoil degradation is serious, but that is not nutrients being washed away
The episode deliberately separates two things: soil degradation is indeed a serious global problem, including carbon loss, salinisation, acidification, contamination by certain elements, and wind and water erosion; but that is a different matter from nutrients being washed away and making vegetables less nutritious. McGrath says there is almost no data supporting the latter claim. This distinction is the key to the whole episode — acknowledging that soil has problems is not the same as accepting that popular causal chain.
— Steve McGrathRice and oats very likely have the same problem, but nobody knows how much
Back to the original question: the claim that carrots are less nutritious than in 1950 lacks good evidence, and so does the claim that soil nutrient density has fallen. But some important nutrients in wheat really have declined, and McGrath thinks the same process very likely happens in other staple crops, such as rice, barley and oats, because they too have used methods that raise yield per plant. Which crops exactly, and by how much, is not currently known. The advice at the end of the episode is plain: eat as varied and balanced a diet as you can.
— Steve McGrathIn their own words · checked verbatim
When you look at the sources of data for these claims that were made, they are often looking at the nutrient composition databases.
Robin Miles1:02
The valid scientific studies have shown that the levels of various nutrients in crops, for the most part, have not declined.
Robin Miles2:02
They stayed the same until around about the 1960s or 70s, when some of those nutrients started to decrease in concentrations drastically, and actually very large decreases in some nutrients like zinc of 50%.
Steve McGrath3:05
You could say that we're pushing more carbohydrates in for more yield, but the nutrients are not keeping up, so the concentration goes down.
Steve McGrath5:15
We analyse these nutrients in soil, and actually, if anything, their concentrations have increased over the 180-year time span.
Steve McGrath5:15
A lot of the micronutrients, in particular, that I'm interested in, they tend to stick to the soil very strongly. They're on the soil particles, if you like, most of it. And they don't wash out. It's not a soluble thing.
Steve McGrath6:18
They've also utilised this way of getting more yield per unit of plant, if you like. And so it's quite likely that the similar sort of thing happens with them.
Steve McGrath7:21
Figures
| Start year of the Broadbalk experiment | 1843 | 3:05 |
| Decline in zinc content of wheat | 50% | 3:05 |
| Decline in iron, magnesium and copper content of wheat | 20% to 30% | 4:12 |
| Duration of the Broadbalk experiment | 180 years | 3:05 |
Glossary
- dilution effect
- High-yield crops pack more carbohydrate into the grain, so mineral concentrations per unit weight are thinned out.
- harvest index
- The share of grain in the whole plant's biomass; dwarf varieties raised this share substantially.
- Broadbalk experiment
- A long-term trial at Rothamsted in the UK growing wheat continuously since 1843, keeping soil and crop samples.
How to listen
Writers and content people working on food, agriculture or health topics, and anyone who wants to understand the evidentiary basis for claims like food is getting less and less nutritious.
The listener letter at the start and the dietary advice at the end can be fast-forwarded; the core is from 3:05 to 7:21.