Chapter III

Printed Books

Word count: 14100

∣📄 p.13 The publishers of commercially successful books have very little occasion to study the cost curve of editions in the range of 50 to 500, this range in edition size that library demand and scholarly interest indicate as the most important in the reproduction of research materials. The cost curve behaves very strangely in this zone, so that the judgment of busines men acquainted with large-edition printing may be of little value here. How can these costs be analyzed most clearly, and the respective places of the various techniques of book production determined?

The stages of book production separate costs into three primary elements: a “first cost,” incurred regardless of the size of the edition; a “running cost,” varying with the size of edition; and a “reprint cost,” incurred when an edition is exhausted and a new printing made from plates. The reprint cost is independent of the size of the reprint edition; it is a kind of second “first cost,” followed by a running cost that varies with the size of edition.

The first cost is itself divisible into two very important elements: composition costs, here defined as all costs that vary with the number of words; and area costs, which are those that vary with the area of paper used in a single copy. Composition costs are found in typewriting and linotyping; area costs are found in some of the pressroom operations. To print in large type rather than in small affects principally area costs, not composition costs, because it means spreading the same number of words over more paper; to reproduce by photolithography is to render area costs paramount elements of first cost. Reprint costs are affected in printing by storing type (rates running around $.01 per page per month), or by making stereotypes and electrotypes; in photo-offset, by saving the metal sheets; in mimeographing, by ∥ saving the stencils.

The librarian receives as a book any bound volume, however it may have been manufactured. Library shelves give their space to many bound volumes of photostated pages (mostly copies of rare books) and many bound volumes of typescript (mostly theses). Occasionally a bound volume produced by the mimeograph finds its way to the shelves. In recent years there have been a number of books produced by the photo-offset method from typescript master copies. Many reprints of old books are manufactured by photo-offset. Such parvenu volumes, invading the library shelves, are regarded with no little contempt by the exponents of established printing press products already in possession. How far are these different techniques suitable for the production of books? How is the field to be allotted among them?

Here it is necessary to note that the book itself is challenged, even in the library. The vertical file, housing an array of unbound memoranda, pamphlets, propaganda matter, broadsides, and other materials, some of permanent value and some ephemeral, has come into the library reference room to take business away from the bookshelves. And now from another quarter comes the possible development of reduced copies on film of material to be read by projection, threatening the books with more serious technological unemployment. Not only the ordinary printed book of the conventional form, but even the book itself is placed in a competitive position where it once possessed a monopoly. It will require no little care to make sure that competition between the book and its rivals is regulated in the best interests of scholarship.

The starting point of an analysis of the technological problem of book production is a fundamental distinction to be made between two types of cost behavior, instanced by the photostat on the one hand, ∣📄 p.14 the printing press on the other. When copies are multiplied by the photostat or in miniature film copies, no matter how large the edition, the cost per copy stands constant. The element of “first cost” is absent; running costs alone prevail. This was the cost situation of all multiplying of writing before Gutenberg. The invention of printing introduced a situation in which the cost per copy fell as the size of the edition increased. For purposes of convenience all those techniques which operate on a cost system resembling printing will be analyzed first, and those techniques which operate on or near a pre-Gutenberg cost basis will be analyzed later.

Of the techniques operating on the system resembling printing (i.e., with a substantial distinction between first costs and running costs), several classifications are possible. The one that will probably appeal to scholars as the most natural would distinguish between those techniques which produce a book printed in the kind of book type with which they are familiar, and those which produce a book printed in typescript.

Of the techniques yielding a product in print-face type, there are two that are available for putting new material into book form: these are printing from raised type on a flat-bed cylinder press, and printing from raised type on a multigraphing machine. Two other techniques are available for making new books from photographs of old ones. These are the photo-offset method, which has many names,¹ and the “dermaprint” method. The photo-offset process works from a negative photographic copy to a metal printing surface, the dermaprint from a positive photographic copy to a mimeograph stencil surface.

There are also four techniques for making books in typescript face. Among them the leading place belongs to photo-offset, but room must also be found for the mimeograph, hectograph, and even the humble carbon paper. These techniques will be discussed in a later chapter. This present analysis is limited to the discussion of making books in ordinary print type from ∥ one of two kinds of copy—a manuscript that is being converted into a new book, and an old book that is being reprinted.

