This is part of a series checking Sutton’s claims regarding his “myth busting”
CLAIM
Myth — The cartoon character, and first American superhero, Popeye ate spinach for its strength providing iron content because his creator was misled by the publication of a widely believed 19th century decimal point error that exaggerated the iron content of spinach.
Fact — Popeye’s creator E.C. Segar had Popeye eat Spinach for its Vitamin A (in fact beta carotene) content, never once for iron. See Segar, E. C. (2007) Popeye: Well Blow Me Down. Vol. 2. Seattle. Fantagraphics Books. Page 162. Spinach is actually a poor source of nutritional iron due to its oxalic acid content. Moreover, the story of the decimal point error is a supermyth *. For the full spinach and Popeye myth-bust See: Sutton (2010) and Sutton (2012a).
- * “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).” (Sutton 2017 Dec. 11)
- Sutton (2010) SPINACH, IRON and POPEYE. Internet Journal of Criminology.
- Sutton (2012a) The Spinach, Popeye, Iron, Decimal Error Myth is Finally Busted.
TEST
This case is convoluted, but it can be split into three distinct claims made by Sutton, and the corresponding test results:
- CLAIM: “E.C. Segar had Popeye eat Spinach for its Vitamin A“
- RESULT: Segar always said Popeye ate spinach for the vitamins. Modern sources have confirmed the connection prior to Sutton.
- CLAIM: “Spinach is actually a poor source of nutritional iron due to its oxalic acid content.“
- RESULT: Spinach is a good source of non-haem dietary iron. The bioavailability of iron from spinach is best assessed as a meal constituent, and is markedly increased by the presence of ascorbic and citric acids.
- CLAIM: “the decimal point error is a supermyth“
- RESULT: There is good evidence of various power-of-ten errors, each interpretable as a “decimal point error“.
The long version deals with each of these three parts in more detail:
1/ “E.C. Segar had Popeye eat Spinach for its Vitamin A“
Originally Popeye stroked the feathers of Bernice, an African Escape Hen a.k.a. W(h)iffle Bird, to give him superhuman strength. In June 1931, he first expressed a preference for spinach, to give him that strength,
Immediately there were plenty of allusions connecting Popeye’s strength, spinach and iron,
Vitamins were also immediately connected to Popeye eating spinach,
“Popeye’s Spinach Diet Leads to State Discovery: … “Popeye” has recently stated to his gallery that his extraordinary pugnacious ability is due to the fact that he “eats, spinach,” that lowly sand sprinkled herb, or vegetable which fond mothers force upon reluctant youngsters In the belief that It is good for the Inner works of the Intestinal tract. “The claim has been made by doctors, veterinarians, vegetarians, or somebody, that Spinach (with a capital “S”) contains vitamins PDQ or SOS, or something or other calculated to Improve the disposition and distemper of the human animal. And now this technical statement, heretofore laughed at by the doubting Thomases, has recently been fully endorsed by “Popeye” And “Popeye” must be believed, for his is no ordinary endorsement to be confused” (1931. The Bee from Danville, Virginia, July 31. p. 9).
When asked about the spinach connection, Segar confirmed vitamins were the reason,
Sutton admits to having carried out a time-consuming, “painstaking analysis” (Sutton 2010 p.13) of “four volumes of Segar’s comic strip cartoons” (Sutton 2010 p. 24), but to have found this connection with vitamins, Sutton reading cartoons was unnecessary. A single internet search shows Segar confirmed the role of vitamins; “The shame of it!” Sutton “Should have used Google!“
Using Google, Sutton would have seen Segar stated Popeye ate spinach for its vitamin content, and also that other modern sources, prior to Sutton (2010), stated Popeye ate spinach for vitamins, and especially Vitamin A. For example,
and,
Sutton cannot claim to have discovered Segar’s reason for Popeye eating spinach was for vitamins, nor to being the first modern source to reproduce that information. However, alongside the millions of Popeye fans and readers of the original strip and the modern-day republished collections, Sutton can claim to have to have seen the frame in the Segar cartoon strip mentioning Vitamin A, albeit within a modern collection published in 2007.
