Monday 11 December 2017

The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted

FOREWORD


The original Best Thinking site that published several of my articles on spinach is sadly gone. However those articles from Best Thinking can be found copied and pasted on this blog site. The first pages of the articles have been internet archived (sadly Internet archive does not save the full article. Blogs from best thinking on this topic are also on this site and have been fully archived.

My Mythbusting articles on Spinach

Spinach, Iron and Popeye: Ironic lessons from biochemistry and history on the importance of healthy eating, healthy scepticism and adequate citation (Sutton 2010) (here and here and also here)

The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted (Sutton, M. 2010) (here)

Spinach Iron Decimal Point Error Myth Busted (Sutton 2010) (Here)

Did Popeye Really Increase Spinach Consumption and Production by 33 percent in 1936? (here)(Original Best Thinking blog post archived in full here)

SPIN@GE USA Beware of the Bull: The United States Department of Agriculture is Spreading Bull about Spinach, Iron and Vitamin C (Sutton 2011) (Here)

Spin@ge II: Does the United States Department of Agriculture’s Publication of Spuriofacts Have its Origins in a Perverse Scientific Paper Written in 1937? (Sutton, M. June 2012) (here and archived here)

 Spinach Iron Decimal Point Error Myth Busted (Sutton 2010/2011)
(here)

How the spinach, Popeye and iron decimal point error myth was finally bust (Sutton 2010) (Here and also here)

Note: The following article is reproduced by cutting and pasting from the orignal Best Thinking site before it closed down below)

~~~
ORIGINAL ARTICLE


The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted 


First page of original article is  INTERNET ARCHIVED http://archive.is/ASqr



A story started thirty years ago by nutritionist Professor Arnold Bender, and famously supported by the immunohaematologist Professor Terence Hamblin, that a decimal point error  made in 19th Century research of the iron content of spinach led to its erroneous promotion, is completely untrue.


Mike Sutton is the author of Nullius in Verba: Darwin's greatest secret the book that famously used big data analysis to prove that Darwin and Wallace committed the world's greatest science fraud.

Postscript (Nov 5th and 6th 2019) 

Although they cite it, in my opinion Mielewczik, M. and Moll, J. are authors of an interesting and most useful paper on this topic (published in 2016) appear - in the most favorable interpretation - to have most weirdly not at all properly read my paper below, which was published online in 2010 and 2011 on the Best Thinking website and  reproduced on this blogsite along with the original BestThinking readers comments on it by academics and my replies to them. I think Mielewczik, M. and Moll, J. for some reason or other accidentally or perhaps even deliberately over-simply my conclusions - focusing their own conclusions about my work in this area on the conclusions of my first paper, whilst not attributing me with the depth of my 19th and 20th century research in this area and the more complex multi-factored findings and conclusions I reach, which are based on that later research I conducted. My original findings and conclusions, and the discussions with academics and their corrections of my errors, and my incorporation of those corrections -  attributed to the academics who made them -  are not properly cited by Mielewczik  and Moll. All are in the paper below and the comments below it. Others are in other later published papers (see above list), which they also fail to cite, that were published years before their work in this area and before they even put pen to paper on the topic.
The journal publishing Mielewczik's and Moll's Spinach in Blunderland and link in that journal to the article in question can be found here

Postscript (26 Jan 2012)

In an earlier paper (Sutton 2010a) I explain how I traced the origins of the Spinach, Popeye, Iron Decimal Error Story (SPIDES) as far back as I could to an article published in the British Medical Journal. That article was written by Professor Hamblin. I emailed Professor Hamblin and asked for the source of his story. His very prompt and courteous reply was that he had forgotten the original source of the story, after so many years, but thought he may have read it in an unknown copy of the Reader's Digest. Unable to find any earlier source for the story, in the Readers Digest or elsewhere, I concluded that Hamblin had probably conjured the story from thin air. As the paper below reveals, subsequent correspondence from the USA came in response to my 2010 paper and gave me a new lead. In light of this new information I was able to trace the story as far back as an inaugural lecture given by Professor Bender in 1972. In his inaugural lecture (at p. 11) Bender attributes the discovery of the decimal error explanation for the exaggerated iron content of spinach to the work of a professor Schupan. In a later letter published in the Spectator (Bender 1977), Bender claims that the correct iron figure for spinach was first discovered in 1937 by Professor Schupan. I can find no reference to any work published by a Professor Schupan on this subject. If any reader can find it then they will have solved another piece of this puzzle about how myths are created and spread (it is possible that Bender misspelled this professor’s name). A further clue to the origin of the decimal error story can be found in Bender's (1977 p. 11) acknowledgment and thanks to a "Professor den Hartog of Holland for tracing the possible origins of the belief." Unfortunately, Bender gave no reference to any published work or this professor's affiliation.

Without any other evidence to the contrary (since Hamblin never referenced his source), it seems fair to conclude that Bender's reference to a Professor Schupan is the source of Hamblin's mysterious German scientists - who he claimed discovered the decimal error myth in the 1930's. As the article below explains, both Bender and Hamblin were wrong about the true iron content of spinach being discovered in the 1930's. They were also wrong about there ever having been a decimal error. And they were wrong about Popeye eating spinach for iron. Because the truth about spinach and iron was known in 1892 and widely disseminated by US scientists as early as 1907. But it is fair to say that Hamblin most certainly never made the story up out of thin air as I earlier, and wrongly, suspected.

The world works in wondrous, though rationally explainable, ways. Had it not been for my original belief in the SPIDES – based on Hamblin’s BMJ article – and my efforts to track down those mysterious German scientists I would have never embarked on a self-taught course in myth busting and philosophy of science. Those newly acquired skills have enabled me to bust two further myths: first, that all beat patrol policing must be ineffective and second, both the Routine Activities Theory (RAT) and Crime Opportunity Theory explanation of opportunity as a cause of crime.

I developed the concept of Supermyths from this work (myths that are, with great unintended irony, credulously believed by scholars and used to argue for the need to be sceptical of widely accepted myths and fallacies). To date I have identified four supermyths - the Spinach Myth (outlned in this article), the Zombie Cop Myth , Crime Opportunity Myth and the Semmelweis Myth.

Supermyths and Bracedmyths Definition:

What are supermyths and braced myths?
Supermyths, of which braced myths are a sub-type - are ironic unintended, or else a deliberate and disingenuous, consequences of fallacy dissemination. Supermyths have three very specific components:
  1. the creation of a fallacy, myth or error by an orthodox expert
  2. it being used by another expert who in turn promotes it as being ‘true’ and
  3. whilst still thinking that it is true, promotes it as a good example of the need to be healthily sceptical of bad scholarship. Moreover, fourthly:
  4. Braced myths are supermyths that have been pointedly deployed by orthodox scholars in order to bust another specific myth or fallacy. The braced myth hypothesis is that using one myth as a specific mythbusting device in this way braces the supermyth to make it further entrenched and therefore more difficult to prevent it being credulously disseminated as veracious knowledge.
Immunomythology?

page 1
page 2
As far as I can tell, the Spinach Myth that Hamblin helped to spread led to no physically harmful impact upon anyone. But the two criminological myths I have busted, since investigating his spinach story, are widely believed to be true and they have had a significant international impact upon policing and crime prevention policy making with all the attendant negative consequences that must flow from ineffective measures in these areas.

In a recent blog on this subject I hypothesize that trivial myths such as the SPIDES might actually work like a vaccination against serious ones. Could Hamblin's role in spreading the SPIDES, and its subsequent myth-busting, work as some kind of beneficial social immunomythology vaccination?

