The Decimal Point

Most people have heard the story. Spinach was never much richer in iron than other green vegetables. A chemist long ago put a decimal point in the wrong place, the figure came out ten times too high, and the error lasted long enough for Popeye to build a career on it.

In 1995 K. Sune Larsson told it in the Journal of Internal Medicine, in an article titled "The dissemination of false data through inadequate citation". His version: "The myth from the 1930s that spinach is a rich source of iron was due to misleading information in the original publication: a malpositioned decimal point gave a 10-fold overestimate of iron content." He cited Terence Hamblin, writing in the 1981 Christmas issue of the British Medical Journal under the title "Fake!". His next sentence was: "Once a paper with misleading information has been published, it is almost impossible to stop citation."

Ole Bjørn Rekdal, who teaches in the health and social sciences faculty at Bergen University College in Norway, read Larsson and believed him. Then he went to Hamblin, and he described what he found in Social Studies of Science in 2014. Hamblin had written that "the discovery that spinach was as valuable a source of iron as red meat was made in the 1890s", and that "German chemists reinvestigating the iron content of spinach had shown in the 1930s that the original workers had put the decimal point in the wrong place." In Hamblin the 1930s is when the error was found. In Larsson's sentence it became the decade the myth began. Rekdal thinks Larsson didn't make that shift himself. A 1989 book, Follies and Fallacies in Medicine by P. Skrabanek and James McCormick, had already credited Hamblin with tracing the myth to "a mistake by the original investigators in the 1930's". Larsson's sentence makes the same combination of mistakes in strikingly similar words, while citing only Hamblin. The decade was in the source the whole time. Somewhere between 1981 and 1995 it was attached to a different event.

That was the smaller problem. Hamblin gave no reference for the decimal point at all: no chemist, no paper, nothing a reader could check. The criminologist Mike Sutton had already followed the trail in 2010 and found nothing at the end of it. He contacted Hamblin and asked where the story came from. Hamblin said he couldn't remember, but he was sure he hadn't made it up. Sutton argued that the high nineteenth-century figures more likely came from other causes, such as contamination in the analysis or confusing dried spinach with fresh. He also argued that Popeye's creator, E. C. Segar, had vitamin A in mind, not iron. (In a 1932 strip, Popeye does say that spinach is full of vitamin A.)

A reader, Bonnie Taylor-Blake, then suggested a source Hamblin might have forgotten: Arnold Bender. In his inaugural lecture at Queen Elizabeth College, London, in 1972, Bender said that "the fame of spinach may well have grown from a misplaced decimal point." He repeated the sentence in 1975. In a letter to The Spectator in 1977 it became "appears to have been based on a misplaced decimal point." In 1981 Hamblin stated it as a fact, with the 1930s German chemists attached. In 1982 a textbook for medical students that Bender co-wrote said that "the belief can be traced back to a mistake in the transcription of analytical results in 1870, when a decimal point was misplaced."

If Bender was Hamblin's source, as Rekdal thinks most likely, the suggestion hardened by steps. "May well have" became "appears to have been", then a flat statement, then a statement with a year. No step cited any evidence. Bender may have known more than he wrote down; nobody can now tell. What each step added on the page was precision, and precision reads as evidence. The dates don't agree with each other either. Bender's textbook put the error in 1870. A 1988 Reader's Digest book put it in the 1890s. Follies and Fallacies put it in the 1930s. The one fixed point in all of it is a number nobody ever sourced.

What makes this more than a story about rumours is who was telling it. Hamblin's piece was a debunking: the decimal point was offered as the correction of a popular error. Larsson used it as his example in an article about false data spreading through citation. Rekdal points out that Bender, Hamblin, Larsson, and Skrabanek and McCormick were all engaged in separating science from bad science. When Follies and Fallacies reached its third edition in 1998, with the spinach section unchanged, McCormick's preface said that rereading the previous edition had turned up nothing "which required deletion or modification." Rekdal's explanations for the story's spread are the authority of the BMJ, and copied citations that look like independent confirmations of each other. I would add one more, as my own reading: it came labelled as a correction. A correction carries an authority the original claim never had. It arrives as the checked version.

But a correction is also a new claim, and it needs a source like any other. Hamblin's had none. And the shifted 1930s would have passed most checks a careful reader runs. Is "1930s" in the cited source? Yes. Is the decimal point? Yes. The tenfold figure? Yes. The date, the decimal point and the tenfold figure can all be found on the page he cited. What had changed was which event the date belonged to (one of several errors Rekdal finds in that sentence), and you only see it by reading the sentence beside Hamblin's. Rekdal describes working that way in general: two monitors, a text on one and its sources on the other.

Hamblin's answer to Sutton, posted on Sutton's site nearly thirty years after the BMJ piece, is the best thing in the whole record: "Thanks for pointing out my mistake of 29 years ago. I never could remember where I had first seen the decimal point story – I thought it was in Reader's Digest. I am very pleased to see that you have uncovered the whole story and very willing to admit that I was wrong. Incidentally my name is Terence not Terrance."

I have a small stake in this. I check my own essays by sampling sentences and reading them against sources, and I correct the ones that fail. This week a correction note I wrote for one of my essays added two things I hadn't checked: a clause about other researchers' work, written from memory, and the city an institute is in, which I confirmed only after the note was live. Nothing about writing a correction makes it true. It only makes it look as if someone has already checked it.

Sources: Ole Bjørn Rekdal, "Academic urban legends", Social Studies of Science 44(4):638-654 (2014), which quotes Hamblin (1981), Larsson (1995), Bender (1972-1982), Sutton (2010), Reader's Digest (1988) and Skrabanek and McCormick (1989); Wikipedia, Spinach (the 1932 Popeye strip).

← Back to essays