Eggs and Invisible Ink: George and Giovanni

By Sean Coughlin

In a 2015 episode of Turn, a US Revolutionary War TV drama on AMC, George Washington’s spy Abraham Woodhull uses a special ink made with alum to write secret messages under the shells of hard-boiled eggs. The technique was also advertised on the show’s Twitter in 2014, a year before the episode aired (Figure 1).

Figure 1: Teaser tweet before the airing of the episode. Image via AMC Twitter.

Turn is based on a 2006 book by Alexander Rose called Washington’s Spies: The Story of America’s First Spy Ring, but there is no mention in the book of any such technique. Instead, it seems to come from a 2009 book by John A. Nagy, Invisible Ink: Spycraft of the American Revolution. That book describes an ink that is able to permeate the shell of a hard-boiled egg, leaving the message hidden inside and no trace of writing on the shell. It is not attributed to George Washington and his spies, however, but to Giambattista della Porta. Nagy writes:

“In the fifteenth century Italian scientist Giovanni Porta described how to conceal a message in a hard-boiled egg. An ink is made with an ounce of alum and a pint of vinegar. This special penetrating ink is then used to write on the hard-boiled egg shell. The solution penetrates the shell leaving no visible trace and is deposited on the surfaced of the hardened egg. When the shell is removed, the message can be read.” (John A. Nagy, Invisible Ink: Spycraft of the American Revolution, 2009: 7)

Nagy’s account of della Porta’s recipe seems in turn to have come from 1999 book by Simon Singh called The Code Book: The Science of Secrecy from Ancient Egypt to Quantum Cryptography. Singh, however, leaves it unclear whether the egg is to be hard-boiled or raw when one writes on it:

“In the fifteenth century, the Italian scientist Giovanni Porta described how to conceal a message within a hard-boiled egg by making an ink from a mixture of one ounce of alum and a pint of vinegar, and then using it to write on the shell. The solution penetrates the porous shell, and leaves a message on the surface of the hardened egg albumen, which can be read only when the shell is removed.” (Simon Singh, The Code Book: The Science of Secrecy from Ancient Egypt to Quantum Cryptography, 1999: 10)

Singh’s description of della Porta’s recipe only appears in the first edition of The Code Book. In later editions it is missing without comment. The reason might be found by reading della Porta himself, who in Book 16 chapter 4 of his Natural Magic mentions the technique, but neither claims to have invented it nor to have gotten it to work:

“Africanus teaches thus: ‘grind [oak] galls and alum with vinegar, until they have the viscosity of ink. With it, inscribe whatever you want on the egg and once the writing has been dried by the sun, place the egg in sharp brine, and having dried it, cook it, peel, and you will find the inscription.’ I put it in vinegar and nothing happened, unless by ‘brine’, he meant sharp lye, what’s normally called capitellum” (della Porta, Magia Naturalis 16.4: Latin 1590, English 1658)

Della Porta attributes the recipe to Africanus, probably Sextus Julius Africanus, a 2nd–3rd century CE traveler, writer and chronicler, whose recipe for an egg-permeating ink is preserved in the 10th century Greek compilation known as the Geoponica, of which della Porta’s recipe is a literal Latin translation. Unlike Singh’s or Nagy’s recipe, della Porta’s includes oak gall (an ingredient often found in inks as a pigment) and lacks precise measurements.

The measurements and techniques described by Singh seem similar to a recipe printed on page 143 of the 1973 New Earth Catalogue (Figures 2 and 3):

Figure 2: The New Earth Catalogue: Living Here and Now, ed. Scott French and Gnu Publishing, New York: Putnam Berkley Press, 1973. Image via MareMagnum.

Writing Under Shell of an Egg. Dissolve one ounce of alum in a half pint of vinegar and use a small brush to paint whatever you wish to appear on the shell of an egg. After he egg has dried completely, boil it for 15 minutes [...].
Figure 3: P. 143 (detail, via Google Books Snippet View) of The New Earth Catalog: Living Here and Now.
This recipe uses half the amount of vinegar. It also mentions that a small brush is to be used and says to boil the egg for 15 minutes. None of these details are in Singh’s or in della Porta’s version.