Professor Hamblin sadly passed away on January 8th 2012. As a immunohematologist, Hamblin was a notable and highly respected and regarded researcher and teacher. He is particularly notable as an early pioneer of stem cell treatment for cancer. He made a difference by making the world a better place.
Dr Mike Sutton 26th January 2012.


Article starts here

The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted

For over thirty years a popular myth has been circulating in academic text books, peer-reviewed scholarly journals, popular discourse, university lectures and on the Internet that an accidentally misplaced decimal point in 19th century calculations of the iron content of spinach, exaggerated its iron content tenfold, which was then accepted as true and cited by a multitude of academic studies - all of which failed to check the validity of the figure. This most famous decimal error, it is claimed, led to spinach being erroneously promoted as a good nutritional source of iron and the reason the carton character Popeye's creator, E.C. Segar, chose spinach as the source of his – and the first American superhero’s - amazing powers. The famous Web site Cracked.com even has it as top of the list of the seven most disastrous typos of all time. A search on any search engine, such as Google, will reveal the extent to which this myth is believed and recycled. Earlier this year (Sutton 2010a) I published a primary research paper in the Internet Journal of Criminology that showed there was no published evidence to support the Spinach, Popeye, Iron Decimal Error story and that it appeared to be a myth. In my paper I proved beyond doubt that Segar chose spinach for its vitamin A content (although in fact we know today that spinach contains beta carotene - which the human body converts to Vitamin A).

In this article, I reveal for the first time the results of several months of historical research on iron and nutrition - much of which involved translating old German nutrition text books and academic papers. This article provides the first ever conclusive evidence for the source of the Spinach Popeye Iron Decimal Error Story (SPIDES) and how it occurred. Moreover it goes on to reveal that there most certainly was no decimal place error - but explains exactly why others thought there was.

In my earlier paper on this subject, I traced the source of the story as far aback as I was able to Professor Terence Hamblin’s (1981) article in the British Medical Journal. However, I can reveal here that the original published source of this story was in fact the famous nutritionist, and self-styled myth-buster - Arnold E. Bender. Bender first mentioned it in his inaugural lecture in 1972 (Bender 1972) and later in an article in the Spectator (Bender 1977). In the Spectator, Bender started the myth when he claimed that a German textbook on nutrition (Noorden and Salomon 1920; 476) replicated an earlier decimal point data mistake made by generations of textbooks that unquestioningly replicated erroneous data first published in 1870 by the German scientist E. von Wolff:

“For a hundred years or more spinach has been (and clearly still is) renowned for its high iron content compared with that of other vegetables, but to the joy of those who dislike the stuff this is quite untrue. In 1870 Dr E. von Wolff published the analyses of a number of foods, including spinach which was shown to be exceptionally rich in iron. The figures were repeated in succeeding generations of textbooks – after all one does not always verify the findings of others – including the ‘Handbook of Food Sciences’ (Handbuch der Ernahrungslehre) by von Noorden and Saloman [1] 1920.

In 1937 Professor Schupan eventually repeated the analyses of spinach and found that it contained no more iron than did any other leafy vegetable, only one-tenth of the amount previously reported. The fame of spinach appears to have been based on a misplaced decimal point.”

My primary research paper (Sutton 2010a) cites many scholarly publications whose esteemed academic expert authors - writing on the very subject of the importance of healthy scepticism - unwittingly believe the spinach decimal error myth to be true and so with unintentional irony they use it as an example to support their exhortations on the need for scientists to be healthily sceptical inquirers and always check published 'facts'.

page 2
page 3
Professor Bender, like Professor Hamblin, is an orthodox authority on nutrition. With further irony, Bender is famously a renowned sceptic of junk science. Perversely, these respected scientists failed to check the 'facts' behind the decimal error story before going into print.

Shortly after publishing the primary research paper I received an email from Bonnie Taylor-Blake in the USA . Bonnie had been researching, on-and-off, the decimal error story herself as a kind of hobby. She wrote to congratulate me on my findings and very kindly emailed some clues she had collected that suggested exactly where the origin of the SPIDES lies. The clue was to be found in an article in the Spectator Magazine (Marnham 1981) and an obscure reference to Professor Arnold E. Bender’s inaugural lecture. “Inaugurals,” as they are known in academia, are quite often lavish affairs, presented by way of an introductory speech by the University Vice Chancellor, or else by one of her or his deans, and supported by free food and wine for well over 100 guests. Bender gave his at Queen Elizabeth College, University of London on 24th October 1972 (Bender 1972). What Bender said before his esteemed academic audience started the myth that he published in the Spectator (Bender 1977) five years later (Bender 1972:11):

“One common belief, that spinach is good for you, appears to be due to experimental error since the belief predates the Hollywood nutrition films based on the muscular development of the film star Popeye. I am indebted to Professor den Hartog of Holland for tracing the possible origin of this belief. It appears to date soon after 1870 when Dr E. von Wolff published food analysis showing spinach to be exceptionally rich in iron, a figure that was repeated in many generations of textbooks; it was in the Handbook of Food Sciences (Handbuch der Ernahrungslehre) by von Noorden and Saloman in 1920. In 1937 Professor Schupan analysed spinach for its iron content with µ-µ’-dipyridyl and found the figure to be one tenth of that reported by von Wolff – the fame of spinach may well have grown from a misplaced decimal point.”

After correcting for the fact that Bender had misspelt Salomon as Saloman in both his inaugural lecture and his Spectator article, which led me to suspect that he had never even read the original volume himself, I was able to purchase a copy of Noorden and Salomon via the Internet from an antiquarian book seller in Germany. As a non-German speaker I found Google’s online translator a most invaluable tool for making sense of the big old German text book that I had in my hands.

Noorden and Salomon (1920) present tables for the iron levels of various food stuffs taken from the published findings of Schall and Heisler (1917), von Berg (1913) and Haensel (1908). These are included in Table 1 (below). Noorden and Salomon’s (1920) figure for the amount of iron in dried spinach is 445 mg of iron per 100g. This figure is derived from Haensel (1908), who presented the iron oxide (note: not iron) content of spinach in percentage terms as 0.445 per cent. To convert this figure to mg per 100g it is simply multiplied by 1000 - which is 445.

Professor Bender was completely wrong about the source of Noorden and Salomon’s data on the iron content of spinach. Absolutely none of it came from von Wolff. Not one single figure. The figure Bender thinks came from the work of Wolff in the 1870’s in fact came from Haensel (1908). The origin of Bender’s myth probably lies in Germany and in his failure and that of others to distinguish between iron and iron oxide measurement by these earlier scientists (a mistake I originally made myself). Had Bender read Noorden and Salomon (1920) he would have seen that they were citing several independent examples of German science as cutting edge knowledge 13 years after Sherman (1907) published the exact reasons why be believed that such data were flawed.

The source of Bender’s decimal error belief derives from the fact that accepted knowledge at the time when Bender was writing quite correctly held that dried spinach contained 44.8 mg of iron per 100g (e.g. Jackson 1938). Subsequent research has recorded a figure of 44.6 (Rewashdeh et al 2009). From this it is easy to see where the decimal error idea in Bender’s spinach story came from. Had Haensel (1908) moved his decimal point to give a figure of 0.0445 per cent iron he would have had an accurate figure for the amount of iron in dried spinach of 44.5 mg per 100g.