2/ “Spinach is actually a poor source of nutritional iron due to its oxalic acid content.”
A major breakthrough in the understanding of iron metabolism came at the start of the millennium, with discovery of the homeostatic hormone hepcidin (Krause et al. 2000), the principal factor in regulating iron plasma concentrations, and ferroportin (Abboud & Haile 2000), a transport protein that acts as hepcidin’s receptor and iron channel, both synthesised by the liver (e.g., Sharma et al. 2005, Ganz 2011); evolution has produced an elaborate and sophisticated mechanism to maintain optimal moderation of a extremely toxic element (see Sangkhae & Nemeth 2017, for a recent review of the current understanding). In brief, the small intestine (duodenum) is where iron is absorbed from digested food into the blood stream, however, these naked ions are harmful, so the body must go to extreme lengths to keep iron clothed. For example, hepcidin-moderated control mechanisms exist to not only limit consumed iron import across the intestinal lining, but also its release from white blood cell macrophages, where it is stored following recycling of senescent red blood cells, and from liver cells where excess iron is stored, and from where this entire metabolic homeostasis is regulated (Ramey et al. 2010).
The numbers are staggering, for example, the destructive turnover of a billion red blood cells daily (there are five billion in a millilitre of your blood, produced by your bone marrow, at a rate of one hundred billion each day), ready for reabsorption by macrophages after about one hundred and twenty days. But, perhaps the most incredible element in this absorption, recycling and storage of iron is the rôle played by ferritin, which comprises twenty-four polypeptide subunits, arranged to essentially form an eight nanometre diameter caged prison, where four-and-a-half thousand iron ions are incarcerated, within a spherical globular cage.
Sutton is under the impression that the oxalic acid in spinach inhibits iron uptake from spinach in human diets. A lot of work has been done in the long term to better understand two aspects in particular of oxalate, the salts of oxalic acid, pertinent to human physiology: whether the observed in vitro action of oxalate has any relevance to in vivo conditions for dietary factors, and what intervention is possible to alleviate oxalate contribution to gall stone formation.
While oxalate interferes with iron binding to transferrin, the protein vehicle that acts as the liver’s agent in controlling concentrations, via distribution from the duodenum, and redistribution from the macrophages to all other tissues (e.g., Halbrooks et al. 2004), there is little evidence of it acting to inhibit iron uptake (e.g., Bonsmann, Walczyk, Renggli, & Hurrell, 2008).
Our dietary iron occurs as haem (unoxidised state) and non-haem (oxidised) iron. Haem iron can be absorbed through the gut wall fairly easily thanks to the presence of dedicated molecular transporters that convey haem iron across cell membranes, to reach the blood system for transportation throughout the body. Non-haem iron is not in the correct molecular form to be transported and so needs conversion into the ferrous form to allow absorption (e.g., Young et al. 2018).
While specific iron content of spinach can be assessed through ash experiments and single constituent in vivo trials (using incubation, enzymes and acids, designed to simulate passage through the gut), this does not reveal the bioavailability of iron. Given that the gut is rarely going to be void of other foodstuffs and a neutral pH environment, it is important to assess that bioavailability of dietary iron in the relevant context, namely, as part of dietary intake. It is important therefore to assess whole meals, rather than individual foodstuffs.
The results directly contradict Sutton’s claim that “Spinach is actually a poor source of nutritional iron due to its oxalic acid content“,
The effect of acids on the bioavailability of iron had been show in experiments where acids have promoted production of ferritin iron-binding protein, for example, ascorbic acid and oxalic acid promote bioavailability of iron, whilst phosphorus is an inhibiter (e.g., Reddy & Malewar 1992). Indeed, supplemental ascorbic acid doubled spinach iron bioavailability (e.g., Rutzke et al. 2004), while a combination of ascorbic acid and citric acid has been shown to allow up to 70% solubility of dietary iron (e.g., Kojima et al. 1981)
Oranges are very good sources of ascorbic acid: 150 g can deliver 120% of a required daily amount (Pao & Fellers 2003). Therefore, with respect to the habits of people taking a drink sometime around eating, tea significantly reduces uptake of dietary iron, whereas orange juice maximises absorption (Kojima et al. 1981).