Printing with Raised Type on Cylinder Presses

From the time of John Gutenberg until that of Ottmar Mergenthaler, the process of printing remained fundamentally the same. Separate types, laboriously set up by hand, were used to make impressions upon paper or other substances. With the invention of the typesetting and typecasting machines in the nineteenth century, however, a revolution in printing methods was effected.

The most common typesetting machines, remarkable in their ingenious mechanism, are the linotype, invented by Mergenthaler, and the intertype, built upon the mechanical principles devised by him. These machines, which produce the greater percentage of composition in this country, assemble matrices containing the faces of letters in intaglio at the side of a mold, in which is cast the complete line in one piece or slug, with the type on the upper edge, properly spaced and justified. The Ludlow Typograph, another slug-casting machine, is used chiefly for display work. The monotype, invented by Tolbert Lanston, casts individual types in the correct order and automatically transfers them to a galley in justified lines.

The type as cast by the compositor is arranged in long columns. These must be proofread and “made-up,” that is to say, divided into pages and arranged in forms in such a way that the printed sheet, when folded, will page correctly.

The printing is then done on a flat-bed cylinder press. The type as laid on the flat surface, with its thousands of slugs and individual types, does not present a true surface, It is, therefore, necessary to prepare the surface of the printing machine cylinder (called the tym- pan sheet) so that even pressure is given on every part on the type. This operation which is a most laborious affair if conscientiously done, is called “make-ready.” Without careful make-ready, satisfactory ∣📄 p.15 results are unobtainable. ∣📄p.[14a] (blank) ∣📄 p.[14b]

∣📄p.[14c] (blank) ∣📄 p.[14d]

∣📄p.[14e] (blank) ∣📄 p.[14f]

Printing: Composition Costs

In printing from raised type on flat-bed cylinder presses, composition costs form a major item of expense in the manufacture of a book. To be sure, the composing machines, by increasing the speed and accuracy of composition, have reduced its cost, so that hand-set type, when used in straight textual matter, has become a luxury. But the elaborate mechanism necessary for casting type makes the cost of the machines so high (from $4000 to $8000) that a heavy overhead charge is necessary. The all-inclusive cost per chargeable hour for the operation of the machines range from $3.00 to $5.00. The composite average for print shops in the United States in 1932 was $3.927, according to the United Typothetae of America. The composite for monotype keyboard operation was $3.466, and for the monotype caster $2.858. These figures include all factory expenses, commercial and selling expenses, overhead, and the wages of skilled compositors, which form 45% of the total. In estimating cost of composition, a standard unit of 1000 ems is used; that is, the amount of text that will occupy the lineage of a thousand of the letter m, or, to be more accurate, the square of any particular type size. If the number of ems per word is calculated at four, there would be 250 words per 1000 ems. Since a compositor will set from 3000 to 5000 ems per hour, the cost of machine composition, reckoning on the basis of $3.00 to $5.00 per machine hour, would run between $2.40 and $6.65 per 1000 words.

Theoretically, the amount of composition is determined by multiplying the number of ems in one line by the number of lines set. It is the general trade practice, however, in measuring up type for the purpose of finding the cost of composition, to include headings, leads, slugs, etc., as composed lines.² Thus the number of ems in the page is found by multiplying the width by the length of the line. The author, therefore, pays in composition for excessive leading. The majority of books appear in 10-point type on 11-point body, the ∥ majority of newspapers in 6½-point or 7-point type on 7-point body. For each additional point of leading the author ordinarily pays about 10% more in composition. It is questionable whether legibility demands more than one point of lead between lines. Certainly types like Caslon Old Style, with its long descenders, can be sparsely leaded without suffering in legibility.

Composition costs can be controlled in some measure by using a compact type face. Most of the commonly used type faces, like Caslon Old Style, Scotch Roman and Modern, have a set corresponding to the face size, that is, an em of these type faces is square; others, like Binny Old Style and Century Expanded, are used at a slightly wider set; while others, like Bruce Old Style, Kennerley, Garamond, and Granjon, are cast on a slightly narrower set. Thus a page of a certain number of ems of Granjon would contain 7% to 10% more words than a corresponding page with the same number of ems of Century Expanded, as is obvious by comparing the word content of Figure I. Two economical print-face formats are shown in figures II and III.

The difference in words per 1000 ems between a compact and expanded type face, plus the difference between minimal and excessive leading, may make an appreciable difference in composition costs for a book of a few hundred pages, perhaps as much as 10% or 20%.