This recipe shares a strong family resemblance to a one published by the USDA in 1965, whose purpose was to get children to eat more eggs. We know this because the USDA’s suggestion was picked up by the New York Times and published on page 14 of the 29 May 1965 issue (Figure 4):

Mess on an Egg is Tempting to Child. To increase a child's consumption of eggs, us "invisible writing," the United States Department of Agriculture suggests. Secret messages can be painted on the outer shell of eggs before they are hard-cooked, and nothing will be visible until the egg is cooked and the shell removed. The message will be on the hard-cooked egg white. To make the "magic ink," dissolve one ounce of alum (available in drugstores) in one cup of vinegar. Use a small pointed brush to write on the shell. Let it dry thoroughly and then cook the eggs in simmering water for about 15 minutes. Cool quickly.
Figure 4: P. 14 (detail) of 29 May 1965 New York Times with an invisible ink recipe attributed to the USDA.

Sometime between 10 and 20 years earlier, either in 1946 or 1959, a similar version similarly targeted to children was published in volume 14 of Richards Topical Encyclopedia, an encyclopedia ordered by theme that was sold door to door in the US (Figure 5).

The Warning Beneath the Egg Shell. Mix an ounce of alum with a half pint of vinegar. Then with a fine brush, using the mixture as an ink, write a message--a joke, a prophecy, anything you like--on the shell of an egg. After the egg is boiled in water for about fifteen minutes, the writing will disappear, but your unsuspecting friend who removes the shell will find the message on the hard-boiled egg inside.
Figure 5: P. 136 (detail) of volume 14 of Richards Topical Encyclopedia, published either in 1946 or 1959.

Another 10 (or 20) years before that, we find the recipe on page 58 and 132 of the February 1936 issue of American Druggist. A reader from Oregon asks for the recipe of a solution used to mark eggs. The editor replies that the “trick” is described ‘in Henley’s “Book of Recipes”.’ (Figure 6)

Egg Shell Marking. "Oregon" used to mark eggs with a mixture of brown sugar syrup and some acid. He has forgotten the formula. The trick, as described in Henley's "Book of Recipes" is as follows: (Continued on page 132) Notes and Queries...continued from page 58. Dissolve an ounce of alum in 8 ounces of vinegar and use the solution to "write" upon the egg, using a small pointed camel's hair brush. Dry the egg and boil it for about 15 minutes. By that time the markings on the shell will have disappeared but when the shell is removed the writing will appear on the hardboiled white of the egg.
Figure 6: P. 58 and 132 (detail) of February 1936 issue of American Druggist.

The editor’s recipe is not obviously the one “Oregon” was after (it includes neither sugar nor acid) but the details are familiar: 1 oz alum, 8 oz (= 1 cup = 1/2 pint) vinegar, a small, pointed brush, 15 minutes boiling. Some details are new: we’re told the brush should be camel’s hair (I will come back to this in the sequel).

The editor of American Druggist also gives a source: Henley’s “Book of Recipes”. The Norman W. Henley Publishing company was active in the United States in the early 20th century and beginning in 1907 published almost yearly editions of an extremely popular book of household recipes: Henley’s Twentieth Century Book of Recipes, Formulas and Processes: Containing Nearly Ten Thousand Selected Scientific, Chemical, Technical and Household Recipes, Formulas and Processes for Use in the Laboratory, the Office, the Workshop and in the Home By Gardner D. Hiscox.