This is a direct contradiction of Sutton’s claim that “Current scientific knowledge does not support the USDA claim that taking vitamin C (ascorbic acid) in any form significantly increases our iron absorption from Popeye’s favourite fast food.” (Sutton 2011)
Clearly the context of people’s food preparation, and intake habits while eating, are relevant. Sotelo et al. (2010) measured iron content in a range of pulses and vegetables and meat common in Mexican diets, before and after cooking. Bioavailable iron content was then estimated by an in vitro treatment . It is worth noting several inconsistencies: most striking, they found cooking spinach had no significant effect on total and bioavailable non-haem content (-6%). This contradicts several studies before and since, albeit subject to a range of analyses (e.g., -32% Tisdall et al. 1937, -75% Kimura & Itokawa 1990, -60% Masrizal et al. 1997, -48% Yadav & Sehgal 2002, etc, also, -33% from cooking after frozen storage Korus et al. 2012). Nonetheless, this simple plot of Sotelo et al. (2010)‘s results, clearly shows the inherent structure in the data.
On comparing any haem and non-haem source, ideally, any member from each group could have been selected at random, with a similar outcome. Therefore, any direct test will tend to find difference, regardless of specific food types being tested. A maximum likelihood approach allows the appropriate amount of variation to be accounted for by the arbitrary fixed term for “group”, so that the remaining error can be structured in terms of the group members. Also, the group sizes don’t balance, with fewer heme samples than non-heme, but a mixed model of this type can account for that, without overly compromising the analysis. Nonetheless, it is a useful collection of data, not least because it illustrates comparison of haem and non-haem groups, immediately suggesting that membership should always be included as a fixed effect in analyses. Failing that, the degrees of freedom for interaction effects across groups will be artificially inflated, giving greater chance of false positives (type 1 errors).
The consequence of this is that haem and non-haem groups are incomparable with respect to iron bioavailability, because making comparisons fails to reflect the realistic choices facing individuals. A vegan or vegetarian diet disallows meat sources of readily absorbable haem iron. An omnivorous diet must be balanced and include vegetables.
As a source of non-haem iron, spinach is a good source. Washed down with orange juice, it’s even better.
3/ “the decimal point error is a supermyth“
There are already two exemplary, comprehensive analyses of this subject:
A series of articles on Joachim Dagg’s Weltmurksbude,
- 2015.
- Sources of the spinach-iron myth: Richardson’s (1848) limbo & Wolff (1871) with a PS on Moleschott (1859) and PPS on Brockhaus (1852). Jul. 30.
- Sources of the spinach-iron myth: Wolff (1871) & Bunge’s (1892) bungle. Jul. 31.
- Sources for the spinach-iron myth: Wolff’s (1871/1880) decimal error. Aug. 1.
- Sources of the spinach-iron myth: Serger 1906, Haensel 1909, Kobert 1914. Aug. 4.
- Sources of the spinach-iron myth: König (1920) confused. Aug. 5.
- Sources of the spinach-iron myth: König (1926) woozily. Aug. 5.
- Sources of the spinach-iron myth: Schup(h)an’s true name. Aug. 6.
- The first chemical analysis concerning the legend of spinach’s iron-richness. Aug. 28.
- 2016. The real decimal error that transmogrified into the spinach-iron-decimal-error myth. Nov. 15.
- 2017. Spinach-iron data transformations: Boussingault (1872) to Berg (1913). May 27.
- 2018. The deep roots of urban legends about the health benefits of spinach. Jan. 15.
and a recent paper,
- Michael Mielewczik & Janine Moll 2016. Spinach in Blunderland: How the myth that spinach is rich in iron became an urban academic legend. Annals of the History and Philosophy of Biology 21, 61-142.
Consequently there is not much to add, except compile a list Sutton’s mistakes. Sincere thanks to Joachim and Michael for helping with this section.