Most of the printing done in America is newsprinting. The American public reads more words in newsprint than in any other form. The publication of scholarly books, if estimated in number of ems, or area of pages, or tonnage of paper and size of edition, is almost negligible in comparison with the corresponding newsprint figures. Yet books, in general, are printed in larger type than newspapers, and with a layout that devotes a much larger proportion of the page area to margins. Newsprint in two-inch columns, as standardized in the American press, is certainly a format of remarkable cheapness and efficiency.

As a factor increasing composition costs, corrections and alterations deserve particular attention. From a description ∣📄 p.16 of the typesetting machines, especially the linotype and intertype, it is obvious that making corrections in the matter already set up is a costly operation, because each error requires the substitution of an en- tire line. With the monotype, corrections are somewhat cheaper, for only a single letter, already cast, need be substituted.

After composition has been made to correspond to the copy, all changes subsequently made by the author or his publisher, whether in the text or in the arrangement of paragraphs, are considered as author’s alterations and are charged against him. Not infrequently the author is presented with a bill for his alterations almost equaling that of the original composition costs. Author’s alterations are usually charged on the basis of time at the rate of about $2.00 to $4.00 per hour, varying with the wage scale, and higher when made by machine than when made by hand. In some cases alterations are billed on a unit basis with standard unit prices for changes, which enable the customer to check his bill. In some cases a free allowance, possibly 10% above the original composition, is made by the printer for author’s alterations, which means simply that this factor has been figured into the original bid.

Author’s corrections would be entirely unnecessary if the author would devote the same attention to the preparation of copy as he would to making a master copy for photo-offset, and be equally satisfied to accept this as a final form. By deliberately preparing a manuscript which would justify the instruction to “follow copy,” the author would contribute to a low cost level in ordinary printing in the same way, though in a less degree, as by taking upon himself the making of the master copy for the photo-offset process. Some editors affirm that authors are so given to habits of carelessness that reform in this respect is possible only by changing human nature.

Printing: Area Costs

As in the case of composition costs, it is difficult for the author to control area costs in raised type printing. Area costs consist of make-up and make-ready. These operations increase in cost with the number of pages, but also greatly increase whenever it is necessary to fit type around ∥ pictorial or non-textual matter during make-up, or whenever the tympan sheet on the press must be prepared for pictorial matter during make-ready.

The fundamental unit of all area costs in raised type printing is the form or signature. The form is the area of type that fills the bed of the press, and is imposed upon the paper in one printing operation. The signature is the sheet of paper, printed on both sides, which is folded into a definite number of pages—usually sixteen or thirty-two. The area costs of printing mount form by form and signature by signature. It is wasteful of area to utilize only a portion of a form or to leave part of a signature unprinted. Printing economy always requires that there be such a calculation of type face, type size, and leading that the text comes out even with the forms and signatures. Just as ems, not words, are the main cost units in composition, so forms and signatures, not pages, are the determining quantities in area costs.

The amount of space a manuscript will occupy can be calculated either by counting the number of words in a suffcient number of lines and then multiplying by the number of lines, or by figuring the number of letters or words to the square inch. The average typewriter, with pica type, writes sixty letters single-spaced and thirty letters double-spaced; and with elite type writes seventy-two letters single-spaced and thirty-six letters double-spaced to the square inch. The average print-face type, on the other hand, has approximately the following number of letters to the square inch:

The average print-face type, moreover, has approximately the following number of words to the square inch:

∣📄p.[16a] (blank) ∣📄 p.[16b]

∣📄 p.17 The number of pages, and hence the number of forms, required will vary with the point size of the type, the compactness of the type, and the amount of leading.

An important difference is to be noted in the effect upon costs of faces and compactness of type on the one hand, and point, size on the other. True it is that resort to compact type faces and light leading will mean fewer pages and hence fewer forms than extended faces and heavy leading. But the greatest difference in area costs is a resultant of two other factors—the proportion of paper given to text as compared with the proportion given to margins, and the point size of type. These elements of cost, more than any others, are likely to be related most directly to the beauty and legibility of text.

According to Vernon,³ there is no great variation in legibility between 6-point type and 10-point type. Readers and reviewers, however, often express preference for the larger type sizes, and their preferences, even if they be regarded as the results of habit rather than the consequences of the physiology of vision, must be taken into account by a publisher. How far it is wise to let considerations of economy in area costs bring about the use of the smaller type sizes must be calculated by comparing two quantities: namely, the percentage of cost saving that can be effected by shifting from the largest to the smallest type, and the sacrifice of legibility or beauty that results from such a shift.

