Archives for category: Book manufacturing

Zoe Cokeliss Barsley of Oxford University Press starts off her blog post entitled “Is publishing sustainable?” with an account of how environmentally expensive book manufacturing is. She overstates her case of course, understandably, given that her job title is Director of Sustainability at the Press. When I first saw the title of the post I thought we were in for an essay on whether publishing could survive economic and technological conditions — which I believe (and hope) it will. But of course it’s not that kind of sustainability: I obviously hadn’t allowed for OUP’s new electronic chops.

There’s no doubt that books do take a toll on the environment — which I’d claim is most significantly incurred in trucking the damn things hither and yon, often in the end to a place where they’ll be incinerated as unsaleable. But we’re already taking action on that. It may be true to claim that “Wood pulp, for making paper, board for book covers, and packaging, comes from forests in countries ranging from Brazil to Finland. Commercial forestry for timber and pulp contributes to deforestation and forest degradation, a leading driver of biodiversity loss that also accounts for around a tenth of global greenhouse gas emissions.” True, but almost irrelevant. If you took all books out of the equation the difference would be unnoticeable: book paper represents a tiny proportion of printing papers, which in turn is a small proportion of overall paper manufacture. Still, as she says, “Publishers can reduce their impact by sourcing certified sustainable paper”, and they do. Most books are already being printed, and have been for most of this century, on paper from sustainably managed forests.

It is of course only to be expected that advocates of digital over print should minimize the environmental costs of electricity, computer manufacturing etc., though in fairness I have to note that Ms Barsley does acknowledge this cost. Whatever the detail of the calculations, OUP is set on a course to achieve by 2025 carbon neutrality from their entire operation, 100% certified sustainable paper usage, and an elimination of any waste to landfill. Obviously if you don’t print any books, this is easier to attain, but the increased use of digital printing also makes controlling your inventory a lot easier. Even when I was there a significant proportion of OUP’s print books were already designated “MD” — which means in their system “manufactured on demand”, i.e. manufactured one-off only after a customer has ordered the book. For an academic publisher this route is truly a no-brainer. Warehouses are expensive, especially for books which only sell in dribs and drabs.

On November the 8th The Bookseller informed us that Cambridge University Press & Assessment* have appointed their first global director for climate education. The Scholarly Kitchen brings us a more general report on what actions academic publishers are taking. That we can do so little in terms of the overall problem should not of course be taken as an excuse for doing less than we can.

See also Environmental sustainability.


* This is the awkward name the Press has adopted since they were smooshed together with the Local Examinations Syndicate — the folks who create all those exam papers. Can’t help unease: Press is an object, Assessment an activity. Better would have been Cambridge University Press Publishing and Assessment, maybe with a colon or a —. As my friend David Tranah points out, a major gain from this would have been the resulting acronym: a cuppa is what we all need. Cupa sounds more like a sort of Amazonian rodent.

EBITDA stands for Earnings Before Interest, Taxes, Depreciation, and Amortization. It is a measure of the amount of operating cash in the business. QuickBooks tells you how to figure it, and why you’d want to. As they indicate, “Those who use the EBITDA formula prefer to analyze a company’s performance based on day-to-day business operations. They disregard debt (interest costs), taxes, depreciation, and amortization.” Printing Impressions has an introduction and the application of EBITDA in print acquisitions. Quite a lot of book publishers now use EBITDA. (My ex-boss, a fan, used to shout at me “Don’t say ‘a lot’. What’s the number?” Sorry, I know not.

To quote the exhibition Building the Book from the Ancient World to the Present Day at The Grolier Club “A sheet folded in half once creates a folio of two leaves; a quarto, folded twice, results in four leaves. Other formats include octavos, duodecimos, sextodecimos (16mos), tricesimo-secundos (32mos), and sexagesimo-quartos (64mos)—the smallest known being a 128mo printed by the Plantins in the 16th century.”

Folio, quarto, octavo, duodecimo

Follow the links on the right side of the exhibition website linked to above.

The actual dimensions of each of these will depend on the sheet size you start with. The relationships will remain the same. Some sizes may be found here.

