Flong time, no see
How a paper mold transformed the growth of newspapers
You’ve heard of letterpress printing. You know about individual pieces of metal type stacked side by side by typesetters to make a line of type. Maybe you’ve even seen a film on YouTube of hot-metal composition, using Linotype or Monotype machines, where reusable molds cast fresh type from boiling lead for magazines and books.
You’ve probably also seen a lot of stock footage of newspaper presses in which paper moves at absurdly high speeds and ink is slammed onto sheets, which are cut, folded, stacked, and thrown into delivery vans.
But you’ve likely never heard of flong. Flong is an unsung hero that arose in the 1800s alongside important developments in high-speed presses, papermaking, and hot-metal typesetting. A flong is a paper mold that’s made by pressing its raw, flat form into metal type and images. With this mold in hand, printers used casting chambers to pour in metal and create its counterpart, the stereotype, a solid plate used for printing.
This missing link is rarely discussed because it wasn’t designed to last: flongs and stereotypes were largely used for perishable tasks, especially in newspaper and advertising printing, and then gotten rid of. Flongs were burned, and stereotypes melted down at the end of a shift or job and the plate metal reused. Little of either remains, in contrast to wood and metal type, letterpresses, and all the other paraphernalia of printing meant to endure for years or decades.
The flong/stereotype process was a critical part of a set of improvements that allowed printing to expand in scope: more words typeset, more pages printed, more copies produced. Newspapers in particular leaned on this combination of mold and plate to expand in every direction.
Flongs and stereotypes largely survive in printing and industrial museums. That’s where I’ve seen all of them in person, though you can find photos online, often with incomplete or misleading explanations. The process may remain in limited use — in some parts of the developing world, a few newspapers continue to use hot-metal type and letterpress printing.
Letterpress printing and hot-metal casting transformed from the 1970s from the heart of commercial printing into a much-smaller but vibrant craft in the last two decades, usually hand-in-hand with a historical component in keeping history alive.
The flong is gone, but its history provides insight into how newspapers grew.
You spin me right round, baby, like a cylinder press
What’s the utility of creating a mold when you’ve already typeset a page and it’s ready to go on press just as you’ve made it? The answer encompasses a lot of elements: cost, durability, speed, mechanics, and efficiency.
We have to travel back to the earliest days of printing, as with many aspects of the art, to Johannes Gutenberg. His truly great innovation, around 1450, was perfecting a way to produce consistent pieces of metal type repeatedly and quickly. This so-called movable type allowed the relatively rapid composition of any arbitrary word. (He also figured out the slightly alchemical formulation of lead, antimony, and tin that cooled instantly on casting, retained sharpness to keep fine details, and was hard enough for repeated printing.)
That paired with his invention of the printing press, an adaptation from wine presses and other screw-driven presses of the day. (Some earlier printing presses existed, used for block printing, typically an entire carved page or illustration.) Gutenberg also had to adjust existing inks to meet the needs of printing, and developed a process of wetting and preparing the typical paper of the day so that it would accept ink under pressure from type without smearing.
Gutenberg’s printing method was largely unchanged until the late 1700s, although it was refined in that period. For most of that time, a page was set and printed as follows: Compositors set type a letter at a time in one or more columns for a page. The type had to be carefully planed, or tapped down to become level, before the type was wedged securely in place. Type that stood higher than the plane of the page would gouge the paper, receive too strong an impression, or break. Type that was too low wouldn’t be inked or produce an impression at all.
It was then locked up into a frame, known as a chase. The combination of type, images, and chase are a forme. A forme was placed into the horizontal bed of a press. It would be inked, paper placed on top (often held inside a tympan with windows cut out for the rectangular areas that should receive ink), and rolled on a carriage under the platen. The platen was a heavy metal plate weighted and adjusted carefully so that a press operator could wit some effort pull a lever and drop the platen onto the paper.
Type planing, inking, paper thickness, and the press adjustments all had to meet with very tight tolerances to produce a well-printed piece. If you look at documents and books printed even as early as Gutenberg’s first works, it’s remarkable how beautifully they were made, given how much that could go wrong. In late 2017 at the St Bride Printing Library, I saw pages in a book printed in 1478 by William Caxton, England’s first printer. It’s gorgeous. A modern printer would be jealous, and Caxton had made everything from scratch that he used to print, including the type. The quality arose partly because these books had to resemble the fine work handwritten by scribes.