How rapidly will area costs fall if type size is reduced? Consider a 6″ x 9″ page, with a printing surface 4½″ x 7½″, printed with 32-page forms. Newsprint in two columns with half an inch between them would yield 800 pages, 24,600 words per form (according to an actual count of words). Eight-point type, leaded one point would put 25,000 words on a form; 100-point type in the same type face, leaded one point, would put 17,500 words on a form. Figure IV shows the increase in word content per page that is made possible by shifting from 10-point Bodoni type to 8-point Bodoni, both being leaded one point. The shift from 10-point to 8-point type would save one form out of three, or one-∥ third of the area cost in a text of 50,000 words.

Printing: Running Costs

Running costs are not decisive in small editions. The difference between the cost of printing 50 and 100 copies is so small that firms often quote the same price for both. It is thus possible to have an overprint, which can be left in sheets until needed, at a very slight additional cost. Even with books of larger editions this procedure is followed. The Encyclopedia Britannica, for example, is printed in lots of 15,000, but is bound in lots of 2500 only. Electrotypes are made which are preserved for further printing, and the type is kept standing for making corrections and for making new electrotypes.

As the size of the edition increases to the point where composition and make-ready costs begin to be balanced by the cost of materials and running time in the press, the established conceptions covering book costs become operative. Many of the great inventions in the printing world since the Linotype have been devices of one kind or another for getting the printing surface on a cylinder and off the flat-bed, so that faster running and longer runs would be possible. This is the story of the stereotype, cast from a matrix, and of the electrotype, resulting from coating an impression in wax electrolytically with copper. But they have no place in the calculations of costs of small editions.

Printing: Analysis of Costs

In analyzing the composition, the area, and the running costs for producing a book by straight printing, one might take the cost behavior in Chicago as an example. A book of 50,000 words in an edition Of 250, on 128 6″ x 9″ pages, could be made for $246.57. Taking these figures as an illustrative sample, it appears that for composing and printing a book, about 50% of the cost goes for composition, about 32% for make-up, 12% for make-ready, and 6% for actual printing in a minimum edition. There is, however, great dissimilarity in the amount which the different departments contribute to the cost of the finished product in various shops of the country. ∣📄 p.18

In general, the wages paid to low-wage skilled labor for composition work and for make-up and make-ready must account for a large part of the printer’s bill. The averages of the United Typothetae of America show that wages account for approximately 45% of the cost of printing. Wages constantly increase in proportion to the total cost of production as the size of the edition decreases. The printing and publishing business, which is under the same pressure as other industries toward decentralization away from the big cities, can be expected to decentralize most rapidly its small-edition publishing. There is now setting in a contrary pressure toward the standardization of wages, but unless the regional and urban-rural differentials are closed up, small-edition publishing ∥ can be expected to drift to the areas.

The Multigraph Process

In multigraphing, the scholar encounters the curious situation that while he is trying to find an economical means of making small runs that will look like printing rather than typescript, the business man is trying to make large runs that will look like typescript rather than print. The multigraph (Figure V) is a small rotary printing machine designed for office use and employed chiefly for printing form letters and circulars of one or more pages in imitation of typescript and for making office forms from print-face type in imitation of printing.⁵

∣📄 p.[18a]

∣📄p.[18b] (blank) ∣📄 p.[18c]

∣📄 p.[18d]

∣📄 p.[18e]

∣📄 p.[18f]

The best adaptation of multigraph ∣📄 p.19 printing to the purpose of book production has been made by the Mimeoform Service under the direction of Dr. Ernest Kletsch, who styles his process the mimeoform process.⁶ He uses a Lanston Monotype for setting the type, which is cast in a low type mold, 8-, 10-, or 12-point. A special caster (the only part of his equipment that may not be found in almost any establishment supplying multigraph shops with composition) automatically delivers the line of short type, properly spaced and justified, to the channels of the multigraph drum. When the drum is filled, the caster automatically stops. It takes only half a minute to fit the drum to the multigraph machine, and the printing can begin. Proofreading is done immediately, and corrections are quickly made from the case of type.

There are certain defects in the appearance of the finished mimeoform product. Within the range of type sizes from 6-point to 12-point, all the type faces available for monotype composition are available for multigraph composition, and pages are produced by the multigraph with all the features of print-face, such as Justified lines, italics, bold-face, etc. But the appearance of the printed page sometimes displays certain imperfections. As the type is transferred to the channels of the drum, it occasionally slips up or down or slightly awry. This results not only in a poor alignment of type, but in some cases in an indistinct impression. In general, the product of the mimeoform does not approach the highest quality of the flat-bed cylinder press.

