快猫短视频

I LOVE playing detective. So when I found myself on a train scrolling through hundreds of shocked responses to yet another viral tweet claiming that the proportion of nutrients in our fruit and veg has collapsed over recent decades, I thought I better do some digging.

This ubiquitous, yet generally poorly evidenced, claim is one that I have explored in a previous column. However, as the new tweet cited a 2018 study that is more recent than many I have looked into before, it seems important to return to the issue, particularly given its new-found influence on social media.

After all, one of the hallmarks of good science is the flexibility to change your stance as new evidence surfaces. So, with a 4-hour journey ahead of me, I started delving into the stats used to underpin this claim.

Centred on a steep line graph extracted from a paper () published in the journal Nutrients, this tweet suggested there has been a catastrophic collapse in the average levels of , magnesium and iron in cabbage, lettuce, tomatoes and spinach of up to 90 per cent in the past hundred years.

With a link to the paper in an academic journal, it outwardly seemed very convincing. So why did it pique my interest?

Well, as a botanist fascinated by food crops and how they have shaped human history, understanding how the handful of plants on which civilisations are built have changed over the centuries is one of my key research areas. The only problem is that it is very hard to study.

One of the biggest problems we have when it comes to answering what can seem like simple questions about things like nutrient density in crops is finding like-for-like comparisons to establish how these change over time.

There are so many factors that have a measurable impact on crop chemistry, including weather, soil make-up, harvest stage and the unique genetics of a crop, that agreeing on a universal value for a 鈥渢ypical鈥� tomato today is a challenge, let alone doing the same for one more than a century ago. It is rather like picking a random New Yorker and expecting them to be representative of not just every other inhabitant of the city, regardless of race, gender, age, class, education or income, but all of humanity.

鈥淓ven graphs from journals may not tell the full story. The data may not come from scientific research at all鈥�

When you start to compare these randomly selected single values over time, created by an enormous range of environmental variables, things get even more complicated.

But let us just imagine the world isn鈥檛 as complex as we know it is, and think of this methodology as a legitimate one. In any case, science inevitably relies on a certain amount of generalisation.

Taking a closer look at the individual data points on the graph revealed that three out of seven were marked with asterisks (1914, 1941 and 1992), which according to the paper indicated that these stats couldn鈥檛 be independently verified.

Now, if it were the case that these asterisked stats were in some way unreliable and should be removed, the graph would actually show a stable level from 2000 to 2018, with just a single outlying data point, showing an atypically high level for 1948, with no consistent trend for a decline. How curious.

So, do these asterisked data points stand up to scrutiny? Well, the extremely high levels of minerals seen in the 1914 stat, which is essential to the 鈥�90 per cent decline鈥� claim, doesn鈥檛 appear to be from a peer-reviewed paper. In fact, it seems to come from a book entitled Nature Cure by Henry Lindlahr, which made for an eye-opening read when I eventually tracked down a dusty copy.

An early 20th-century physician, Lindlahr believed that sunbathing could cure cancer, vaccines caused smallpox and only 鈥渃ivilised races鈥� could succumb to ill health. Even at the time, he was widely criticised as a quack for his outlandish views, so his book may hardly be a good source for a scientific paper more than a century later.

I also found the affiliations of the three authors of the paper interesting. Two of them were listed as working for a corporation that sells nutritional supplements, and under the conflicts of interests section in the paper, the third author was described as being a paid consultant for the same company and sitting on its scientific advisory board.

I suppose the moral of the story is that even graphs from published journals may not tell the full story of the facts or any possible vested interests until you start digging. And sometimes the data behind the claim doesn鈥檛 come from scientific research at all. But who has 4-hour train journeys to do the detective work to check out each viral 鈥渟cience鈥� tweet?