Most interestingly, when we study what Bender wrote, we cannot be certain whether he actually meant that the error in the iron content of spinach was due to von Wolff, or his assistant or a typesetter misplacing a decimal point. To be fair, at a stretch, he might even have meant that flawed analysis produced a figure that would only have been correct if someone came along and moved the decimal point. Later myth mongers would have it very clearly that this is what happened as they embellished the tale with various versions that add goofy assistants, sloppy typsetters and von Wolf blotting his copybook. But I’m not sure what a jury would make of Bender’s words. Personally, I do not find the professor guilty beyond reasonable doubt. Let me explain.

In his inaugural lecture Bender writes: “ – the fame of spinach may well have grown from a misplaced decimal point.” And later in the Spectator he writes more boldly: “The fame of spinach appears to have been based on a misplaced decimal point.”

page 3
page 4
Bender might have meant that Wolff’s analysis was bad science (although as we now know it was in fact Haensel’s figures Bender was referring to) that led to a figure ten times greater than that later found to be accurate. In other words it is unclear whether Bender means a decimal point was put in the wrong place or whether he meant instead that erroneous science produced a figure with a decimal point in the wrong place.
The strongest evidence of all that Bender did not start out with the intention of creating the misplaced decimal point part of the myth is that he begins talking about the subject in his inaugural lecture by saying: “One common belief, that spinach is good for you, appears to be due to experimental error…”

Clearly we can see here that it is early inaccurate scientific techniques that Bender appears to be blaming.
In support of this conclusion, we should note that Bender fails to provide a citation to any work by Professor den Hartog, whom he thanks for telling him the story. The fact that Bender says he is indebted to Hartog for “tracing the possible origin of the belief” suggests, in the absence of any evidence to the contrary, that the two merely communicated informally - either verbally or in writing - about the story.

What Bender is guilty of, as evidenced firstly in his inaugural lecture, is a foggy style of writing that creates a greatly increased chance of potential misinterpretation upon the message he really wishes to convey about the history of nutritional research into spinach. Five years later in the Spectator Bender’s “Early inaccurate scientific techniques” sentence completely vanishes from his account. Perhaps Bender was seeking to use the Spectator to deliberately start an urban myth? We will never know. But what is most interesting is that eight years later he published a book somewhat ironically entitled: “Health or Hoax?” (Bender 1988).

In Health or Hoax? Bender mentions the fact that spinach is not a good source of iron. And once again gets his facts wrong about Popeye eating spinach for iron. But most tellingly – in a chapter entitled “Old Wives Tales True or False? – Bender makes no mention whatsoever of his earlier misplaced decimal point story.
Perhaps by 1988 Bender had actually undertaken some arduous research and found that in fact it was von Bunge in 1892 who discovered the problems with von Wolff’s (1871) research (see Table 1) and not Professor Schupan in 1937.

Lack of clarity in Bender’s writing may be the key to how the myth first started out on its journey to becoming a culturally embedded super myth. For example, Bender’s Health or Hoax? book is infuriatingly every bit as cryptically unhelpful as his inaugural lecture on the subject of iron in spinach. Even though the book is written for consumption by the general public - it is not until page 197 that one sentence alludes to the fact that what Bender writes on page 132 about spinach being “rich in iron” may not mean those who consume it will get much iron out of it: “Meat is a better source of iron, not only in quantity but because the iron in meat is better absorbed than that of plant food.” Bender writes, while never informing his readers why this is the case, and offering zero advice on whether or not “rich in iron” spinach is a good choice of food for those wishing to eat greens as well as, or instead of, meat to ensure they get enough iron in their diet.

Spinach whispers

As Bender’s SPIDES was re-told by others and re-printed it evolved at twists and turns so that there are so many different variations on the theme.

One version of the story suggests that these turn of the century scientists were not clear about whether they analysed fresh or dry spinach (Walter 2004). This is certainly true in relation to Noorden and Salomon’s (1920) tables because, while they do very clearly distinguish between fresh and dry spinach when presenting tables based on von Haensel’s research findings, they fail to do so for Schall and Heisler (1917) and for von Berg (1913). And their lower finding of 44 mg per 100g, for what we can only reasonably presume to be fresh spinach – given that it is labelled as leaves, stalks and juice - would be accurate were it dry spinach (e.g. Rewashdeh et al 2009).

The data in Table 1 reveal how easy it is to see how a popular variant of the myth that would have it that von Wolff recorded dry spinach as fresh (Coultate 2009) has become entangled with the notion of a misplaced decimal point and other accounts that von Wolff mixed up his figures for dried and fresh spinach. Because moving the decimal point for Wolff’s finding of the amount of iron in fresh spinach one place to the left would give a figure of 5.0 which as Table 2 reveals is close to the range of acceptable recent findings for the mg of iron in 100g of dried spinach (see: Olade and Obarisaide 2009). Moving his figure for dried spinach two places to the left would give a reasonable figure for fresh spinach. However, the idea that Wolff confused dried and fresh spinach and/or made decimal point errors in two observations seems fanciful in light of Sherman's (1907: 43; 53) explanation that poor science knowledge at the time was to blame. This is described in more detail below.

page 4
page 5
Table 1: Older data: Accurate and inaccurate knowledge regarding iron and iron oxide levels in spinach
Type
Units
Value per
100gs
Knowledge Source
Reasonably
Accurate
Discredited Ash Method
Raw Fresh
Dried
mg
mg
50
3,350 (Fe203) 386 (Fe)
von Wolff (1871) ♣
von Wolff (1880)♣
.
mg
2170.6 (Fe203)
Von Haensel (1908)
Other Methods
Raw Fresh*
mg
3.8
Sherman (1907)
Raw Fresh
mg
3.0
Sherman (1907)**
Raw Fresh
mg
36.3
von Haensel (1908)
Raw Fresh***
mg
44-60 (Fe203) 31-42(Fe)
Berg (1913) ♣
Schall and Heisler (1917)♣
Raw Fresh*
mg
4.3
von Bunge (1892)****
Dried
mg
37.0
Atwater and Woods (1895)
Dried
mg
445.0 (Fe203)
von Haensel (1908)
♣ Revision Note: 29 March 2011: Jan Willem Nienhuys very kindly commented on this paper (see comments section) to point out that I made an error in using Wolff's tables for iron oxide here whist assuming them to be for FE (iron). I think my own mistake is interesting - although it may be a pure coincidence - in that the figure of 3,350 mg per 100g (FE203 iron oxide - not iron - FE) would correspond very closely to what we know today to be the real level of actual (FE) iron in spinach if the decimal place were moved two places. In effect, as Jan's comment (see end of this article) reveals, von Wolff (1871) quotes an 1848 result of 50 mg iron per 100 gram of spinach 'fresh substance', and in a later account (1880) he found 3.35 percent iron oxide (5.52 gram) in the ash of 100 gram dry material, corresponding to 386 mg iron. Here then we can see that if the decimal point were moved one place to the left of 386 we would have a pretty good figure for what we know today to be the mg of iron in 100g of spinach. But there is no evidence of a decimal error in von Wolff's work, just the inferior science of the day. As Jan reveals and neatly explains in his comments on this paper, it is important to note that von Wolff and others produce and present different findings for iron and iron oxide. von Wolff (1871) measured 50 mg iron per 100g of 'fresh substance'; and also calculated an average iron oxide content of the ash of 100 gram dry spinach (of two measurements) to be 3.35 percent, which works out to be 386 mg iron per 100g. Wherever I have given a value of Fe203 (iron oxide) - followed by a value for Fe (iron) - in Table 1, these values have very kindly been supplied by Jan Willem Nienhuys. His research results and reasoning, which appear sound to me, are set out in the comments section at the end of this paper.

page 5
page 6
*Reported by Sherman (1907) as spinach obtained from a grocer that contained some 3 to 4 times less water than a perfectly fresh sample which resulted in the average iron content being a third higher than in a perfectly fresh sample.