Gutenberg’s operation could produce at least a few hundred pages a day, though some estimates place the number higher. By the late 1700s, speeds had climbed to hundreds of impressions an hour, although it often involved two impressions to print a single side of a sheet with multiple pages on it.
It took until when the industrial revolution was underway for dramatic changes to appear, and quickly. Around 1800, a process in France for making paper not a sheet at a time but in continuous rolls, called the Fourdrinier machine, was perfected. By the early 1800s, metal type manufacture had shifted from a cottage industry into foundries that began to offer centralized industrial scale. Up until the end of the 18th century, presses remained mostly wood with metal parts, but the all-iron Stanhope press in 1800 — the iron hand press — doubled the surface area that could be printed at one time while shedding effort by 90%.
Newspapers had began to emerge as their own medium in the 1600s, and started to spread widely in the 1700s. The Stanhope press allowed production of larger editions in the early 1800s, which reduced the price, which led to even higher demand. Higher demand then naturally fed the need for issues with more pages and more copies of each issue to feed a growing readership with diverse social, political, and financial interests.
To speed up printing even further for these kinds of fast turn-around, high-volume runs, press tinkerers shifted from human power to steam and later electricity. New presses relied on cylinders around which type was fixed, with paper passed between the rotating cylinder and a flat surface. Paper could be now fed into the press—from several inlets at once!—improving efficiency.
But recall that up to this point, all type was locked up flat on the horizontal bed of a press. How did printers make the transition from a flat forme to a curved one? It seemed like you’d need to make substantial changes to get this to work reliably or type would go flying everywhere.
Sadly, that was the case. There was no magic solution in 1832, when Richard Hoe released the first cylinder press. Arthur Winter said in a rather understated way in his industrial tome, Stereotyping and Electrotyping (1948):
Efforts made from the middle of the last century to lock movable type on the outside of rotating cylinders was only partially successful although many presses were made on this principle.
More directly, Charles Henry Cochrane related in The Wonders of Modern Mechanism (1904) that:
On a big central cylinder were wedged the pages of type, secured by ingenious devices in “turtles” to prevent the letters from being flung out by the centrifugal force of rotation. With the best of them some types were always sure to work out, marring the print.
These turtles were wedge-shaped rules, the strips of lead used to space lines of type (hence the term leading). By making them wedge-shaped, the type could be ostensibly locked into place. But in fact, as The Encyclopædia Britannica noted rather frankly in an article in its 1896 American edition, referring first to a press designed in 1837 by David Napier and then generalizing:
Napier’s press had a tendency to throw the type out, as did indeed all the presses up to his time.
Cylinder presses, which still required manually feeding paper and which printed on one side only gave way to more sophisticated rotary presses that could pass paper through from end to end. Not long after their introduction, rotary presses could print from a continuous roll of paper on two sides and cut it after printing with an automatic knife into sheets, and fold it, too.
From 1,000 impressions an hour with the first steam-powered cylindrical presses, rates of production shot up to tens of thousands of impressions across the first generations of rotary presses fed by rolls of paper. By 1887, Harper’s Magazine reported that the New York World newspaper could produce a 28-page Sunday edition of 250,000 copies. (See also an afternote on typesetting.)
With standard type locked around a cylinder, these faster presses would just have broken down more frequently and potentially caused more damage to equipment — and people.
But what allowed that huge jump in speed didn’t arise from prayers that type remained fixed in place instead of becoming high-velocity pointy projectiles. Rather, what you needed was something that could be attached to a cylinder as a single piece: a metal plate.
And that’s where the flong and the stereotype come in. A flexible paper mold allowed casting a plate that conformed to the needs of the rotary press.
Far-flong ideas finally come to fruition
The idea of stereotypes didn’t arise just because newfangled presses required curved plates. Rather, printers had been tinkering for centuries on a way to create molds and reusable plates — flat ones befitting contemporary presses. The term stereotype means “solid” or “durable” type, and was coined long after people tried to develop them.
