As you can see from my comments under this earlier blog entry, the data of Thomas Richardson (1848. Beiträge zur chemischen Kenntnis der Vegetabilien. Annalen der Chemie und Pharmacie LXVII Bd. 3.)* may well be this first source that has caused this widespread opinion. Soon after 1848, educational publications started to spread the myth. For example, an encyclopaedia published by Brockhaus (1852. Die Gegenwart. Eine encyklopädische Darstellung der neuesten Zeitgeschichte für alle Stände. Siebenter Band. Leipzig: F. A. Brockhaus) stated at page 172: "Weiße Rüben enthalten nur eine geringe Menge von Eisen, Spinat dagegen viel." [White turnips contain only a little iron, spinach however a lot. My translation.] One year later, Aaron Bernstein published a popularisation of scientific findings in a work called "Aus dem Reiche der Naturwissenschaft: ein Buch für Jedermann aus dem Volke" (Berlin: Franz Duncker, 1853). At pages 157-158, Bernstein praised spinach as an iron rich, organic alternative to medicine for pale children.
*[this publication hangs in a digital limbo, because it has been attached to the end of the preceding article by C. List (1848. Ueber das sogenannte Terpentinölhydrat. Justus Liebigs Annalen der Chemie 67(3): 362-376.]
Richardson gave the values of various chemical compounds of various vegetables as percent values in relation to the raw ash and as percent values in relation to the pure ash (raw ash minus carbonic acid, charcoal and sand). The iron content, in particular, was given as the percent values of "Phosphorsaures Eisenoxyd," which literally translates as phosphor-acidic iron oxide but chemically means iron(III) phosphate. Richardson's data sheet also provides the percent values of the ashes in relation to the fresh matter.
As I have argued elsewhere, it is false to calculate the portion of, say, iron compound in the fresh matter by simply multiplying the portion of the iron compound in relation to ash with the portion of ash in relation to fresh matter. The reason why this leads to false values is, because the ashes gain mass during combustion. In particular, gaseous oxygen combines with the burning matter and while some products of combustion end up gaseous themselves (e.g. carbon dioxide) others end up as solid ash (e.g. magnesium oxide). That is, the ashes gained mass through combustion that was not part of the fresh matter.
It is not anachronistic to call this a mistake, in relation to Richardson's time, because Phlogiston theory had been questioned in the second half of the 18th century already, and experiments had shown that metals gain mass during combustion. Hence Richardson's contemporaries and followers should have known that simply multiplying the portion of iron compound in ash times the portion of ash in fresh matter would yield false values for the portion of iron compound in the fresh matter.
Nevertheless, Bunge (1892) has committed this particular mistake in manipulating data from Wolff (1871) as shown here. Therefore it seems likely that others have also simply multiplied the percent values of Richardson's data and drawn false conclusions. Or, anyway, it seems interesting to reconstruct what conclusions contemporaries of Richardson might have drawn from such a data manipulation.
As you can see from the table below, spinach comes out second after radish herbage (values in scientific notation: 1,76E-03 means 1.76*10-3 equal to 176mg in 100g). Assuming that the herbage of radish was usually not eaten, however, spinach would be the edible item with the highest iron content in Richardson's data set. Hence Richardson (1848) may well be the first source from which the widespread opinion sprang that spinach was a good source for dietary iron in turn.
|Item||portion of iron phosphate in raw ash times portion of ash in fresh matter||portion of iron phosphate in pure ash times portion of ash in fresh mater|
|Ananas, ganze Frucht|
|Feige, ganze Frucht|
|Brocoli (Kohl), Herz||2,02E-04||2,14E-04|
|Kastanie, ganze Frucht||1,77E-04||1,93E-04|
|Erdbeere, ganze Frucht||3,69E-04||4,56E-04|
|(Kidney Beans) Bohnen||3,09E-04||3,56E-04|
|Pflaumen (greengages), ganze Frucht||1,80E-04||2,42E-04|
|Orleans-Pflaumen, Haut der Frucht||5,39E-04||6,63E-04|
|Orleans-Pflaumen, Fleisch derselben||9,50E-05||1,49E-04|
|Kirschen, ganze Frucht||1,47E-04||1,61E-04|
|ditto Stiel derselben||4,74E-04||5,57E-04|
|Birne, ganze Frucht||6,60E-05||8,00E-05|
P.S: The Genealogical World of Phylogenetic Networks has a new post with a phylonetwork illustrating the whole dataset of Richardson (1848).