** Here Sherman (1907) calculated the mean average of the results of a series of earlier studies.
*** Noted by Noorden and Salomon (1920) as leaves, stalks and juice, presumably this is fresh and not dried material but we cannot be certain.
**** As cited by Sherman (1907: 53).

Today (see Table 2) we know that the real figure for fresh spinach is around 2.75 (US Department of Agriculture, 2010andthe real figure for dried spinach is around 45mg per 100g (Jackson 1938; Rewashdeh et al 2009).
Table 2: More Recent Data: Accurate knowledge regarding iron levels in spinach
Type
Units
Value per 100gs
Knowledge Source
Raw Fresh
mg
2.71
USDA (2010)
Raw Fresh
mg
2.5
Council on Foods (1937)
Raw Fresh
mg
1.7 to 3
Stiebling (1932)
Raw Frozen
mg
1.89
USDA (2010)
Raw
Canned
mg
2.30
USDA (2010)
Cooked
mg
3.57
USDA (2010)
Dried
mg
44.8
Jackson (1938)
Dried
mg
44.6
Rewashdeh et al (2009)
Dried
mg
34.0
Olade and Obarisaide (2009)

So is there any possible truth at all in the spinach, in decimal error story?
On the basis of the evidence presented so far, there is at least the possibility that von Wolff made two errors: (1) he recorded dry spinach as fresh, and that in addition he compounded this error when (2) somehow he or a typesetter, or some other person accidentally moved the decimal point in his findings two places to the right. An alternative possible explanation is that von Wolff worked on fresh spinach just like he said he did and that either he or a typesetter, or some other person, accidentally moved the decimal point a massive three places to the right.

Rather than simply accept these simple and compellingly feasible explanatory possibilities, as others have done, and present them as entertaining facts we should look for alternative possibilities. A good place to start is to examine whether perhaps the science of measuring iron levels in foodstuffs was not particularly good at the time and whether other scientists – working independently – came up with similarly high measures of the amount of iron in spinach.

Writing in the first decade of the 20th century the German biochemist Haensel (1909: 9) notes that he is aware of only two studies that have analyzed the iron content of spinach. Paying tribute to but failing to cite: “recent studies on iron in food by G. von Bunge”, Haensel (1909:9) writes that all findings in his paper are based on his own analysis of the precise amount of iron in spinach.

page 6
page 7
Haensel analysed the iron content of spinach by various methods, and each time presenting in his paper two of each of his findings for the percentage of iron oxide. Haensel (1909: 14) writes that, contrary to (non-cited) earlier findings of others his own analysis, using the incinerated plant ash determination method, found that spinach was not in fact the richest plant source of iron and that lettuce, winter cabbage and endivien contained more iron.

If Emile von Wolff’s hugely exaggerated reported findings of iron in spinach were due to a decimal error alone then the decimal point was out by three decimal places. Moving his decimal point three places to the left gives a respectably accurate figure of 2.17 mg per 100g for fresh spinach. If by some amazing coincidence Haensel (1909) made a decimal place error as well – one that was reproduced by Noorden and Salomon (1920) without question – in his non-ash analysis of spinach then he made the same basic accounting error in recording the iron level for every plant he analysed; because using various methods of analysis Haensel consistently has winter cabbage, lettuce and endivien as containing slightly more iron than spinach.

Haensel’s findings for iron in fresh spinach are similar to the lowest levels found by Berg (1913) and Schall and Heisler (1917) (see Table 1). Given the similarity in their exaggerated findings, it seems more likely that the analysis method used by all three scientists, and some 45 years earlier by von Wolff, is to blame for these exaggerated figures, rather than a peculiarly coincidental series of separate decimal place accounting goofs. And, most clearly, the data presented in Table 2 exonerates von Bunge from the SPIDES because his 1892 figure is accurate. Bunge then is at least one early pioneer of nutrition who finally can stop turning in his grave as the truth is finally revealed for the first time in this article

Bad Science was to Blame for Exaggerations in the Iron Content of Spinach

More than 100 years ago Sherman (1907: 43; 53) in the US Department of Agriculture’s Office of Experiment Stations-Bulletin 185 explained in plain English the various ways that earlier biochemistry methods – such as those employed by Wolff and later by Haensel - exaggerated the iron content of foodstuffs:

“Little weight can be attached to the statements regarding iron which are to be found in the standard compilations of ash analysis, as these are based largely upon results obtained by methods which greatly overestimated the iron. Generally speaking, it is only since the discussion of iron in food materials was begun by Bunge in 1885 that analyses have been made with special reference to the determination of iron, and the amount of data which appears trustworthy is not yet very large.”

“Undoubtedly…many of the ash analysis of Wolff and Konig greatly overestimate the iron content of green vegetables, as was pointed out by Bunge in 1892.”

The various reasons Sherman’s Bulletin 185 gave for early bad science in the early high determination of iron levels of various foodstuffs, including plants and meat, can be summarised as follows:
  • Contamination of precipitates (the dissolved or suspended substances in solution)
  • Using less than completely fresh, and therefore somewhat dehydrated “fresh” plant material.
  • Volatilization of iron as ferric chloride during the ignition of the sample.
  • Contamination with iron from utensils.
  • Iron contamination from laboratory dust, including that from charcoal used in the heating process.
  • Iron in the reagents used.
  • Small amounts of platinum from dishes used in burning of the samples.
Sherman reveals then that at the turn of the 20th century American scientists were quoting von Bunge’s accurate findings of the iron levels of spinach and were well aware that von Wolff’s hugely exaggerated findings were due to unreliable methods, rather than any kind of simple decimal place transcription error.
Sherman shows us that by 1907, American scientists were more accurate, knowledgeable and ahead of the Germans on the iron in spinach issue. And because their clearly written knowledge was made available by way of the US Government’s Printing Office, it seems improbable that Noorden and Salomon’s (1920) tomb, which was never translated into English, would have influenced American scientists.

page 7
page 8





Clearly then the popular myth that is being credulously recycled by journalists, nutritionists and scientists (e.g.Gates 2010) that: “The mega-iron myth first began in 1870 when Dr. E. von Wolf misplaced a decimal point in his publication which led to an iron content figure that was ten times too high. Although investigated in 1937 by the Germans, the rumor remained strong for decades (thanks to a pipe-smoking sailor man).” Is a myth about a myth.

The myth busting Jan Willem Nienhuys, who is secretary of the Dutch organization Skepsis, kindly points out in his comment in the comments section on this paper that:

“It is not quite correct to say that Wolff used bad methods: he reports various values for an enormous number of substances all or most of them derived from measurements found in the literature. He must have known about these wildly varying iron contents, but what can one do if one reports research of others?”

Discussion and Conclusions

When respected academics and sceptics, with painful irony, erroneously believe a myth to be true and so use it as an example of the need to be healthily sceptical they create the unintended consequence of bracing the myth (Sutton 2010b). Braced myths are a kind of reinforced super myth; quite what this means in terms of how difficult it will be to kill such myths remains to be seen. I suspect that supermyths are particularly hard to kill because they originate in an atmosphere of supposed concern with veracity and mythbusting.

Bender was wrong that the figures for iron in spinach in Noordon and Salomon came from Von Wolff. They, came from another scientist called Haensel, and he, like other 19th century and early 19th century German scientists produced figures that were for the most part wrong, because the methods used to measure iron were inaccurate.

In the early 20th century, German scholars were using old methods that American's such as Sherman had discredited. This may be because, for the first two decades of the 20th century, American advances had not reached the German scientists cited in this article.