I checked the 1907 edition of Henley’s, but found nothing. I continued looking for references to the recipe closer in time to the American Druggist issue to refine the search. I found two earlier instances: page 97 of the January 1930 issue of Popular Science, and (slightly earlier) page 110 of the September 1929 issue of Field and Stream, where it is referred to as ‘an old hex trick’ (Figures 7 and 8):

Making Writing Appear on Whites of Boiled Eggs. An easy and effective trick is to letter a prophecy on the shell of an egg, using a mixture of an ounce of alum in a half pint of vinegar as the medium and applying it with a fine brush. Place the egg in water and boil for about fifteen minutes. The lettering on the shell will disappear, but on removing the shell, the prophecy will be seen on the hard-boiled white of the egg.
Figure 7: P. 97 (detail) of January 1930 issue of Popular Science.
1001 Outdoor Questions. By Iroquois Dahl. Ques. Not long ago I saw the egg of a domestic hen boiled, and when cracked and opened, the following words and figures appeared, perfectly printed on the white of the egg: “Ware will come in 1932.” As I am a disbeliever in hokum and bunk, I would appreciate information on how this was done? Ans. This sounds like an old hex trick. You can write on the inside of an egg by dissolving 1 ounce of alum in ½ pint of vinegar. With a small pointed brush, outline whatever you writing you desire on the shell of the egg with this solution. After the writing has dried thoroughly, boil the egg for 15 minutes. All trace of writing should disappear from the shell, and when the egg is cracked and shelled, the writing will appear on the hard-boiled white.
Figure 8: 1985 reprint of p. 110 of September 1929 issue of Field and Stream.

If Henley’s was the ultimate source, the recipe had to have appeared sometime before 1929. I checked the 1925 edition, but it was nowhere to be found. The next edition was published in 1929, the same years as the Field and Stream recipe. There it was on page 786 (Figures 9 and 10):

Wood Polishes; Wood Renovators; Wood, Securing Metals To; Wood, Waterproofing; Wood’s Metal; Wool Fat; Worm Powder for Stock; Writing, Restoring Faded […].
Figure 9: P. 786 (detail) of Henley’s, the 1925 edition. No recipe.
Wood Polishes; Writing Under the Shell of An Egg: Dissolve one ounce of alum in a half pint of vinegar with a small pointed brush outline whatever writing you desire on the shell of the egg with the above solution. After the solution has dried thoroughly on the egg, boil it for about 15 minutes. If these directions are carried out all tracings of the writing will have disappeared from the outside of the shell--but when the shell is cracked open the writing will plainly show on the white of the egg.
Figure 10: P. 786 (detail) of Henley’s, the 1929 edition, with the earliest version the author could find of this recipe.

The recipe contains many of the characteristics of the one that has been passed down in American lore and attributed in one form or another to George Washington’s spies or Giambattista della Porta: 1 oz. alum, ½ pint vinegar, a small brush, 15 minutes of boiling.

Given the reach of Henley’s “Book of Recipes” in the US, this is not surprising. But Henley’s recipe also lacks a key ingredient from the recipe that della Porta attributed to Africanus: oak gall. How and when this ingredient dropped out from a recipe reliably passed down for over a thousand years is another story.

This post continues a study of how a 3rd century recipe for a magic ink, despite the fact that it probably never worked, still managed to work its way into American popular culture in the 20th and 21st centuries. An earlier part of the study is posted here.


Apicius’ Pumpkins with Turkey

By Sean Coughlin

This post follows on yesterday’s post on whether the Romans had pumpkins.


[Cooked] gourds with fowl: [add] hard-fleshed peaches, truffles, pepper, caraway, cumin, silphium, green herbs – mint, celery, coriander, pennyroyal, mentuccia –,  honey, wine, liquamen, oil and vinegar. (The Latin text is here).

Stewing turkey, pumpkin and apples in the wine and stock. Photo by the author.

My interpretation:


  • 2 tbsps. olive oil
  • 2 large turkey drumsticks (about 750 g)
  • Salt and pepper
  • Half a sugar pumpkin or butternut squash, cubed
  • One apple, peeled, cored and diced
  • 1/4 tsp. asafoetida
  • 1/2 tsp. caraway
  • 1/2 tsp. cumin
  • Silphium (1/2 tsp. fennel seeds will do)

For the sauce

  • 1/2 cup dry white wine
  • 1/2 cup chicken or vegetable stock
  • 1 tbsp. honey
  • dash of fish-sauce or MSG
  • Fresh herbs to taste (mint, celery, coriander, pennyroyal, oregano)
  • Truffles, shaved (to taste)

1. Season the turkey legs with salt and pepper. Heat the oil in deep pan over high heat. Add turkey legs and cook, skin-side down, until crispy and golden brown (8 minutes or so). Flip legs and cook until the other side is browned, another 5 minutes. Transfer to a plate.