Paper flongs came late in the process. Instead, most efforts before around 1830 involved plaster of paris, sand casting, or other materials that could set hard to create the mold. These often involved adapting existing processes used in professions like statue making and goldsmithing. Some inventors’ techniques showed promise, like William Ged of Scotland’s in the 1700s. But he was sabotaged by printers, compositors, and type foundries for decades to keep him from creating a method that would save labor!
Even the best of these molding methods lacked fine detail, and thus couldn’t reproduce even type (much less images) with enough fidelity in comparison with other printing of the time. They often were barely legible, based on surviving samples.
The Didot family of France — famous in the printing world across centuries — developed and commercialized a far better process that relied on setting type using individual type molds. A stereotype could be made directly from these typeset matrixes. This method required an extraordinary amount of effort, and was extremely difficult to make corrections once plates were first made. But the firm claimed it made thousands of these a month under contract from 1810.
It wasn’t until around 1829 that a wet flong method emerged; the year is known due to a patent issued that year. However, it took decades longer to develop into a reliable process.
A wet flong is effectively papier mâché (“mashed paper”), combining a paste and several layers of tissue and blotter paper. The type forme has to be prepared so that it’s clean and has a layer of oil or other substance that allows the flong to be lifted off when dry without breaking or flaking. The flong can be forced—literally beaten—into all the crevices in type and images. Once knocked into shape, it had to be fully dried before it could be removed. The result was a flexible and resilient paper mold. (I can find no illustrations or photographs of the wet process; books on the topic describe making flongs, but don’t illustrate what the result should look like.)
Since I know you’re wondering, flong is how the English pronounced flan, the name given in France ostensibly because the material’s wet form resembled a custard. A flong may also be called a mat, short for matrix, a term used in printing for any kind of mold used for casting.
Scientific American described in 1873 what was still a relatively new process:
This consists in beating into the face of the type, with a heavy brush, a prepared sheet, with a body almost like paper pulp, and somewhat thicker than heavy railroad card. The type form, with this wet blanket kind of mold beaten into it, is then placed on a steam bed to drive out the moisture and harden the mold, which, in a few minutes, can be taken off almost as hard as a sheet of card board, but holding a perfect impression of the type.
Flong was ahead of its time. Paper mills had to produce softer papers that could take the impression better, while printing boffins had to work out a way to produce curved stereotypes from the flexible flongs. (Say “flexible flongs” five times fast.) By the 1860s, all the pieces were in place.
Beating flong with a flong brush — I just enjoy that sequence of words, too— later transitioned into using a rolling press backed rubber blankets and heavy boards that provided a faster and more even impression. (A lot of this technology co-existed, so you would find some printers relying on older hand techniques where they couldn’t justify the capital cost versus labor, while others were cutting edge, especially where high volume was required.)
A flong was then trimmed, built up with backing materials in low areas that would otherwise sag and fill with metal, and put into a casting machine. A stereotype emerged. The casting device for rotary presses was curved to the exact needs of the press, and a flong mounted along that curve.
This plate would then be — according to Stereotyping, the Papier Mache Process (1892) — sawed, shaven, trimmed, routed, beveled, and finished. (This also sounds like how men had their facial hair manipulated at the time.) Each plate had to conform perfectly to the press, but all these steps had specialized machines and printing-plant workers who knew how to employ them quickly and expertly.
Flongs could be made from type along, as well as—by themselves or locked into formes—wood cuts, copper halftone blocks (etched plates that substitute different sized dots for shades of gray or color), “zincos” (etched zinc plates), “lino” (linoleum) cuts, and plastic originals, something that had emerged itself as a duplication technique.
At the time that the flong/stereotype process took over, both type and typesetting remained expensive. Stereotypes were thus used in quantity not just by newspapers, magazines, and book publishers, but also in other areas of printing. It kept type from wearing down, and allowed plates to be cast, printed from, and retained for reprinting without tying up stores of type. Even as the Linotype made type effectively cheaper and faster to produce, and Monotype later brought the same savings (and speed) to book publishing and finer work, stereotypes retained their value for the speed of printing.