Initially, Sutton accused Hamblin of fraud,
This is a great shame as Hamblin had simply echoed Bender’s first mention of the “decimal point error“,
“One common belief, that spinach is good for you, appears to be due to experimental error since the belief predates 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 analyses 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 Ernährungslehre) by von Noorden and Saloman in 1920. In 1937 Professor Schupan analysed spinach for its iron content with α-α’-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.” (Bender 1972. The Wider Knowledge of Nutrition. Inaugural Lecture, Queen Elisabeth College. p. 11)
It is important to note that the first part of the highlighted sentence is a statement, “the figure to be one tenth of that reported by von Wolff“. The second part is a conjecture, “may well have grown from a misplaced decimal point“.
Hartog confirmed the “possible origin of this belief” with the mention of Wolff c.1870 (there are two compilations by Wolff that qualify, within reason, 1871 and 1880, presented as a single volume online) and “food analyses showing spinach to be exceptionally rich in iron“, that was later shown to be ten times typical measurements, hence why Bender suggested that this might be as a result of a “misplaced decimal point.“
There are several potential sources from which the iron content of spinach may have become artificially inflated by a factor-of-ten, through say, miscomprehension of experimental conditions; for example, figures are very different for fresh and dried samples, and vary according to growing conditions and growth stage (e.g., Ancuceanu et al. 2015). These various origins for spinach’s high iron content, including Wolff (1871, 1880) are discussed independently by Dagg, and Mielewczik & Moll.
In terms of Sutton’s claims, it is only necessary to focus solely on Bender’s mention of the iron content of spinach; it is not necessary to locate a source for high values other than in Wolff (1871, 1880) who reported values for dried spinach. For that reason, the focus will be on the iron content of dried spinach.
If it can be shown that Bender was correct in saying that Wolff’s data for iron content of dried spinach was ten-times that for equivalent measurements by the 1930s, then Bender’s reasoning was also valid to suggest a “decimal point error“. This, in turn, would exonerate Hamblin of Sutton’s accusation.
Two of the examples involving the iron content of dried spinach compiled and adjusted by Wolff (1871, 1880) are worth highlighting here. The outcome of each combination of factors is for the iron content of dried spinach appearing to be ten-times the typical range evaluated in subsequent nineteenth century experiments, and confirmed by modern measurements (e.g., in 100 g of dried spinach: 32.7 mg, Bunge 1892; 39.1 mg, Boussingault 1872; 37 mg, Sherman 1907; 53 mg, Sherman et al. 1934; 35.9 mg, Karmakar et al. 2013). Let us call these our “benchmark values”.
Example 1: Wolff (1871) compiled the iron content of spinach from measurements made by Saalmüller and Richardson, and presented them in a table, expressed as percentages of 100 parts of pure ash (p. 101),
The Reinasche value is the percentage pure ash in the dried spinach sample (what remains after burning off the 90.53% water content, mentioned in the footnote on p. 101, and deducting the percentages of “Sand und Kohle“, sand and coal, and “Kohlensäure“, carbonic acid, from the “Rohasche“, raw ash).
The iron content in the sample can be roughly estimated by multiplying the percentage of Fe2O3 in 100 parts of the Reinasche by the amount of that pure ash. Taking a rough estimate that assumes 100 parts of pure ash is equivalent to 100 g of pure ash sample, we have,
Saalmüller: 2.1% of 16.27% = 0.021 x 0.1627 x 100 = 0.34 g in 100 g = 340 mg in 100 g
Richardson: 4.6% of 16.70% = 0.046 x 0.1670 x 100 = 0.77 g in 100 g = 770 mg in 100 g
Whilst these are not directly comparable to other estimates, it should be immediately obvious that these figures differ from the “benchmark values” above (32.7 to 53 mg in 100 g) by a factor-of-ten.