One thing is certain and that is that 20th century Americans were certainly never influenced by old German exaggerated iron levels in spinach, and there is no evidence that the British were either. Moreover, there never was a decimal point error in any reported figures, although it is quite easy to spot many different examples in old papers and books where sloppy modern scholars reading those early accounts might mistakenly draw that false conclusion for a variety of reasons discussed in this article.

At least perhaps now at last, Segar and the much maligned von Wolff and von Bunge can stop turning in their graves as the SPIDES is laid to rest.

There is always a little variation between different studies and samples, but the currently accepted general level of iron per 100mg of fresh, canned and dried Spinach and per 100 mg of fresh spinach are well represented in the table below.

Raw Fresh
mg
2.71
USDA (2010)
Dried
mg
34.0
Olade and Obarisaide (2009)
Raw
Canned
mg
2.30
USDA (2010)
Looking at the difference between raw and fresh spinach you can see where the latest myth of a decimal point error came from. But remember Bender never said that the error came from comparing dry with fresh spinach. That new variety that sprang from Bender’s original myth came along as the myth evolved with more story telling by other people seeking, as do all mythmakers, to fill in the gaps in our knowledge.
Segar, Popeye's comic strip creator never once had Popeye eat spinach for Iron, it was only for vitamins A (which is in fact the beta carotene in spinach that our bodies convert to Vitamin A. (see Sutton 2010).

page 8
page 9



Popeye only ever ate spinach to get Vitamin A - never for iron!
Confusingly, however, Popeye with his iron-anchor tattoos and “I yam an Iron man” patter simply adds more fuel to the confused mix in the minds of the public to this day. Perhaps it’s a good job “steel” is not a nutritional element in foodstuffs, or else Superman “man of steel” would be causing problems too!
Segar was comic strip artist only, he had no say and no control over the imagery and story-lines of the
Iron Metaphor from the
 superbly innovative 1930's Popeye Movie
 "A Dream Walking"
Drawn by Mike Sutton
Popeye movies. Those movies contained many visual spinach and iron metaphors. For example in the 1933 cartoon movie "A Dream Walking " Popeye eats spinach and his bicep turns into an old clothes iron - in other films his muscles become anvils and his fists hammers etc - all of which may have misled many people to think that spinach was a good source of iron.

People, it seems, want simple and entertaining ironic stories to fill their knowledge gaps. But the truth is often inconveniently complex and even more ironic. Deeply embedded in this story is the facts that the myth, though simple in the telling, hides a multi-layered, complex reality. And the myth about spinach iron and Popeye, decimal point errors and German scientists has evolved through various diverging ramifications to be best circumstance suited to survive in our popular culture. 

If you would like to know more about the story and mythology of iron and spinach, my first paper on the topic is Iron Spinach and Popeye ; others are:Spin@ge USA: Beware of the Bull

Timeline For Busting the Spinach Popeye Iron Decimal Error Myth

1871
  • Von Wolff’s 'bad science' is said to have exaggerated spinach iron content
1892 - Swiss
  • Von Bunge gets it right
1907 - USA
  • Sherman explains 19th Century bad science
1920 - Germany
  • Noorden & Salomon still citing old German bad science
1972 - UK
  • Bender publishes spinach Popeye iron decimal error myth
1981 - UK
  • Hamblin braces Bender’s myth
2010 - UK
2010 - UK


page 9
page 10
(
Aschenbestandteile). Dresden. Taken here from: Noorden, C. and Salomon, H. (1920) Handbuch Der Ernährungslehre Erster Band Allgemeine Diätetik. Berlin. Julious Springer.
Bender, A. (1972) The Wider Knowledge of Nutrition. Inaugural Lecture. October 24. Queen Elizabeth College., University of London. Somerset. Castle Cary Press Ltd.Bender, A. (1977). Iron in spinach. Spectator. p.18. July 9.
Bender, A. (1988) Health or Hoax? London. Elvedon Press.
Berg, R. (1913). Die Nährungs- und Genußmittel
Coultate, T. 2009. Food the Chemistry of its components. 5th Edition. Cambridge. Royal Society of Chemistry Publishing.
Gates, D. (2010) The Body Ecology Guide to the Ten Healthiest Greens http://www.bodyecology.com/07/08/30/healthiest_greens_guide.php
Haensel, E. (1909) Über den Eisen- und Phosphorgehalt unserer Vegetabilien. Biochem. Zeitschr 16. 9.
Hamblin, T.J. 1981. Fake! British Medical Journal. Volume 283. 19-26th December. pp. 1671-1674.
Jackson, S. H. 1938. Determination of Iron in Biological Material. Industrial & Engineering Chemistry Analytical Edition10 (6), 302-304.
Marnham, P. (1981) Postscript Counteract. The Spectator. Jan 3. page 26.
Noorden, C. and Salomon, H. (1920) Handbuch Der Ernährungslehre Erster Band Allgemeine Diätetik. Berlin. Julious Springer.
Oladele, O.O. and Aborisade A.T. (2009) American Journal of Food Technology.Volume: 4. Issue: 2. pp.: 66-70
Rewashdeh et al 2009. Iron Bioavailabilty of Rats Fed Liver, Lentil, Spinach and other Mixtures. Pakistan Journal of Biological Sciences. 12 (4) 367-372.
Schall, H. and Heisler, A. (1917) Nahrungsmitteltabelle . 5. Aufl. Taken here from Noorden, C. and Salomon, H. (1920) Handbuch Der Ernährungslehre Erster Band Allgemeine Diätetik. Berlin. Julious Springer.
Sherman. H.C. (1907). Iron in Food and Its Functions in Nutrition. Office of Experimental Stations Bulletin 185. May 25. p56.
Stiebling, H. K.(1932) United States Dept. of Agriculture, Circular No. 205.
Sutton, M. (2010a). Spinach, Iron and Popeye:Ironic lessons from biochemistry and history on the importance of healthy eating, healthy scepticism and adequate citation. Internet Journal of Criminology (Primary Research Paper series). http://www.internetjournalofcriminology.com/Sutton_Spinach_Iron_and_Popeye_March_2010.pdf
Sutton (2010b) Discovery of Braced Myths. Supermyths blog. September 24th. Available online: http://super-myths.blogspot.com/search/label/Discovery%20of%20braced%20myths
Wolff, E. (1871) Aschen-Analysen von Landwirthschaftlichen Producten Fabrik - Abflällen und Wildwachsenden Pflanzen. Berlin. Wiegandt & Hampel.
Walter (2004) Astronautics – that’s what you get out of it. Institute of Astronautics. http://www.lrt.mw.tum.de/en/interessierte/fs_nutzen_der_raumfahrt.phtml

[1] This is a misspelling by Bender of Salomon.

Alan Crozier
May 6, 2016 at 9:22 am
K Sune Larsson
Reading your excellent article in the Internet Journal of Criminology, I see you refer to K. Sune Larsson as "she" and "her". Here is his obituary:
http://avestatidning.com/narakara/minnesord/1.2224141-en-hembygdens-rost-har-tystnat-k-sune-larsson
Author's Post
Mike Sutton
May 7, 2016 at 2:59 am
Daer Alan
Many thanks
Yes its my mistake.
Unfortunately we have a strict policy of not altering articles in the IJC once published.
When I write the story fully up from the several articles and blogs I have written on the topic I will be sure to address my error.
Many thanks for pointing it out.
Jan Willem Nienhuys
April 12, 2011 at 10:16 am
I have checked what E. Haensel (no "von") wrote in 1909. Haensel applies the ash method and he gives the result for 27 vegetables. Interestingly he mentions (on p. 14) that the often prevailing opinion that spinach is the vegetable that is the richest in iron is incorrect. Lettuce contains (so his analysis) about twice as much.