And flongs were durable enough that multiple casts could be taken from a single form. The process didn’t just transform a difficult medium into an easier one — pieces and lines of type and figures blocks into a plate — but it allowed for duplication. A publisher could use multiple presses at the same time to print the same pages without setting the copy twice. Or produce multiple flongs and distribute those to cast plates elsewhere. (In fact, flongs could be made from stereotypes, producing new stereotypes, though with the same degradation of quality as with multi-generation photocopies.)
(There’s another kind of duplication technology called electrotyping, which relies on a wax mold put into a chemical bath suffused with electricity to create a hard metal shell later backed with type metal that’s a high-quality duplicate of the original engraving, type, or other material. This technique takes hours, and is extremely fussy, making it inappropriate for newspaper work, but could be used for magazines. Electrotyping was also widely used to make duplicates of engravings of ads that would be sent to newspapers for printing via flong and stereotype!)
You can see a problem, however, with making flongs: excessive heat used in drying could surely damage the type and images from which molds were made. And that was certainly the case. In the Journal of the Royal Society of Arts (1890), a transcribed lecture by Thomas Bolas notes:
The question of damage to type during the process of stereotyping is one of some importance, and it mainly steps in when a high temperature is employed for drying. If the forme is very tightly locked up in the chase it may, in expanding and softening under the heat, become elongated while on the other hand it may become shortened by the pressure of the drying press. These two circumstances tend to make a newspaper fount become of unequal height, and the fount is rendered useless.
Again, this was less problematic once the Linotype arose for many newspapers, since the type didn’t retain any value after the day’s newspaper was printed, but not all papers had converted by the end of the century. larger headlines and other parts of newspapers remained set by hand. Flongs were also made for job work (the customary stuff a printing office would produce for business, like letterhead and pamphlets) and in book publishing, where type was more precious or might need to be corrected and new flongs made.
Fortunately, help was on the way in the form (not forme) of dry flong.
It’s not the heat, it’s the humidor
Rather than use a complicated process of multiple layers of paper and paste, dry flong was spongy in its raw form. It was often “made from wood pulp or from a mixture of rag and chemical pulp or repulped waste papers containing these fibers,” according to Stereotyping and Electrotyping. It was kept damp (but not wet)in a humidor before it was placed under pressure in content with a forme or components to be duplicated.
It didn’t require extensive preparation, heating, or drying before it could be removed, thus dramatically speeding up production. Several contemporary accounts talk of shifting from tens of minutes to a few minutes.
The exact year of its introduction is hard to tell from the historical record. Stereotyping and Electrotyping claimed in 1948 that someone in Germany and someone in England simultaneously came up with a product in 1893.
It took dry flong a few years to catch on in newspapers despite its advantages, because it required new suppliers and new expertise. But once it took hold after the turn of the century, it became the only mold used for newspapers of any scale. (Wet flong apparently continued to be used at smaller papers and in job shops due to cost.)
In my research and purchases, I’ve found that most of the dry flong and stereotypes that remain were used in advertising and for letterpress clip art. National advertisers, such as movie studios and appliance makers, would send the flong to their customers or to newspapers. The newspapers would set copy for the local address or business name, and leave space in the forme to cut in the ad mats or clip art mats after the mold was made for the rest of the page.
This likely sounds familiar to anyone in the photographic era of print production — it’s like paste-up, only for making relief plates instead of the flat metal plates etched by light exposure for lithographic offset printing. Even down to the clip art, which was routinely employed in paste-up days.
Cartoon syndicates also distributed comics via flongs. This answers the question I’ve long had about how strips were syndicated in the relief-printing era. In later years at least, syndicates would distribute a sheet that contained the six weekday strips (Monday to Saturday) and then the separate color plates for the larger Sunday edition.
(For those versed in print production, dry flongs also explain the limitations on halftones. Newspaper halftones never exceeded 85 lines per inch in the letterpress era, and I find references to 60 lpi in the first part of the 20th century. That’s to do with the fidelity of flong, not limitations in halftone etching. Newsprint also required a coarser screen, but that wasn’t the sole determinant.)
