Via Daniel Lemire. It turns out that spinach is not, in fact, unusually rich in iron (and what it has is not bioavailable), but the story of how this truth spread is rich in irony, and illustrates several pitfalls in scientific literature.

I remember when I was blogging regularly on health science following down a reference in an article which made a particular claim, and discovering that the original source (which was vaguely cited) supported exactly the opposite conclusion. That’s an extreme, but this article suggests several ways in which it could have come about.