Haensel gives the amounts of iron oxide as percentages of fresh weight (his Table I), of dry weight (his Table II) and of ash weight (his Table III). As percentage of ash he provides 2.1937% (first measurement) and 2.1476% (second measurement). His Table V gives the same numbers, but in Table V the vegetables are arranged in decreasing order of iron content. I remark that the average of these percentages is 2.17065%, which would mean 2170.6 mg iron(III) oxide in 100 gram of ash. This explains the somewhat mysterious number in the summary of the 'Discredited Ash Method' on page 3 of this exposition.
I am no chemist. Haensel gives a detailed description of the manner in which he did the iron determination. If experimental error introduced iron from other sources into the sample, then I don't understand why there are such wildly varying amounts in different substances. Why should for example bananas only yield about 100 mg iron oxide per 100 gram of ash? This not once, but two different measurements giving about the same result. The data are apparently wrong, but I suspect a different reason than just a bunch of impurities from the lab.
In the story above the impression is given that after the publication of von Bunge almost all scientists were in the position to the know the truth, but it is not so clear that between 1920 and 1935 scientists really knew what were the real iron contents of many vegetables, and implicitly also the reason older measurements were off so dramatically.
Author's Post
Mike Sutton
April 12, 2011 at 11:58 am
Jan
Many thanks. You make two distinct points and I think these need two distinct replies from me.
(1) I agree the banana conundrum does require some explanation and does raise questions about Sherman's (1907) explanation for the exaggerated findings of earlier scientists. That looks to me like a useful avenue for further research.
(2) That said, Sherman does quote Bunge's figure as accurate and thereafter most scientists in the USA (as Sherman and the various USDA reports and scholarly articles show) provide fairly accurate figures by today's standards. Outside of the USA, however, the scientists referenced 13 years later in Noorden and Salomon (1920) seem to remain curiously way-off the mark still from 1908 to 1913. And from that cause, the point I make above in the article is that the SPIDES myth is wrong in its main assertions, which are, (a) there was a decimal place error that was credulously believed by Americans and the rest of the world (b) that was only disproven in the mid 1930's by German scientists re-checking the old data. In both of those main strands the myth is certainly bust.
I do agree of course that the reason why the early figures are so high most certainly needs to be examined in more depth than it appears to have been.
Reference
Noorden, C. and Salomon, H. (1920) Handbuch Der Ernährungslehre Erster Band Allgemeine Diätetik. Berlin. Verlag Von Julious Springer.
Sherman. H.C. (1907). Iron in Food and Its Functions in Nutrition. Office of Experimental Stations Bulletin 185. May 25. p56.
Jan Willem Nienhuys
April 14, 2011 at 8:17 am
Mike,
I think you summarize the points at issue correctly. One extra point: from my experience as physics student, science students all over the world practice by redoing old analyses, just like pupils in high school reprove the theorem of Pythagoras. That in itself guarantees that any serious mistake is caught. On the one hand this means that when the ash method is wrong because of impurities, people will try to redo the analysis with better methods, but when there is another basic error it may not be caught. On the other hand, it is outright silly to think that something like the mineral content of a popular food is measured once in 1870, and that no one (agricultural and health authorities, students in universities or researchers who develop new techniques of analysis) will ever bother to check such a thing.
In alternative medicine such silly thoughts abound: the ancient healing methods of a mythical king in the Chinese Stone Age, the strange teachings of 18th century physicians are promulgated as if faithful copying of their ideas guarantees the preservation of the wisdom of the ancients. In cancer quackery one will find recommendations to follow diets devised in the early 20th century. Conversely it is customary to scorn 'science' thusly: 'So science says this and that, but who is to say that it won't change opinion next year, they are constantly revising their ideas, so the present ideas are probably wrong. That's why I stick to age-old reliable wisdom that has stood the test of time.'
May I add some remarks concerning spelling:
1. Von Julious -> Julius (von means just "of" or "by", so Verlag von Julius Springer just means "Published by Julius Springer"
2. Nahrungs und Genubmittel (Ashenbestandteile) -> Nährungs- und Genußmittel (Aschenbestandteile)
(Instead of Genußmittel you can also write Genussmittel)
3. Haensel, von. E. (1909) uber den eisen- und phosphorgehalt unserer Vegetabilien. -> Haensel, E. (1909) Über den Eisen- und Phosphorgehalt unserer Vegetabilien.
4. similar correction is the ref. of Schall, H. and Heisler, A.
Jan Willem Nienhuys
April 4, 2011 at 11:16 am
Another correction. I checked von Noorden and Salomon. This is a huge book of 1237 pages. On page 476 there is a table with 58 vegetables, and the iron listings ar obviously for fresh substance. But they are for iron(III)oxide, so they refer to the weight of iron oxide in the ash. So I think that if one wants iron rather than iron oxide, one should shange 44-60 into 31-42.
The same holds for the data on 24 different kinds of vegetables taken from E. Haensel, Über den Eisen- und Phosphor gehaklt unser Vegetabiloien, Bioch. Zeitschr. vol. 16, Sept. 1909. There we see 445 mg iron(III) oxide in 100 gram of dry substance; as fresh spinach contains 8.07% dry matter, this amounts to 311 mg iron in the dry matter, and 25 mg iron in the fresh spinach.
The two tables differ quite a lot (numbers all in terms of iron oxide)
the first table has celery stalks and leaves: 5.4 mg/100g the second table has celery leaves 39.5 mg/100g;
the first table has kohlrabi stems listed as 35 mg/100g, the second table has 6.2 mg/100g
the first table has chanterelle at 29 mg/100g, the second table says 10.9 mg/100g the first table has winter radish (Raphanus sativus subsp. niger) at 64 mg/100g, the second one 4.1 mg/100g
the first table has tomato at 23 mg/100g and the second table at 0.7 mg/100g.
The authors remark that the given averages are not reliable. 'For no other mineral this is true in the same degree as for iron.'
Incidentally, as remarked, von Noorden and Salomon hardly refer to von Wolff. They mention him twice, not in connection with this subject, and only through quotations of others.
I think that the number 2170.6 mg of E. Haensel is wrong. It probably means 2.17% iron oxide content of the ash of 100 gram dry material, which would be about 250 mg iron per 100 gram of dry material, which is in the right order of magnitude for this results from ash analysis.
Obviously one shouldn't trust von Noorden and Salomon, but they have elsewhere something to say about preparing spinach. Even if you boil it for only 5 minutes, about half of the mineral content goes away. If you boil it even longer als half of the 'organic' substance goes away. The best way of preparing is boil for 1 minute and then steam it.
Jan Willem Nienhuys
March 31, 2011 at 5:30 am
Tho values in Wolff kept nagging me. I looked again, and I think I know where the data on page 128 of the Zweiter Theil come from. Wolff has just averaged the data of the two lines 51 and 52 on page 101. So, at least for spinach, the 1880 book doesn't contain new data.
The only exception that the data are not precisely averaged is the amount of chlorine, but as I remarked before, the calculation of chlorine is somewhat mysterious. Apparently the ash contains chlorine in bound to metals, but the analysts determined the total amount of K, Na, Fe and so on, and then somehow converted this back to a value for the oxides. That is the reason why the percentages do not exactly add up to 100 but to a bit more. Possibly the chlorine that escapes in gaseous form during the combustion of the samples was measured in a different fashion.
New proposal for correction:
von Wolff 1871: 50 mg iron per 100 gram 'fresh substance'; average iron oxide content of the ash of 100 gram dry spinach (of two measurements) 3,35 percent.