In a post from 5 June 2020 I wrote

In so far as they think about it, I imagine that most book-readers believe that when a publisher decides to print 5,000 books, 5,000 books is what they’ll receive. Wrong, wrong, wrong! Indeed one might almost say that of all the possible quantities which will be checked in at the warehouse, 5,000 is probably the least likely!

This is not because publishers and book manufacturers are really bad at counting: it’s inherent to the manufacturing process. If a car manufacturer sets out to make 5,000 cars, that’s what they’ll make. They set up 5,000 bits of metal on the conveyor belt, and stop adding more bits once they’ve reached 5,000. (All right; I dare say it’s a little more complicated than that.) A book is made differently. It’s made as a collection of different parts which end up being brought together and united at the very end of the process. And it’s made of rather flimsy materials any of which can easily get damaged at any point in the process. I’ll use as my example the now rather old-fashioned procedure of a book printed on a sheet-fed offset perfecting press, bound as a sewn hardback with a jacket. (Books printed in other ways have to go through the same steps but they are often made with a different sequence and significance.) For my example I’m assuming a 5% over/5% under allowance on final delivery.

Lets assume our book has 320 pages, and is imposed to deliver 32-page signatures. Each of these ten signatures/sections will be printed separately, or possibly as two-up for five separate passes through the press. If there are two sigs on each sheet, the sheets will then have to be cut in half after they have been printed. Once you have your ten stacks of flat sheets each containing one signature, you send them off to the folding department where they will be folded down to ten little booklet-like sections. Let’s say there’s also an 8-page 4-color insert, which has been printed at a different plant and delivered already folded to the bindery. Now each of these eleven sections (10 for the text, and one insert) will be placed in sequence in a series of pockets on a gathering machine, which takes one from each pocket to make a book’s-worth of eleven sections, arranged we all hope in the right sequence without any duplications or omissions. Prior to gathering the first and last sigs will have had the endpapers tipped on. Now that you have lots of little piles of sigs, each making up an 11-sig book, they will go into the sewing department where they’ll be joined up into untrimmed book blocks. Cases are made from board and cloth, then stamped with foil. Thereafter the book blocks and cases go to the binding line where the book blocks are squeezed, gripped and trimmed, get their spine glued, and a liner applied, then get glue spread down the sides and have the case cover dropped and pressed onto the pages before they roll off the end of the machine. Here they receive a rapid examination as the jacket is applied, and the books travel on on a conveyor belt to where they are packed into cartons meeting the weight requirements of the warehouse to which they will be traveling.

At every one of these steps there is a risk of damage and loss, so you have to figure out what the loss risk might be at every stage and aim for a quantity at step one which will get you to the finish line with at least 5,000 copies of the book. Printing presses are big pieces of equipment and don’t stop and start on a dime. Makeready is required for every impression: you have to get the paper path perfect and adjust the ink/water balance to ensure an even black impression across the whole sheet. You’ll know from past experience how much paper you’ll be likely to need for makeready on this press: so you start off aiming for, let’s say 5,550 copies. (This is just a notional number used for this example. I don’t have access to any book manufacturer’s spoilage allowance data — and in any case it will vary from plant to plant, and press to press, maybe even from crew to crew.) You don’t want to make the spoilage allowance too big, because you have to pay for the paper you use, but you don’t want to make it too small, or you may be forced into a little (and killingly expensive) reprint of one sig to make up the count. Additionally there are accidents which can occur on press and have to be allowed for: a lump of dust flies onto the plate, or even a fly, and gets squashed onto the blanket where it starts printing. This we call a hickey — and as soon as they notice it the press minders will start pulling pages out of the press and throwing them away, while one of them cleans up the blanket and on we run guessing at how many impressions we need to add at the end of the run to keep our count up. You don’t stop the press if you can avoid it: if you do, there’s that makeready to do all over again.

So here you are, off press with between 5,250 and 5,650 copies of each of ten sigs. Sod’s law will determine that you actually have about 5,500 of eight sigs and 5,250 of two of them. You may loose a few sheets to makeready on the folder: it’s got to get the fold in at perfect right angles to the edge of the sheet: start with the sig with the highest count! On their way to the gathering department one of the skids of folded sheets may get bashed by a passing forklift truck, or waiting overnight, they may be dripped on when the roof leaks in a thunder storm — hey, accidents do happen. At each step you’ll lose a few more overall: fingers crossed the major loss will not be on the sigs you started out with 5,250 on.