Even as dry mats took over newspapers, companies that made them — like the Wood Flong Corporation — tried to expand their market to more common commercial printing, especially contrasting their product against electrotyping. In an ad in the trade journal The Inland Printer in 1922, the company said in bold type, “The Electrotype Is Doomed,” and said electrotyping cost 10 times as much and took 20 times as long. Both may have been true, but quality was the typical difference. The ad noted, “Under proper conditions the dry mat stereotype plate will print just as well and save much money and time.” That’s a bit of a hedge.
A few pages earlier in the same issue, the International Association of Electrotypers posed the rhetorical question, “…ask yourself what would happen to the printing industry if the whole load of modern letter-press printing were suddenly thrown upon any process or method which aims to compete in any sense of the word with the real electrotype.” It concluded, “Just so long as the result includes quality, no process of duplication thus far invented should be considered.”
Even by 1948, the electrotype still won out on quality. Ben Dalgin, the New York Times production chief, wrote in his book that year, Advertising Production, that “mats” (flongs) were used especially during WWII due to shortages in electrotype metal. “The results we were able to get from them were far better than we ever expected.” However, “we feel that we get the best impressions when electrotypes have been used,” partly because flongs were made inconsistently. (He also noted some newspapers wouldn’t accept either a mat or a stereotype for advertisements at all.)
The need for speed in newspapers remained paramount, however. Dalgin wrote that dry flongs were made in a few minutes, prepped quickly, and seven stereotypes cast every two minutes. He noted that as many as 30 plates might be made from a single matrix to feed that many presses.
For the final pages of the day—page 1 and inside pages that contain the continuation of front-page stories—Dalgin wrote that it took a remarkable 15 minutes from those pages being locked up to pass through molding, platemaking, and printing, until “printed, collated, folded, and cut” newspapers came off the press! Placed on waiting trucks, 15 minutes later those papers arrived at a train terminal for distribution.
Fling the flong
There are very few flongs or stereotypes left in the world. When flongs and stereotypes were in heavy use, there was little reason to retain them, and they weren’t precious.
Dalgin in Advertising Production noted that the Times cast 90,000 pounds of metal nightly for daily editions and 300,000 pounds every Saturday night for the Sunday paper. It was all melted down.
In late 2017, Sue Shaw, the founder of The Type Archive in London, said to me while discussing newspaper history, “They burned all the flong.” (The St Bride Printing Library over in Fleet Street has a couple of very-late-era newspaper flongs and stereotypes in their letterpress shop, however.)
Full-page newspaper flongs appear on auction sites from time to time. However, because the ads and clip art I mentioned above were prepared on a large scale and only a portion was ever cut into newspaper molds and used, more of it survived than news pages. I originally thought there’s little available, but people post new pieces on eBay constantly. And some printers and collectors have sent me photos of stacks of clip-art flongs wrapped in their original packaging.
I mentioned above that syndicates distributed comic strips as flongs, and it seems that very little of that still exists, either, as they were almost always assembled onto pages and then discarded. You can find a historical conversation here and there. However, I managed to purchase a sheet of Doonesbury flong from May 1973 that survived in uncut form. The reason? It’s from the week that John Ehrlichman resigned from the Nixon administration, and the strips had to be pulled because they were no longer sensible. The syndicate sent a new flong. (Trudeau reworked two of the six strips, but I believe the other four never appeared.)
Stereotypes are in far shorter supply. Occasionally you’ll find a stereotype for an ad that was intended for flat-bed reproduction, so it would be locked into a forme, or a flong would be made from it and then inserted into a newspaper. Recently, I’ve spotted a couple of stereotypes on eBay of newspaper front pages from the day after JFK’s assassination. A current auction for the plate of page 1 of the Dallas Morning News has a starting bid over $10,000. I can only suspect that printing plant workers or executives saved these stereotypes for sentimental — or collectible reasons.
Flongs also show up in strange places. At least a couple of houses were insulated with it, because it was thick wood pulp. A Spokane TV station ran a story in 2008 showing new owners of a home uncovering siding to find the entire house covered with flongs, and a Reddit user found them lining their attic. You can find interesting examples, too, like this wall decoration. In a discussion on Twitter among rare and early-book curators, I learned of a 1904 book bound in dry flong. I was able to acquire an interesting and high-quality example of Fairy Tales Up To Now, which retells old stories in modern dress in verse and typeset as if for a newspaper.