Example 2: Wolff (1880) proceeded to take the averages for his collected data, and presented them in another table (p. 128), still expressed in the same units of the percentage of 100 parts of pure ash from the dried spinach sample,
Wolff (1880) added one final summary table for the group of vegetables that included spinach (p. 147). This table expressed the average iron content (p. 128) from the two studies (p. 101), as the the percentage content in 1000 parts of dried spinach sample. In other words, he carried out the following calculation, introduced in Example 1,
Average of two studies: 3.35% of 16.48% = 0.0335 x 0.1648 x 1000 = 5.52% in 1000 parts
Making the same assumptions as before, 5.52% in 1000 parts is the equivalent of 5.52 g in 1000 g, which is also 0.552 g in 100 g, or 552 mg in 100 g.
Whilst this simply confirms the arithmetic behind Wolff’s tables, in addition to the factor-of-ten difference to the “benchmark values”, it’s also worth noting for this table (p. 147) that, even though Wolff (1880) stated in the title for the section, several pages beforehand (p. 141), the values would be in terms of a 1000 parts of sample (“III. Mittlere Menge der Asche und Aschenbestandtheile in 1000 Gewichtstheilen der Trockensubstanz von land- und forstwirthschaftlich wichtigen Stoffen” / Average amount of ashes and ash constituents in 1000 parts by weight of the dry substance of important agricultural and forestry materials), it is important the table’s header still says the same as the previous table (p. 128), “In 100 Theilen der Reinasche” (in 100 parts of pure ash).
To make the numbers consistent across tables (the same magnitude, and displaying two decimal places), it would seem Wolff had to adjust the Reinasche amount from “16.48” to “164.8”, so that the value for dried spinach could appear as “5.52”.
However, for anyone assuming the tables to be equivalent, as informed by their headers, the appearance of “164.8” in the Reinasche column would create a discrepancy by a factor-of-ten, and so appear to be a “decimal point error“. We can see this by working back through the calculation,
Normal calculation: 5.52 / (100 x 0.1648) = 0.335 g = 33.5% in 100 parts of pure ash
Wolff’s adjustment: 5.52 / (100 x 1.648) = 0.0335 g = 3.35% in 100 parts of pure ash
Aside: Bender refers to Noorden & Salomon (1920. Handbuch der Ernährungslehre) about which Sutton (2012a) says, “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.”
Noorden & Salomon (1920) did include Wolff’s measurements on iron in meat and root vegetables, but the values presented for the iron content of spinach were obtained by Haensel (1908. Über den Eisen- und Phosphorgehalt unserer Vegetabilien. Biochem. Zeitschr. 16, pp. 10, 12). Considering that Haensel’s figures for dried spinach (450 mg and 440 mg, in 100 g) are in the range reported by Wolff (340 mg to 770 mg, in 100 g), Bender’s conflation of the two original sources is understandable.
Sutton acknowledged publication of Mielewczik & Moll‘s paper, but dismissed its contents, saying the authors haven’t read his updates, and specifically the correspondence he held with others in the comments section to his article,
Sutton’s blogpost correspondence was primarily about spelling mistakes and recalculation of iron values. There has been no revision of Sutton’s central claim that a “decimal point error” never existed; he simply shifted his target from Hamblin to Bender.
Consequently, Sutton’s claim that no such error occurred is nullified by the evidence, in particular that presented by Dagg, and Mielewczik & Moll. Indeed, a number of possible origins have been identified for the factor-of-ten error in values for the iron content of spinach, including the two direct examples in Wolff (1871, 1880), given above.
Bender’s speculation over the existence of a “decimal point error” therefore seems reasonable, even if the actual error only gave the impression of having been caused by a “misplaced decimal point“, but Bender’s statement was clearly only ever conjecture, “the fame of spinach may well have grown from a misplaced decimal point“.