The manufacturer will be aiming to deliver the maximum number of books permissible under the purchase order. After all the more they ship the more they’ll be able to bill: but there are limits. Most orders allow for a 20% slop, which tends to resolve into 10% over and 10% under, but more powerful customers will negotiate that down to 5% each way. In this example, worked on 5% over/under, the book manufacturer will be striving to be able to ship and bill 5,250 copies — and will be tearing hair out if the count drops below 4,750. The same consideration has been motivating the jacket printer, and the insert printer. They will aim to over-deliver against the ordered quantity, but not by so much that their materials costs will cause them a loss on the job. Publishers often like to receive extra jackets for refurbishment in the warehouse, so running out of jackets at the bindery line shouldn’t be an issue, though like everything else it has been known to happen. If you go into the gathering line with one or two sigs totaling only 5,075, you know you’re not going to be able to overdeliver too many, or any at all, and you also know that all those extra sigs from the other parts of the book are just wasted time and paper. They’ll be sent for recycling after the job’s completed. Storing partials is no longer an economically viable aim.

Book manufacturers will negotiate adjustment to the overage allowance. What I mean is they may accept an order for not-more-than, in which case the overage allowance of all goes to unders none to overs. In other words, in our example, the 5,000 will be priced at rates which would apply on a 4,500 run multiplied up. If the publisher ordered 5,000 exactly, the job would be priced as if it were a 5,500 run. In any case, if the manufacturer comes off with too many copies, they will be on the phone asking the publisher if they’ll accept overs over and above the overage allowance. If the book has subscribed well (i.e. if bookstores have ordered better than expected) the publisher may well decide to accept extra copies, so that the time when they’ll need to order a reprint is moved out a little. Tough negotiators will offer to accept these overs at no charge: I’d have agreed to paid for them at run-on cost; but I’m just a softie when it comes to supplier relations. Unlike many I regard the relationship as a partnership: if this sounds a bit old-fashioned, I fear it is.

The advent of digital printing, plus the forces of economics, which to some extent may be held responsible for the development of digital printing, have reduced the significance of overage. At the extreme end, true print-on-demand, overs aren’t an issue: the customer orders one copy, one copy is produced. If you’re printing a couple of hundred copies though, some of the pitfall-points listed above do come into play. Obviously damage can occur at random, but the crisis points are fewer, and it’s easier to get close the the ordered quantity on a digital press run than with other processes. Of course you may still have the spoilage allowances on the folding and binding lines to negotiate.

This is good news: less spoilage is good for us all, ecologically and economically, and being able to more accurately control your inventory makes publishing more profitable, or at least removes one hindrance to profitability.

Paper is sold by weight. In the days when most of our printing was done sheetfed, you had to be able to get to a number of pounds from the required number of sheets. This you could calculate in two steps: firstly by working out the number of sheets needed for each copy — if it’s printing 32 to view, i.e. 64 pages per sheet, your 256-page book will require 4 sheets per copy. Multiply that by the number of copies you plan to print, and add a spoilage allowance, which, if you didn’t already know it from experience, was something the printer could be coaxed into giving you, but would always be a “best guess”. With these numbers — Bob’s your uncle — a number of sheets. How much does each sheet weight though? This information is contained in the basis weight designation for the paper: if it’s 50# basis that means that a ream of this paper, if your sheet were to measure 25″ x 38″, would weigh 50 pounds. But of course your book is almost certainly not going to print on a sheet as small as 25″ x 38″ — this is just the standard sheet size used when we talk about book paper basis weight. These sorts of quaint old-fashioned things are there because they were first thought of eons ago. Who talks about reams now? But because it all works nobody’s changed it. (Metrication in Europe has greatly simplified this sort of thing.) So

M weight = Length of sheet x Width of sheet x Basis weight ÷ 950/2.

The number 950 is the product of 25 x 38 (the standard sheet size area). You’re calculating the area of the sheet you need to buy and dividing it by the area of the standard sheet which links you to weight. The 2 comes in because a ream is half of a thousand — 500 sheets. The answer to this equation, the weight of a thousand (M) sheets, is then multiplied by the number of thousands of sheets you need: and now you have it in pounds.