A limitation in format is also to be noticed. The space between the channels on the multigraph drum, into which the short type is transferred by the caster, is fixed, permitting not more than six lines to the inch, regardless of size of type. Thus all four type sizes available (6-, 8-, 10-, and 12-point) are cast on a 12-point body (note Figure VI, set in 12-point and 8-point type). This means that lines of ∥ 10-point type, the most commonly used for book composition, would be leaded two points; 8-point type, four points; and 6-point type, used for footnotes, six points. For spacing between paragraphs, a channel of the drum would be left empty. In 10- point composition the loss in compactness over ordinary composition leaded one point is about 30%, and this percentage increases as the size of the type decreases. The waste of space is particularly great in footnote work. The effect of this wastage of paper area can be neutralized by using the full surface of the drum and printing upon a large page, 8½” x 11”. Running costs are the same for this large page as for a smaller one, and indeed, in the writer’s opinion, the best possible format for mimeoform work is an 8½” x 11” page set in 10-point type in two columns (Figure VII).

One further limitation of the multigraph as a book-printing device must be mentioned: it is quite unsuited to any publishing in which illustrative material is to be put on the same page with textual matter. Although it is possible to incorporate line cuts or coarse screen half tones in a page of matter produced on the multigraph by purchasing small curved electrotypes of the illustrative material, the cost is such that it is more economical to use some other process—e.g., photo-offset—for highly illustrated text. On the other hand, if illustrations are on separate pages, the multigraph is well suited to their use, because they do not have to be interleaved in signatures but can be gathered for binding like any other pages.

The mimeoform process of multigraph printing is useful for the production of scholarly books when the edition is in the region of 100 to 500. Considerable savings over. ordinary printing are here effected because composition costs are reduced, in some measure, by wage-scale factors and by the very difficulty of making author’s corrections which tends to enforce upon authors practices that would be equally effective in reducing printing costs.⁷ ∣📄 p.20 Area costs are held down because make-up is eliminated, being combined with typesetting, and because there is no make-ready, and it is in area costs that the principal savings occur. Running costs are higher in proportion than printing press running costs because each page is run separately; for that reason, printing will always catch up with and pass mimeoform, if the edition is large enough; but in the region of 100 to S00, running costs are so small a part of the total costs that the economy of the flat-bed press in printing sixteen pages at a time ∥ is of small importance.

The following figures show the comparative prices of mimeoform and flat-bed cylinder publication, and show the mimeoform process lowest in cost as a means of presenting straight textual matter in print-face type in small editions. The advantage of the mimeoform fades out when the size of the edition passes the 600 mark. The costs per page as here quoted in 1934 are within the range of those that Bean found normal in 1927—from $2.25 to $4.00.

Note: The quotation on the “economy format” from one firm turned out to be higher than the average of five quotations on a less economical format. If the specifications had been sent out for quotations to the five firms who estimated for the upper table, the price would be lower. The point of this table is to compare estimates on a newsprint format with estimates on a good sound book format.

∣📄 p.[20a] ∣📄p.[20b] (blank)

∣📄p.[20c] (blank) ∣📄 p.[20d]

∣📄p.[20e] (blank) ∣📄 p.[20f]

∣📄 p.21

Alternatives to Printing

This is the cost picture that printing presents to the scholar. If the problems of scholarly publishing are not met at this cost level, the alternative to be studied is the use of typescript reproduced by photo-offset, the mimeograph or hectograph, or by some photographic process. Of these, photo-offset is the most frequently applied to book production. The first use of these processes is found in bringing back into print materials already available in printed form; but they can also be used for publishing new matter from typescript. However, these processes can be most clearly analyzed as devices for reprinting, and they will therefore be described in that connection.

Reprinting: the Photo-offset Process

The process known as photolithography, photo-offset, lithoprinting, offset printing, planograph printing, and a number of trade or firm names, has developed in the last thirty years from the combined application of the principles of lithography and photography, together with a new type of printing press.