Sutton has made many mistakes in his dealing of the iron content of spinach, most documented by Mielewczik & Moll,
- Sutton is wrong in his account of events,
- Sutton is wrong that Wolff mixed up fresh and dry weight,
- Sutton is wrong that Bunge was first to obtain lower measurements,
- Sutton is wrong in claiming there is no evidence of a “decimal point error” #1,
- Sutton is wrong in claiming there is no evidence of a “decimal point error” #2,
- Sutton is wrong in his historical narrative #1,
- Sutton is wrong in his historical narrative #2,
- Sutton is wrong in his conclusions,
- Sutton is wrong to claim that he bust myths. He has solved no puzzles, but instead, through his many mistakes, he has created and proliferated the myths that there was no “decimal point error” and that he was first to rediscover Segar’s incorporation of vitamins into the Popeye story,
A particularly galling element of Sutton’s revisionist manipulation of history, is his use of Sherman’s 1907 USDA report, Iron in Food and Its Functions in Nutrition (Bulletin 185), to claim that reasons had been identified, to account for high spinach iron content in early German measurements,
- “Sherman (1907) published the exact reasons why be believed that such data were flawed.” (Sutton 2012a)
- “Sherman’s (1907: 43; 53) explanation that poor science knowledge at the time was to blame.” (ibid.)
- “More than 100 years ago Sherman (1907: 43; 53) … explained in plain English the various ways that earlier biochemistry methods … exaggerated the iron content of foodstuffs:” (ibid.)
- “various reasons Sherman’s Bulletin 185 gave for early bad science in the early high determination of iron levels of various foodstuffs” (ibid.)
- “Sherman reveals … Wolff’s hugely exaggerated findings were due to unreliable methods, rather than any kind of simple decimal place transcription error.” (ibid.)
- “Sherman shows us that by 1907, American scientists were more accurate, knowledgeable and ahead of the Germans on the iron in spinach issue.” (ibid.)
- “earlier erroneously high measures—such as those made by von Wolff in 1871—were explained in the USA by Professor Sherman in 1907 as resulting from iron contamination from heating dishes and other bad science” (Sutton 2016. HealthWatch Newsletter 101. p. 7)
Sherman did comment that early ash-based measurements returned inflated values for methodological reasons. Dismissing Boussingault (1872), Sherman (1907, p. 8) considered improvements were only forthcoming with measurements from Bunge (Sherman 1907, p.43). Sherman (1907, p.53) identified Wolff’s (1871) compilation as one of the early erroneous sources but, contrary to Sutton’s claims, did not comment specifically about any of the causes for those errors.
Instead, the sources of error that Sutton (2012a) then goes on to list, and claim Sherman identified as the causes of those high values, were either solely referring to Sherman’s own laboratory procedures, or in relation to studies on meat and shellfish, and in vivo metabolic methods, and were unrelated to early German ash-based estimates of spinach iron content.
Sutton fabricated the fiction that Sherman (1907) identified methodological causes for high values in early measurements of spinach iron content. It suited his chauvinistic narrative that,
“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” (Sutton 2012a).
RESULT
The issue is not whether whether Popeye’s strength became associated with the iron in spinach. That is well known. Nor was there any question that a myth existed for spinach having an extraordinarily high iron content. Again, that was beyond doubt.
More central to Sutton’s claims are whether the Segar cartoons were the source of the association of iron with Popeye, and whether there was an error in the reporting of spinach iron content, equivalent to Bender’s supposition of a “decimal point error“.
- While there are overt references to iron in the earliest strip cartoons, Segar made the explicit connection with vitamins. Various modern sources have subsequently confirmed that connection with Popeye’s spinach eating and vitamins, before Sutton had published.
- Evidence from independent sources identify several possible origins for the high iron content in spinach. Some of these are interpretable as a “decimal point error“. Bender was correct in his conjecture, and Hamblin repeating the story was wrongly accused by Sutton.
- Compared with meat, all vegetables are inferior sources of dietary iron. However, when assessed realistically, in the context of eating choices and habits, and augmented with, say, a drink of orange juice, spinach is a good source of dietary iron.
Each of the above points directly refutes Sutton’s claims regarding the cultural history of spinach, the science of its nutritional properties, and the history of that science.
So what did Sutton contribute? He found a cartoon frame linking Vitamin A, but that was needless because Segar had said so in interview. All Sutton needed to do, was use Google. Ironic.
EVIDENCE & REFUTATION 