Author's Post
Mike Sutton
April 6, 2011 at 5:58 am
Dear Jan - once again - very many thanks.
I made the changes in both the table and the explanation - with attribution to your work. It certainly clears up a lot of the confusion at last.
Mike
Author's Favorite
Jan Willem Nienhuys
March 30, 2011 at 3:48 pm
I think to be most fair to von Wolff, you would need two lines:
1. von Wolff 1871: quotes 1848 result of 50 mg iron per 100 gram 'fresh substance'
[ the computation for this: item 52. on page 101 part 1: iron constitutes 111.7/159.7 = 0.699436 of Fe2O3. So 100 g fresh stuff gives 9.47 g dry stuff (see footnote on p. 101). 16.7 percent of this is the weight of the pure ash, makes 1.568149 g of ash, of which 4.6% is iron oxide, so 72.13 mg oxide, giving 50.45 mg iron. You better check it, because I do make mistakes. In my 5.37 AM post I mixed up the data from the two lines. I can understand why von Wolff in 1880 as service to the reader provided tables where the constituents of the ash were given in gram rather than percentages. His readers didn't have calculators ready.]
2. Wolff 1880: 3.35 percent iron oxide (5.52 gram) in the ash of 100 gram dry material, corresponding to 386 mg iron
[calculation from page 128: 3.35 percent iron oxide by weight of 16.48 gram pure ash is 0.55208 g iron oxide (just as page 147 says except there the basis is 1000 parts of 1 gram each), is 386 mg iron in 100 gram of dry material. The reason to mention von Wolff 1880 is that anyone thinking that von Wolff would be quoted in the twentieth century, surely would think that the latest edition was referred to. Moreover 386 mg is about 9 times the true value. Together these two lines should make clear what von Wolf did, and the reader can look up von Wolff, just type Spinat in the search box and you get all pages where that word occurs. ]
Author's Post
Mike Sutton
March 31, 2011 at 11:23 am
Jan
Once again - many thanks for providing this data. It seems to me to check out. I've updated my comment on von Wolff's figures in my Table 1 - with the attribution to your additional research.
As far as I can tell, the most likely explanation for Wolff's later 9 to 10-fold exaggerated figure of the iron content of spinach is inferior scientific techniques of the kind explained by Sherman - rather than a decimal error. Would you agree?
Author's Post
Mike Sutton
March 30, 2011 at 12:10 pm
Dear Jan
I'm glad to see I'm not the only one who has struggled with making sense of this data.
I'm going to bow completely to your mathematical superiority to my own. And your superior interpretation of Wolff’s publication as well.
Therefore, so that I can be sure of getting this right, can I ask very simply:
(a) What figure should I type in my Table 1 in this article for Wolff's mg per 100g for FE and
(b) what figure should I type in that table for Wolff's mg per 100g of iron oxide?
Again - your help with clarifying and rectifying the information in this paper is very much appreciated.
Mike
Jan Willem Nienhuys
March 30, 2011 at 12:01 pm
"Reported by Sherman (1907) as spinach obtained from a grocer that contained some 3 to 4 times less water than a perfectly fresh sample which resulted in the average iron content being a third higher than in a perfectly fresh sample"
Here something most be wrong. If some kind of vegetable contains 90 gram of water and (say) 10 mg of substance X, then reducing the water by one third, gives you: 60 gram of water, 10 gram of nonwater, of which 10 mg of X. Then the X-content of the dried out stuff is (10 divided by 0.7) mg per 100 gram.
So if the iron content goes up by a factor 4/3, the weight loss by drying out should be a factor 3/4. If it true that fresh spinach contains about 94% water, then we have 100 gram fresh spinach = 94 gram water + 6 gram nonwater gives 75 gram dried out spinach = 6 gram nonwater + 69 gram water, so about a quarter of the water is lost.
If the same spinach loses 3/4 of its water then we get 29.5 gram of spinach of which 6 grams of nonwater, and the nonwater contents are multiplied by 1/0.295 = about 3.4
Incidentally, did Bender really spell 'Hundbuch der Ernahrungslehre'? That should be 'Handbuch der Ernährungslehre' if he copied the title correctly. Hundbuch means 'dog's book'.
Author's Post
Mike Sutton
March 30, 2011 at 12:23 pm
Dear Jan
It looks like you found something there.
Actually I just checked Bender in the Spectator magazine article - and found that was my spelling error. Looks like I made a bit of what we English call a "dogs dinner" of that one.
Many thanks Jan. I've just corrected it,
Jan Willem Nienhuys
March 30, 2011 at 11:31 am
It is my turn to be a little red-faced about misinterpreting results. I mentioned different results on page 128 and 147. But after examining the differences carefully and also after reading the foreword to part 2, it turns out that on page 147 the amounts given are in the same dimension as total ash. So 1000 gram of dry material gives 164.68 gram of purified ash, which then contains 5.52 gram of iron(III)oxide. You can see this must be the correct interpretation if you add up all numbers in the row. You get something close to 164.8.
I think I know why von Wolff was implicated. It is a book with an enormous amount of data, possibly very famous, so if someone is claiming that 'old data' are wrong without looking up the details, then 'von Wolff' would be the logical choice.
Jan Willem Nienhuys
March 30, 2011 at 8:07 am
One more remark: the Wolff source actually consists of two parts: the fist part is published 1871, and the mentioned page 101 refers to this. The Zweiter Theil is subtitled: results 1870-1880. I still don't see the relation between the data on p. 128 and p. 147. In the first part he quotes sources, in the second part he merely states how many analyses (presumably by others) are included.
Author's Post
Mike Sutton
March 30, 2011 at 10:01 am
Dear Jan
Thank you so much for helping me correct and improve this paper.
I had hoped that putting it up on a peer-to-peer site such as this one was a good way to go. It turned out to be a good hunch. This is most useful information that actually puts von Wolff in a better light.
I hope you don't mind, but I quoted some text from your comment below in the paragraph just prior top the conclusion - mentioning your association with Skepsis
Regards
Mike.
Jan Willem Nienhuys
March 30, 2011 at 5:37 am
Hello Mike,
I took another look at your reference
http://www.archive.org/details/aschenanalysen01wolfrich
but most of those tables of the Aschen-Analyse have in the column heading between MgO and P2O5 the heading Fe2O3. In some tables (of trees) Mn2O3 occurs. Only on page 131 the column heading has the misprint F2O3. Throughout the ash-analysis Wolff gives the percentages of the composition of the ash (all percentages add up to 100), and naturally there are only oxides in the ash. Only chlorine is not listed as oxide. On page 3 it is explained how to interpret this, but I must admit I cannot - as a non-chemist - quite understand it.
I missed one page with spinach: on page 101 it says that 100 parts of dry material leaves in the case of spinach 19.76 parts of 'Rohasche' (unpurified ash), of which 10.07% sand and carbon and 7.41% carbondioxide, which leaves after subtraction 16.30 parts (my computation), but in the table 16.27 parts (some kind of rounding error or small experimental error). Of these, 2.1 parts in 100 are iron(III) oxide.
This means 1.19 parts per thousand iron of the original dry material, i.e. 119 mg per 100 gram. This is the measurement of a certain Saalmüller, and/or Strumpf in 1851 (I think Saalmüller as quoted in Strumpf). See footnote. Below that item 52 is also about Spinat, but it is a different measurement, obtained from Richardson, 1848.
Unpurified ash: 21.52 parts obtained from 100 parts of dry material. Purified ash 16.58 parts, iron (III) oxide 4.6% of the ash, making 226 mg iron per 100 gram of dry material. The reference is quoted as saying that the original fresh substance contained 90.53% water, and this would mean 25.26 mg iron per 100 gram of fresh spinach.