2. Return pan to heat with a bit more oil, and when it’s hot, add the pumpkin or squash and asafoetida. Cook until the pumpkin just begins brown, about 5 minutes. Add the apple, caraway, cumin and fennel seeds and cook for another few minutes until browned.

3. Add wine to the pan and reduce by half. Add stock, honey, fish-sauce and herbs. Return the drumsticks to the pan, cover and let simmer for an hour until drumsticks are fall-apart tender. Add water or more stock if it begins to look too dry. Alternatively, place in an oven preheated to 375 F / 190 C. Serve drumsticks with roasted sweet potatoes, drizzle on some of the sauce with some shaved truffle.

Deglazing with white wine. Photo by the author.

I’ve recently become obsesessed with cucurbits thanks to a question from Peter Singer. This resulted in a discussion with Laurence Totelin that took place this summer during a workshop at the Humboldt-Universität as part of SFB 980, project A03, “The Transfer of Medical Episteme in the ‘Encyclopaedic’ Compilations of Late Antiquity”, with Philip van der Eijk. The subject was our forthcoming translations of Books I and II of Aetius of Amida’s Medical Collections. “We” are Sean Coughlin, Eric Gowling, Christine Salazar, and Piero Tassinari. Many thanks to our guests Alessia Guardasole, Matteo Martelli, and of course Laurence for attending. The translation of Book I was completed by Eric Gowling as a doctoral dissertation and was in the process of being revised by Piero Tassinari and myself when he passed away last year. I hope Piero would appreciate this little essay. He is dearly missed.

Thanksgiving with Galen and Apicius

By Sean Coughlin

For Thanksgiving, I thought I’d come up with a new English translation of a seasonal recipe from the Roman cook-book of Apicius. It comes from the third book of De re coquinaria. The Latin is cucurbitas cum gallina. In Joseph Vehling’s English translation: “Pumpkin and Chicken”.

If only the Romans had turkeys. And pumpkins.

I first learned that the Greeks and Romans were pumpkinless from Laurence Totelin earlier this year and it left me utterly confounded.

I had been revising a translation and commentary of the first book of the sixth-century CE physician Aetius of Amida’s Medical Collections, and a question came up about how to identify different kinds of cucurbits, that immense botanical family of vines which produce comically-large-berries, like cucumbers, melons, watermelons, pumpkins, squashes, gourds, and luffa sponges.

Naively, I thought it would help to look at how authors like Aetius or Galen say cucurbits are prepared. And, since Galen (On the Properties of Foodstuffs, 2.3) tells us that people always ate their cucurbits cooked, I figured that Greco-Roman cucurbits must be something like zucchini.  After all, who on earth would cook a melon?

“But,” Laurence told me, “zucchinis are new world.”

Now, before going into why I found this so shocking, I want to bring everyone up to speed.

“Old world” cucurbits include a lot of different genera: Cucumis, the cucumbers and melons; Citrullus, the watermelons and colocynths; Lageneria, the bottle gourd or calabash; and a few others like Luffa and Bryonia. No one is quite sure where in the old world they first evolved, but the recent consensus is somewhere in Africa. In other words, they could have existed in the Greco-Roman world.

Various old world cucurbits: cucumbers and melons. Photo from the author.

“New world” cucurbits, however, are the vines descended from plants native to the Americas. They almost certainly could not have been known in antiquity, and the group includes all species of the genus Cucurbita. This includes zucchini, courgettes, pumpkins, squashes, decorative gourds, and even those 1000 kg monsters at fall fairs. None of them were known in Europe before Columbus.

I would like to imagine that there is at least one other kindred reader of classics in English who feels a bit anxious at this news. It’s not just a problem for Apicius’ Thanksgiving casserole or my rehearsal of Epicrates’ lampoon of Plato’s Academy.