Flong, but not forgotten
Why devote 5,000 words to flong and stereotypes? Almost entirely because they were forgotten. They require a scale of operation no longer needed for letterpress along with an astonishing amount of specialized equipment and training to use them.
As a bridge technology dedicated to speeding the production of the printed word in depth and reach, they were invaluable—and seemingly inevitable. Without them, huge constraints would have been placed on burgeoning commerce by the limited speed of printing.
The Linotype is no longer in production, but remains firmly fixed in memory, and dozens—maybe hundreds—of Linotypes remain in some form of operation. Rotary presses and roll-fed paper dominate printing. Craft and pedagogical letterpress (and some commercial) is charming and once again widespread.
Flongs and stereotypes, left out of the party, deserve a memorial for their place in history. They received Viking funerals while they lived set afire nightly, and now that they’re gone, this electronic remembrance of their one-time greatness.
Glenn Fleishman writes about technology past and present from his home in Seattle. In 2017, he was the School of Visual Concepts first artist in residence, printing a letterpress book of his type and language, Not To Put Too Fine a Point on It. In early 2018, he published the book London Kerning about living typographic history in that city. He’s currently making the Tiny Type Museum & Time Capsule, a collection of printing and type artifacts in a custom case that trace the history of printing and put it in your hand.
This is a work in progress relying on sources stretching over more than two centuries. I welcome expert feedback and all questions!
An afternote on typesetting
As a former/recovering typesetter, I couldn’t let the transition from handset type to Linotype go without a brief mention in the context of newspaper efficiency and production.
While stereotypes increased the speed and reliability of printing, typesetting held back the quantity of news that could appear in each issue. Until the late 1880s, when the Linotype hit newspapers, typesetting remained intensely manual.
Newspapers simply couldn’t hire enough people or purchase and maintain enough type to set larger editions. The 1929 U.S. Bureau of Labor Statistics report, Productivity of Labor in Newspaper Printing, noted:
Before the introduction of the linotype it required 16 compositors for approximately seven hours to set sufficient type for four pages of a representative newspaper at that time. Distribution of the type required about one-half that number for the same length of time, while other hands necessary in composing-room work would probably bring the personnel on a 4-page daily newspaper to about 40.
The Linotype machine let an operator set type five to ten times faster than the typical hand compositor. It took up far less space, allowing for more composition, too. And distribution, or putting away each character of type in the cubbyhole in a type case or drawer that it came from was unnecessary. It was simply melted down after use.
That increase in speed coupled with efficiency allowed for bigger editions:
In one typical establishment, for example, the average daily issues consisted of 12 pages in 1896, 24 pages in 1916, and 36 pages in 1926. The Sunday issues contained an average of 48 pages in 1896, of 54 pages in 1916, and of 60 pages in 1926.
The report indicated that there was an initial massive displacement of compositors, about 60% to 80% fewer people engaged in that part of operations. But the growth in newspaper scale increased demand and revenue, and “in a comparatively short time more compositors were employed than formerly.”
- Bureau of Labor Statistics, U.S. Department of Labor, Productivity of Labor in Newspaper Printing (Government Printing Office, 1929)
- The Encyclopædia Britannica with American Revisions and Additions (1896, page 1276)
- Harper’s Magazine (July 1887, page 176)
- The Inland Printer (June 1922)
- Journal of the Royal Society of Arts (August 22, 1890, page 844)
- Scientific American (May 31, 1873, page 337)
- Albert Sidney Bolles, Industrial History of the United States (the Henry Bill Publishing Company, 1878)
- Charles Henry Cochrane, The Wonders of Modern Mechanism (J.B. Lippincott Company, 1899)
- Ben Dalgin, Advertising Production (McGraw-Hill Book Company, 1948)
- George Kubler, A New History of Stereotyping (self-published by his firm, the Certified Dry Mat Corporation, 1941)
- Charles Summer Partridge, Stereotyping, the Papier Mache Process (A.N. Kellogg Newspaper Co., 1892)
- Frederick Wilson, Stereotyping & Electrotyping (E. Menken, 1898)
- Arthur Winter, Stereotyping and Electrotyping (Sir Isaac Pitman & Sons, Ltd., 1948)