Lithography was developed by applying the principle that grease repels water, but attracts a greasy ink. The process was discovered in 1796 by Alois Senefelder of Munich, who used the limestone of Solenhofen, Bavaria, which was especially good for absorbing water and taking grease. A stone surface was covered with the greasy marks of a lithographic pencil, then moistened and then inked. The ink would adhere to the pencil marks, avoid the moistened spaces, and transfer to paper as a lithographic print. This direct method from stone to paper necessitated the image to be drawn or imposed on the stone in reverse.

The same principle is applied in the photo-offset process, which is fundamentally lithographic in character. A photographic image is transferred, via a negative, to a metal plate covered with a light-∥ sensitive emulsion (albumen with ammonium bichromate). An exposure to an arc light hardens the portions of the emulsion under the transparent parts of the negative so that they remain on the plate after development as an image of the copy to be reproduced. Thus, the process is photographic. A sample sheet of the negative (Figure VIII) and a sample piece of a used metal sheet from the multilith (Figure IX) are bound herewith.

The emulsion image is of such a character that it is receptive to greasy ink. The plate, therefore, may be placed upon a press which moistens it (the moisture adheres to the parts of the plate not carrying the hardened emulsion) and inks it (the greasy ink avoids the moistened surfaces of the plate, but sticks to the emulsion image). The printing surface is flat; it has no raised surfaces like type faces, no sunken intaglios; hence the name “planographic.”

The third element of the process is the offset press. The first practical offset press, upon which flat-surface metal plates could be used, was developed in 1906. A few years earlier Ira W. Rubel, a lithographer of Nutley, New Jersey, conceived the idea of developing an offset press while watching a feeder operate a cylinder press. When the feeder missed a sheet, as sometimes happens, there was made on the tympan sheet an impression which offset on the back of the next sheet and produced a better print than the one Rubel was getting by direct impression. He thereupon developed an offset press in which a reversed image was printed by the flat-surfaced plate on a rubber blanket, from which the wet impression was offset on the paper sheet. It is this last feature which has given the whole process its name of “offset printing.”⁸

For scholarly publication, two small offset presses, the “rotaprint” and the multilith (Figures X and XI), which are used by business offices doing small ∣📄 p.22 runs of advertising matter or publishing a house organ, deserve special attention. When soliciting business, the users of the large offset presses regard these smaller units as little better than office appliances; the users of the small presses challenge this judgment. But in the presence of the NRA the large offset press people tried to make rotaprint and multilith come into the Graphic Arts Code, while they held out on the claim of being merely office appliances. Examples of their work will enable the reader to judge of their relative quality (see Figures XXV and XXVI in Chapter V). This volume is planographed with a Harris press (Figure XII), which has also been used in reproducing pictures in the chapter on illustrations.

Photo-Offset: Composition Costs

The composition of the text from which the photographic negative is produced is not actually part of the photo-offset process: in preparing a new book by photo-offset, either ordinary raised type composition or typewriting would be used, and the factors affecting these costs are elsewhere discussed. If ordinary linotype composition were resorted to with this process, no saving would be possible unless a considerable number of illustrations appeared; the production of typescript books by photo-offset will be examined in Chapter V.

In the present discussion photo-offset is being considered as a reprinting device, and in the case of a reprint of a book by this process, there is actually no composition whatever, the old book serving as master copy for the photographic part of the process.

Photo-Offset: Area Costs

In the photo-offset process cost behavior is controlled by the fact that first costs in the shop are essentially area costs. It is to the area costs, that is, to the preparation of the photographic negative and to the making of the metal plates, that chief attention must be devot- ed.

(a) The Photographic Negative

In the preparation of the photographic negative, various kinds of materi∥ als can be used. The photo-offset process was originally developed with wet-plate photography. This technique came to the photo-offset shop from the photoengraver. The photoengraver would coat a sheet of glass with albumen, collodion, and nitrate of silver in the order named, thus making a “wet plate.” This wet plate would go into the camera and be developed and fixed as a photographic negative. It would then have its surface strengthened by applications of a liquid rubber solution and collodion. The result would be a “sandwich” of the photo- graphic silver image between two sheets of collodion adhering to the glass plate. Then the photoengraver would soak off the sand- wich with a bath of acetic acid and use it as a negative in preparing the metal sheet printing surface. More recently it has become the practice to use film rather than wet plate at this stage of the process, and now paper negative is taking the place of film for some kinds of work. The cost relationships of these three photographic media are approximately:

When the material to be copied is of maximum legibility, with ample contrast between its black and white, the paper negative (Photostat N-26, Kodalith, Contrasto Polygraphic Company’s product or Haloid Company’s Litholoid) is wholly satisfactory. According to an estimate made for the Joint Committee, the paper negatives required in preparing metal sheets for a book of 128 pages, 6″ x 9″, would cost $5.12. If film were used, the cost of the films would be $17.92. Thus a variation of $12.80, or approximately 10% of a job priced at about $125.00 (exclusive of binding and paper costs) is controlled by the choice of film or paper negatives. On the other hand, when the copy is poor the photo-offset printer may be forced to use a wet plate negative.⁹

It sometimes happens that there will be no way of bringing the original copy of the book to be reproduced to the photo-offset printer. In such a case, a skilled photostat operator or commercial photographer, acting under instructions ∣📄 p.[22a]

∣📄p.[22b] (blank) ∣📄 p.23

given him for the purpose, can prepare the paper negatives. The situation may bring about a separation of the photographic costs from the strictly photo-offset costs. The trade does not usually work on this basis, but figures received from several sources indicate that a discount of 18% to 30% can be given if satisfactory paper negatives are provided to the photolithographer.¹⁰ Another basis for calculating this cost is found in the difference between the original and reprint quotations when the photolithographer has preserved the negatives for the book, but not the metal sheets. This discount when given runs near to 15%.

(b) The Preparation of the Metal Plate

However, in computing area costs, the preparation of the metal plate is a more important factor than the making of the photographic negatives. The price per ∥ area of photo-offset printing surface has been calculated by Fred W. Hoch, author of the Standard Book of Estimating for Printers, and published in the American Printer for June, 1932 (p. 55), as follows. According to information recently received, these prices hold today. From these figures it appears that the price of photo-offset printing surface should be somewhere be- tween $2.14 and $2.41 per 100 square inches. These figures agree roughly with those obtained from certain firms in 1933 and with quotations obtained in 1931.

(c) Full Utilization of Metal Plate Surface

Since the printing surface of the metal plates constitutes the most important factor in the determination of area costs, photo-offset work pages should be planned to use all of the surface of the sheets.

∣📄 p.24

This raises the question of the relation of page, plate, and press sizes.

There are on the market some eighty different sizes of offset presses, using metal sheets of eighty different dimensions. Yet such is the force and pressure of standardization in the trade that most operators will aver that all these variant sizes are best adapted to the two standard page sizes, 6″ x 9″, and 5½″ x 8½″ (and its double, 8½″ x 11″). These are, in fact, the page sizes that will divide most evenly into the standard cuts of book paper and bond paper. They are convenient, but they may turn out to be wasteful.

For example, the rotaprint metal sheet, now officially known as Rotaplate, measures 11″ x 17″. It will take perfectly two pages of normal typewriter size paper, 8½″ x 11″. A working margin of ½″ may be needed. But suppose a book of 6″ x 9″ should be run on this press. No less than 42% of the available printing surface would be wasted. In printing a book of 128 pages, 6″ x 9″) on this machine, sixty-four plates would be required, for no more than two pages could go on each plate. At $1.50 a plate, the cost would be $96.00, and of this about 42%, or $40.00, is wasted. Whoever would print 6″ x 9″ pages should find a printer with another size of press. In bringing back into print some old book that happens to be printed in large-size type, it might be advisable to reduce the size to one that will fit evenly, or with minimum wastage, on the metal sheet of the press that will be used on the job.

(d) Area Costs and Legibility

The principle which permits this economy has wide application. Since the printing surface for the photo-offset process, is prepared. photographically, the material to be reproduced can be enlarged or reduced to any degree desired. The purchaser buys the use of a certain number of ∥ square inches of printing surface; it is almost immaterial to the printer what is contained within these dimensions. This opens the publisher to an interesting temptation. He can enormously reduce the cost of reproducing any given text by reducing the size of the print. If he cuts the print in half; he inereases fourfold the amount of printed matter on his area of metal. Thus, it has been proposed that the New York Times be issued in a half-size edition for permanent preservation.