Probably I’ve wasted too much space on “the largest”, “the smallest” etc., but here’s another contender, one which makes the explicit claim on its publisher’s website: “It is literally the largest book in the world by far and can be admired as the materialization of an ecosystem saturated by media. ONEPIECE exists as an object of pure speculation.” I call the claim explicit, but apart from the first few words, the claim here is more than confusing and far from explicit. The price of the “book” is €1,900 (plus shipping) though the printer/publisher, JBE Editions informs us it’s sold out.

The book’s dimensions are 12 x 18.5 x 80 centimeters (4.7″ x 7.2″ x 31″) and weighs 17 kilograms (about 38 lbs.). A 31″ spine width does make for awkward reading even if you were up for 21,540 pages of classic manga.

One Piece is a long-running Japanese manga. At its immense article Wikipedia tells us “One Piece set the Guinness World Record for ‘the most copies published for the same comic book series by a single author’,” Eiichiro Oda. JBE’s edition was of fifty copies only. The Siliconera post referenced by my original source, Shelf Awareness for Readers of 30 September, maintains JBE printed the monster, which means they’d have to have cleared copyright permission with the original publisher. Easier would have been to buy 50 copies (if they’re all in print). As it is the only manga that had an initial print run of above 3 million continuously for more than 10 years, this could well be less unlikely than one might assume.

Those who care will already know, and those who don’t know may not care: but the difference between manga and anime is as follows. Manga is a comic book form, while anime is the video version. No doubt this is an oversimplification: if you need more detail, check out Fictionhorizon.

In the olden days when you went to a lithographic print works, off to the side there would be this black metal revolving door, contained within a sort of huge tightly-filling barrel. You’d squeeze into the little space provided by the door and tiptoe around pushing the revolving door till it span you out into a wonderland filled with red light and tightly excluding all the white light from outside. In the middle sat a gigantic camera. As far as photographic film is concerned red = black, so here, in the darkroom, film could be handled freely in the open. (If you did this on the other side of the door all your film would instantly become exposed.) I wonder what effect life in a red-light environment might have on a person working there. Back in the days when photography was still analog, many keen photographers had their own darkrooms at home where they’d splash around with chemicals and photographic papers. I knew a few, but never detected any red-light craziness.

Gutenberg Museum Fribourg – Historic Repro Camera
Gutenberg Museum Fribourg – Historic Repro Camera

Here are a couple of photos of the repro camera in the Gutenberg Museum in Fribourg, Switzerland. In the lower photo you can see the copy, clamped vertically before the camera lens. Above you see the camera which will shoot it in order to create a negative which can be stripped up to make a flat. The focal length (enlargement) can be adjusted by moving the lens closer or further. Once things had been cranked into position, the arc lamps (white light) would flash exposing the film in those areas where there was something to be seen in the original — like words. Develop and fix the film and there’s your negative of a page of text, which can now be looked at in white light.

If the original copy wasn’t the whole page, but just an illustration, after the camera had shot it, someone would have to cut the neg into a negative of the type page which had had a blank space left where the picture should go. X-acto knives and red adhesive tape were the tools of the trade for a stripper.

Here’s a view of another camera showing the whole set up:

Photo Deutsches Kameramuseum

To shoot a photograph a gridded screen would be placed between the artwork and the lens, and this would break the picture down into a series of dots. A process camera would also add color filters to drop out all colors except the four needed for four-color process printing: cyan (blue), magenta (red), yellow and black. After the camera operator had done his work the resulting film might be corrected by a color retoucher working principally with a tiny paintbrush. By the late seventies and eighties process cameras were becoming a bit of a rarity as color separation migrated to color scanners, and eventually the computer.

Color separation arranges the dots so they appear in a rosette pattern — so that they don’t fall one on top of the other blanking colors out. This effort begins at the camera with the different halftone screens being placed at different angles.

From Pocket Pal

My copy of the Library of America edition of Henry James: Novels 1903-1911 is printed cross-grain. It’s the second printing, and I bet the first printing (like all their other books) is printed right-grain. You can see the effect of this “error” in the cockling of the pages. The book resists being opened, and when it is, it protests by wrinkling its spine.

The book just doesn’t want to open properly, all the way, so that you cannot, as you should be able to, balance it in one hand with the pages fully open. A cross-grained book’ll want to mouse-trap — to flip shut if you don’t break the spine thus forcing it to stay open, though this tends to be more visible in a paperback book on a heavier stock. Not only is the cross-grainedness of the volume shown in these strange opening contortions, but the book as a whole just doesn’t want to shut properly. Thus although the cross-grained book resists opening fully, it also refuses to shut fully.

It is indeed a tortured volume.

So why is it cross-grained, and what does cross-grained mean? In an ideal book universe the grain of the paper will lie parallel to the spine. In a cross-grained book it is at right angles to the spine. The grain of the paper, the direction in which the fibers in the paper are aligned, is easily established if you dare. Just take a page and tear it. Paper will tear in a straight line when you tear in the direction of the grain, whereas your tear will wander all around when you do it the other way, against the grain.*

This tear test isn’t any kind of magic. In a Fourdrinier paper machine the very watery pulp is poured onto a moving wire-mesh belt; the movement of the belt causes the fibers in the slurry to align themselves with least resistance to the motion, i.e. in the direction of travel. Thus the fibers in a roll or sheet of paper will (almost all) lie pointing in one direction. For the same reason as tearing across the grain is harder, so too is folding across the grain. The only fold which remains after the book has been bound and trimmed is the spine fold, and it’s here you’ll detect trouble with grain direction. (In extreme cases you’ll also see a cockle in the forage of the book, a waviness of the pages which you can see when the book is just lying there.)

So why would Library of America have allowed their book to be printed cross-grain? Maybe on their reprint, their quantity being lower, they sought to save a little by using up an oddment — or maybe the printer decided to do that on the qt. In the latter case the publisher may well have screamed when they discovered the problem, and have ended up accepting the run along with a bit of a discount.

When you buy paper in sheets you have to remember to specify long-grain or short-grain, in other words, to indicate whether the grain should be parallel to the length or the breadth of the sheet. With rolls of paper this isn’t a consideration: the grain will always be parallel to the direction of the roll, never across it. Adjustments to the imposition and to the roll width are what’ll be used on a web press to ensure your book is printed right-grain.


* The same phenomenon will be known to any woodworker. Sawing across the grain of a board is much harder work than ripping it along the grain direction.

OUP’s sewing department in the 60s

Apparently compositors (all male, and, as the aristocrats of labor in the print works, determined to keep things so) were not allowed into Oxford University Press’s sewing department, shown here.

In the immediate post-World-War-Two period, many of the employees in book manufacturing, in the bindery mainly, were women. The beauty of employing women was that you got to pay them less, so in departments which demanded less of the “tote that barge, lift that bale” sort of thing, “girls” would rapidly become a majority.

I recently asked a retired British printer when it was that it became cheaper to print and throw away a portion of a job rather than to execute a more or less elaborate inserting operation: i.e when did materials become cheaper than labour? Without hesitation he named 1975 when the provisions of 1970’s Equal Pay Act came into force, and it was no longer possible to cheat women out of their just desserts.

As to what today’s gender balance in sewing departments might be, it’s hard to pontificate, since so few of them survive, but the many women working in book manufacturing now are getting more equitable pay.

In Japanese, suminagashi apparently means “floating ink”, not a bad name for this almost magical printing process. Our instructor Ling My Truong takes us through seven different techniques. Note that she refers to surfactant at the outset but never tells you when to add it — presumably the water she pours into her dish at the outset already has surfactant mixed in.

If you don’t see a video here, please click on the title of this post in order to view it in your browser.

This looks like something one could easily do at home. Hope we all succeed.

You don’t often come across real marbled endpapers in books nowadays. Just costs too much to have an individually printed piece for every copy. If you do find a marbled endpaper it’s liable to be a printed reproduction from a photograph of a single marbled sheet (look for the halftone dot), and see that the endpapers in all copies are identical, which they wouldn’t be if genuine marbling was in question. (Tristram Shandy did have a marbled page inserted into every copy of its first edition.)

See also Marbled papers.