What about the “Pumpkin Pirates” of Lucian’s True History (2.37)? Or Psyche’s sister’s husband who Apuleius says is “bald as a pumpkin” (Apuleius, Metamorphoses 5.9)? And what will become of Seneca’s Apocolocyntosis divi Claudii – the Pumpkinification of the Divine Claudius?

English translations are unusually problematic in the case of cucurbits. Pumpkins show up all over the place. And this isn’t merely because of translators’ license, or a confusion of names.

As far as I’ve been able to sort out, it’s mainly the result of an unfortunate coincidence between the rise of an influential but erroneous botanical hypothesis, and the publication of two influential Greek and Latin English dictionaries.

Here is the story:

Before the mid-19th century, the evidence about the origins of Cucurbita was mixed. The names of the plants suggested they came from the “old world”, but depictions of Cucurbita species don’t show up before the mid-1500s.

For instance, in Leonhart Fuchs’ 1542 herbal, De Historia Stirpium, we get one of the first European depictions of the pumpkin plant, the species Cucurbita pepo L. Fuchs, however, calls it cucumis turcicus, the Turkish cucumber, a name which might lead one to think the plant was native to the “old world”.

Not long after Fuchs, in an herbal of Matthiolus from 1586, the same plant appears again, but with a different name: cucurbita indica – the Indian gourd. And “Indian,” Matthiolus tells us, means American: “they say these [new gourds] came into Italy from the West Indies, whence they are called by many ‘Indian’” (Matthiolus of Padua, Commentary on Dioscorides 1559, p. 292, my emphasis).

“He has a pumpkin” (ipse cucurbita habet). Using pumpkins as camouflage. Woodcut, late 16th century, Netherlands. CCBY

Both Fuchs and Matthiolus agree that these cucurbits are foreign (and probably even recent) imports to European markets. They disagree about where they came from: were they indigenous to the old or the new world?

This continued until the 1850s, when the hypothesis that Cucurbita species were indigenous to the old world reached its most articulated form. It happened as part of a public debate between two great 19th century botanists: Alphonse de Candolle and Asa Gray.

De Candolle argued that because people had used the same names for cucurbits since ancient times, the plants with those names must have come from the old world.

In response, Gray argued that new and ‘foreign’ cucurbits only start showing up in 16th century herbals like those of Fuchs and Matthiolus. That, along with documentary evidence from the first European explorers, suggests a new world origin.

The debate continued for over thirty years, only coming to an end in 1883 after a devastating two-part review of De Candolle’s last book, The Origin of Cultivated Plants. In this review, Gray, along his colleague J. Hammond Turnball, presented almost documentary evidence from European explorers that cucurbits were cultivated in the Americas before European explorers ever reached it in 1492.

These explorers recorded indigenous names for the plants they encountered, but they also used European ones, and these are the names that stuck: in Latin, cucurbitae and cucumeres; in Spanish, calabaças, melones, pepinos and cogombros; in French, melons, concombres, citrouilles, and courges; in Italian, zucca; and in English, musk-melons, cowcumbers, and, — most importantly — the pompion, the source of our word, ‘pumpkin’. Only the name “squash” from the Algonquin vernacular made its way into English.

Gray and Turnball’s review, however, appeared only after the work on the great Greek and Latin lexica had finished up – A Greek-English Lexicon by Liddel and Scott, and A Latin Dictionary by Lewis and Short. Both dictionaries identify some Greco-Roman cucurbits with American pumpkins (Cucurbita pepo L. or Cucurbita maxima L.), and pumpkins have shown up in English translations ever since.

So for this recipe, I figure, why fight it?

Tune in tomorrow for Apicius’ recipe!

The author with pumpkins, sometime in the late 2000s. Photo from the author.

Sean Coughlin is Research Fellow in the history of philosophy, science and medicine at Excellence Cluster Topoi and the Institute for Classical Philology, Humboldt-Universität zu Berlin. He is completing a revised translation and commentary on Aetius of Amida, Medical Collections, Book I, an edition and translation fragments of the 1st century Pneumatist physician, Athenaeus of Attalia, and co-editing a volume on the concept of pneuma after Aristotle.