Observe that the modern values (2,75 mg and 44,6 mg) per 100 gram of fresh/dry material imply that the water content of fresh spinach nowadays is about 94% . It is difficult to guess why English spinach in 1848 would have less water than modern American spinach, but maybe the stuff wasn't quite fresh anymore, and had lost already about half of its water by the time it was on the lab scales.
Anyway, this kind of answers your question about iron versus iron(III) oxide.
It is not quite correct to say that Wolff used bad methods: he reports various values for an enormous number of substances all or most of them derived from measurements found in the literature. He must have known about these wildly varying iron contents, but what can one do if one reports research of others?
The data in the appendix are a kind of averages, but I don't quite see the relation between these four numbers 119, 226, 193, 318.
Author's Post
Mike Sutton
March 29, 2011 at 12:00 pm
Jan Willem Nienhuys

New Article Comment

Jan Willem Nienhuys posted the following comment on The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted:

I find this article extremely interesting. However, I don't quite understand what it says about Emil von Wolff. Fortunately the book is online, but it says two contradictory things, which do not correspond to anything in the article.

1. In the table in the Zweiter Theil, p. 128 it states that the ash of Spinat (spinach) is 16.48 percent of the dry material, and that it (the ash) contains 3.35 parts Fe2O3 per 100. My calculation says that this amounts to 193 mg per 100 gram of dry material.

2. On page 147 (also of the Zweiter Theil) the amounts of ash are given as parts per 1000, and the fraction Fe2O3 is given as 5.52 parts per 100, amounting to 318 mg per 100 gram of dry material.

The 3.35 parts per 100 do superficially correspond to 3,35 gram (3,350 mg) per 100 gram, but they refer to a fraction of the ash, and also to iron(III) oxide rather than to iron. It is not clear to me what the source of this ‘3,350 mg per 100 gram’ is.
Author's Post
Mike Sutton
March 29, 2011 at 12:54 pm
Dear Jan
I am not sure why your comment was not automatically published (it came into my mail box and I published it here myself.
Anyway, thank you for your comment. It is very much appreciated.
You are correct in that I took the figure of 3,350 mg per 100 g from the table on page 128 of Dr Emil Wolff's Ashen-Analysen von Landwirthschaftlichen Producten (1871).
I printed the relevant tables off from the internet and I can see now that the table is indeed as you say (at least it looks like) for F203 - not FE203. I'm not a chemist - and I had to use Google Translate to translate the text so I can understand it (I'm afraid I don't speak German) but I'm guessing F203 and FE203 are both Iron oxide. Is that right?
It looks like I most certainly made an error (not being a chemist) and have indeed used Wolff's tables of iron oxide here whist assuming them to be for FE (iron).
My own mistake is interesting - although it may be a pure coincidence - in that the figure of 3,350 mg per 100g (FE203 iron oxide - not iron - FE) would correspond very closely to the real level of actual (FE) iron in spinach if the decimal place were moved two places.
Does Emil von Wolff have any tables for FE in his book do you know? If so then I would dearly love to know what his calculations were for FE in spinach – converted to mg per 100g.
I will add a footnote to the table in my article here on Best Thinking - by way of a revision. With an attribution to Jan Willem Nienhuys
Kind regards
Mike Sutton
Author's Favorite
Jan Willem Nienhuys
March 29, 2011 at 11:06 am
I find this article extremely interesting. However, I don't quite understand what it says about Emil von Wolff. Fortunately the book is online, but it says two contradictory things, which do not correspond to anything in the article.
1. In the table in the Zweiter Theil, p. 128, it states that the ash of Spinat (spinach) is 16.48 percent of the dry material, and that it (the ash) contains 3.35 parts Fe2O3 per 100. My calculation says that this amounts to 193 mg per 100 gram of dry material.
2. On page 147 (also of the Zweiter Theil) the amounts of ash are given as parts per 1000, and the fraction Fe2O3 is given as 5.52 parts per 100, amounting to 318 mg per 100 gram of dry material.
The 3.35 parts per 100 do superficially correspond to 3.35 gram (3,350 mg) per 100 gram, but they refer to a fraction of the ash, and also to iron(III) oxide rather than to iron. It is not clear to me what the source of this ‘3,350 mg per 100 gram’ is.
Author's Post
Mike Sutton
March 29, 2011 at 12:58 pm
You are correct in that I took the figure of 3,350 mg per 100 g from the table on page 128 of Dr Emil Wolff's Ashen-Analysen von Landwirthschaftlichen Producten (1871).
I printed the relevant tables off from the internet and I can see now that the table is indeed as you say (at least it looks like) for F203 - not FE203. I'm not a chemist - and I had to use Google Translate to translate the text so I can understand it (I'm afraid I don't speak German) but I'm guessing F203 and FE203 are both Iron oxide. Is that right?
It looks like I most certainly made an error (not being a chemist) and have indeed used Wolff's tables of iron oxide here whist assuming them to be for FE (iron).
My own mistake is interesting - although it may be a pure coincidence - in that the figure of 3,350 mg per 100g (FE203 iron oxide - not iron - FE) would correspond very closely to the real level of actual (FE) iron in spinach if the decimal place were moved two places.
Does Emil von Wolff have any tables for FE in his book do you know? If so then I would dearly love to know what his calculations were for FE in spinach – converted to mg per 100g.
I will add a footnote to the table in my article here on Best Thinking - by way of a revision. With an attribution to Jan Willem Nienhuys
Kind regards
Mike Sutton
Author's Post
Mike Sutton
January 3, 2011 at 4:41 am
Timeline
Many thanks Erwin.
That’s a good idea.
Off the cuff, I think there might need to be two timelines shown together: one that shows the time line of the myth itself and the other that shows how it runs parallel and often (but not always) contrary to what was/is current knowledge. I’ll certainly have a go at it.
Meanwhile, in a few days time I’ll be publishing on Best Thinking my third article on the subject. The findings for this next article came out of my research into the SPIDES myth. This one will show that the United States Department of Agriculture is spreading potentially harmful bull**** about spinach on the net – using standard pseudo-science tricks.
I’ll have a go at that timeline next.
Regards
Mike
Erwin
January 2, 2011 at 3:06 pm
Instructive
Awesome article, thanks!
It would be very useful to have something like a timeline at the end of the article that summarizes the various steps in this crazy story. It would make it much easier to spread the gospel :).
Author's Post
Mike Sutton
February 5, 2011 at 5:19 pm
Hi Irwin
I added that timeline you suggested into the spinach decimal error article. Hope its useful.
Best wishes
Mike
Author's Post
Mike Sutton
December 23, 2010 at 12:08 pm
Dear Professor Hamblin
I must admit to being somewhat perplexed at the amount of work that I put into getting to the bottom of the story – in what started out as something that was meant to be no more than a 30 minute fact checking exercise into a famous story of the impact of bad data on policy.
I’m not sure how useful or interesting the end result is for others but the research got me involved in reading a lot of literature on pseudo scholarship, bad science etc. I have found a couple of other examples of myth-busters ironically reinforcing myths. One of them is in my own field of criminology, which has had an enormous impact on the decline of beat policing in the UK and USA. So the concept of braced myths (myths braced by orthodox sceptical authorities) might be worth some further investigation.
As someone who seems to be embarking on busting myth busters I’d better do my best not to hoist myself by my own petard. And if I do, I hope I can be as forthright and generous as you in thanking the critic.
Author's Favorite
Terry Hamblin
December 23, 2010 at 7:36 am
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

1 comment: