Faking It


The December 19, 2013 issue of The New Yorker included an interesting article by Nicholas Schmidle about forging antiquarian books. The focus of the story was a copy of Galileo's Sidereus Nuncius (the 1610 first edition) that appeared on the market in 2005 and included what were thought to be original watercolor illustrations (vs the engravings in the published edition) by the author. If the book was right, and the illustrations could be proven to be by Galileo, the copy's value would catapult from six figures well into the sevens.

The copy was subjected to extensive forensic and academic scrutiny by scholars "using such tools as long-wave ultraviolet radiation (to identify inks) and X-ray fluorescence (to determine the paper's composition." It was known that about 30 proof copies had been sent to Galileo, all lacking the engraved illustrations. The hypothesis, which the experts ultimately endorsed, was that this was a proof copy that had been embellished with original watercolors that were then copied by the engraver. (This claim that it was a proof subsequently allowed the experts to dismiss inconsistencies with authentic copies.) With these bona fides in place, the bookseller who owned the copy planned to offer it for $10 million.

The copy had come to the bookseller from an Italian dealer-cum-impresario named Marino Massimo De Caro. It won't ruin Schmidle's article to report that this copy of the Sidereus Nuncius was a forgery conceived and orchestrated by De Caro, and that its eventual discovery contributed to his conviction earlier this year of embezzlement (selling books he stole from libraries).

Getting to the point: reading the article, it sounds as if the only person who didn't look at the Sidereus Nuncius was someone who knows anything about printing. Obviously that cannot be the case, but some of the details leave one wondering. The answer probably is that the technical details were glossed over by the magazine - after all, it's not a printing journal - and this post is the equivalent of complaining about the logic of time travel in a Star Trek episode. Nonetheless...

Schmidle succeeded in getting De Caro to discuss his forgery, which was achieved in Buenos Aires. De Caro had visited papermills in Italy over the years and studied how artisans "recreated seventeenth-century watermarks and other vintage elements. He took this knowledge to Buenos Aires" and worked with a local artisan to make paper that looked authentic (i.e. 17th century).

Stop right there: making paper that can pass inspection by some of the world's leading booksellers and scholars isn't something you do easily. Making a good sheet of paper, and then repeating that sheet a few hundred times, is not a skill one picks up quickly.

Reproduction of 16th century engraving printed at HM from polymer plate for The WunderCabinet (2011)

Carrying on: De Caro scanned a copy of the Sidereus, "cleaned" the scanned letterforms, and made polymer plates of each page. To replicate ink of the era, he found bottles of 19th-century India (i.e. writing) ink at an antique shop.

Stop again: writing ink and printing ink are very different. One's runny & one's sticky.

After inking a plate, De Caro "misted the paper with a spray bottle to prevent cracked paper and bleeding ink. ('That was one of the secrets," he said.) They lined up a plate on a page and applied pressure; after a few seconds, they removed the plate. The process left impressions slightly deeper than those of seventeenth-century letterpress, but De Caro believed the effect was accurate enough to pass off onto an unsuspecting client."

Couple problems there, but this could be where the editors were going for brevity in a part of the story that won't really matter to the majority of readers. One doesn't "mist" paper to prevent "cracking." If cracking was a concern, the paper they were making would never be confused for the lovely linen paper being made in 17th-century Italy. This matters because the feel of the paper tells you as much as what's printed on it.

The description of printing sounds like they did it one page at a time, with the back of a spoon. That cannot have been the case: there's no way it would have fooled anyone.

The issue of impression is an interesting one. Unless there was dramatically more, or an almost complete absence of impression, that probably wouldn't raise any flags. A bigger concern would be a noticeable difference in the quality of inking from copy to copy: that's a much harder factor to copy and maintain than impression.

One final nit to pick: a scholar identified another of De Caro's forgeries - or possibly an adulteration - in which the date on the title page had been "inked over," adjusting the date from 1649 (third edition, $) to 1640 (second edition, $$$$$). That should probably have read "over inked:" a strategic blob of ink, as if from poor inking, on the 9 could make it look like a 0.

For our immediate purposes, here's the particularly interesting aspect of the article: the use of photopolymer to reproduce metal type. De Caro's play was ultimately unravelled thanks to the work of a scholar from Georgia State named Nick Wilding. He wasn't included in that initial brain trust that certified the Sidereus (and then took great offense when Wilding began to question their conclusions). He seems to be the first person in the story who knows how to look at letterpress printing:

"In 1610, when a printer was inking his lead type, ink occasionally seeped into crevices along the edges of the type*; during printing, this caused splotches and faint lines to appear along the 'shoulders' of the page. Such marks lacked the depth of letterpress type. But, if someone used a high-resolution image of the original page to create photopolymer plates, the polymer would make the black marks equal in depth to the text."
* ink doesn't seep. It's too stiff - HM

This is a technical issue with any kind of letterpress printing. Here's how Rummonds explains it in Printing on the Iron Handpress (p. 277):

"Even with type-high roller bearers, the roller will have a tendency to bump on and off the type blocks as it moves across the type, causing the first and last letters in each line [i.e. the 'shoulders' mentioned in the above quote] to be over-inked. A smoother transition can be gotten by putting small pieces of masking tape with feathered edges on the bearers, aligning the first piece of tape with the first and last letters of the lines of type..."

Ink should be applied only to the face of a piece of type. Over-inking causes some ink to adhere to the edge, or "neck" of the type. (The above image is taken from A Short History of the Printed Word, W. Chappell & R. Bringhurst, Hartley & Marks, p. 56.) As shown above, the neck slopes out and down from the face to the shoulder; it's not perpendicular to the face. This is why some ink can squeeze (or 'seep') down from an over-inked roller (or ink ball). The photo below help illustrate the point: look along the right shoulder, and particularly the bottom right corner. It's not the most egregious example on HM's shelves, but it's the closest to hand (a 1940s reprint of Poe's essay on anastatic printing).

When the type is impressed to a sheet, all of the ink is transferred, resulting in a letterform that is thicker than it should be, like the difference between a regular and bold face. But the inked shoulders would not leave impression as deep as the face. Here's a photo of a particularly grungy piece of type (not HM's!) to further illustrate the point:


If you were printing from the original text block, and took care, this over-inking of the shoulders could be eliminated. If, however, you were reproducing in polymer a page that had been over-inked on the shoulders, the polymer would reproduce the source page exactly as it was printed, over-inking and all. (And if you aren't careful in printing the polymer, you could amplify this same over-inking.)

A related issue with all platen-type presses, whether they be of the common or handpress style or more modern clamshell, is that the shoulders of a form receive greater force (impression) from the platen. Even if inking is consistent across the form, this will result in the edge letters printing more heavily. Here's an example from a 1711 book on Roman coins; note the visible impression from the other side, especially along the outer shoulder.

This is easily fixed with the makeready and, on a handpress, platen bearers.

Wilding understood how to look at the back (i.e. verso) of a letterpress page, and distinguish impression caused by the face of a type verses the (lesser) impression caused by the neck; and how this would be different when looking at a polymer reproduction of an over-inked original.

In actuality, however, comparing & discerning these variations in impression would be difficult, and variations from one copy to another could legitimately be the result of the way printing was done in the 17th century. (To this day that's the trick of using a handpress; minimizing variation from impression to impression where everything is controlled entirely by the printer's hand and eye.) The kind of paper and the method of printing (damp vs dry) have much greater influence on the fidelity of a type's reproduction. Dampened paper requires significantly less ink and gentler impression than dry, and demonstrably gives the sharpest result. As Schmidle's article details, there were much more obvious signs of forgery in the Sidereus that should not have been so quickly dismissed.

Which brings us to this: as characterized by De Caro in the article, the forging comes off sounding like a lark. But without even trying to commit deception, making paper that could pass for what was being made in 17th century Italy; printing on it in a manner that would pass for authentic; and then binding the pages in a style that is makes it look 400 years old, all take a small team of highly skilled craftspeople. There's no way the people who were fooled, starting with the bookseller who purchased the forgery, would have been easily fooled. Morals and ethics aside, this book was a technical accomplishment any printer would be proud of. It seems that, like many forgers, De Caro's ultimate undoing was at least partially the result of his own pride in the work.

Geek-out completed. Track down Schmidle's article, it's an engaging read with interesting insights to the rare book world and the economics of forging. If you want to read a great novel about forging antiquarian books, check out Arturo Perez-Reverte's The Club Dumas (1993). But don't watch the film adaptation. And don't bother with John Dunning's novel about forging, The Bookman's Wake, which suggests an implausible degree of ignorance about how printing is actually done from an author who'd once been a bookseller.