The ease with which photo-offset printers can enlarge or reduce the size of the matter they copy, taken in connection with the controlling position of area costs, gives to the problem of legibility a special importance. The size of the type-face can be fixed at any point, from microscopic to above normal, and while the diameter is shrinking in arithmetical series—1,2,3,4,—the cost of producing the copies is shrinking in geometrical series—2,4,8,16. This situation creates a temptation to pull the size of the print down to the lowest point of legibility. There would seem to be several thresholds: first, the limiting size of the type that can be read without special strain or attention (accordime to Vernon, it is 6-point type); below the there is perhaps the smallest type that can be read without the use of a reading glass; and below that, the lower limit of reading glass efficiency.¹¹ Efforts have been made to use photo-offset at all these levels. The Hispanic Society has reprinted a rare Spanish dictionary in a reduction to one-sixth (one-sixth of its original size.¹² The experiment of the New York Times with a miniature edition reduced 50% from 7-point type has been mentioned. These experiments together with others are appraised in Chapter X in connection with various optical devices for reading. For the present, it is necessary to collect such information as) may be available on the effect of photo-offset reductions of type upon legibility with the naked eye.

∣📄 p.25 The Joint Committee on Materials for Research secured the help of Professors Donald G. Paterson and Miles A. Tinker of the University of Minnesota in examining this problem. By a speed-of-reading test these scholars studied the legibility of newsprint of normal size (7-point) and newsprint reduced to 80%, 50% and 30% of normal size. The result of their work is a curve of diminishing legibility, as shown in Figure XIII.

To summarize: the relationships of legibility, type size, and area cost in photo-offset are such that costs of multiplying materials by this method must always be reckoned to include the type size and legibility factor. Just as the scholar with a new book to be set up in type for the printer will find costs depending on the number of words in the book rather than the size of the letters, so the scholar with an old book to be copied by photo-offset will find the costs depending on the size of the letters rather than the number of words.

Photo-Offset: Reprint Economy

An important cost element that must also be taken into account in evaluating the photo-offset process is that of reprint economy. The idea of “reprint economy” is one of basic importance in all publishing accountancy. It deserves, therefore, a close analysis. It has been shown that publishing costs are of two kinds: first cost and running cost, and that the first cost is a composite of two elements, composition cost and area cost. If a book can be made to pay its way in an edition of 2000, a reprint of the book, manufactured by the same process, could also be expected to pay its way provided 2000 new purchasers should appear after the exhaustion bf the first edition. The number of prospective readers that can make a first edition pay can always make a second edition pay. But under certain conditions, the number of readers required to make a second edition pay need not be as large as that required to make the first edition pay. This result will follow whenever some of the work that has been paid for, as the “first cost” of the first edition can be held over and used again without additional cost in a second edition. The revolutionary importance of this element in the ∥ cost structure of text reproduction is realized most completely in microcopying, but the element is also present in relief printing and in photo-offset work.

The printer from raised type can hold his type or change his form into an electrotype by printing it in wax and depositing electrolytic copper in the wax matrix. He can then preserve the electrotypes and make a second printing without incurring any composition costs. But electrotypes and stereotypes are almost out of the question except for large editions. The electrotype for a 6″ by 9″ page, in a book of 128 straight-type pages, will cost approximately $1.50 a page unmounted, $2.00 a page mounted. The corresponding photo-offset device is to preserve the metal sheets, protected with a coating of gum and ready for a second run.

Some firms do not save the metal sheets, but save rather the photographic negatives from which they were prepared. The policies of photo-offset companies differ in the matter of passing on to the consumer the reprint economies made possible by saving the metal sheet or negative. Some policies and economies are here described for purposes of illustration:

Edwards Brothers, Ann Arbor. Preserves metal sheets free of charge. If cost of first run is $1.50 for first 100 copies, cost of second run will be $.83, a difference of almost 50%.

National Process Company, 75 Varick Street, New York. Saves the metal sheet, charging $1.50 for cost of storing. This is usually a sheet covering sixteen pages, so that the cost is about $.095 a page. If the cost of the first run is $1.50 per page for 100 prints, and sheets are saved, cost of second run, including charge for storage, is $.80 per page for 150 prints, a difference of 42%.

Chicago Planograph Corporation, 517 South Jefferson Street, Chicago. Retains all negatives, and, if customer desires, plates as well. “Certain percentage” of reduction given on reruns.

The Copifyer Corporation of Ohio, 1111 Power Avenue, Cleveland. Keeps plates and negatives until the customer releases them. No ∣📄 p.26 charge is made for negatives, but a charge of $1.00 per plate is made after three months.

The Tudor Press, Inc., 251 Causeway Street, Boston. Holds plates and negatives for one year, and after one year, if ordered. Reprints made at approximately $.50 per page less. If cost of first runs as $125.00 (exclusive of binding and paper costs) and sheets are saved, cost of second run is $61.00, a difference of 51½%.

William B. Burford Printing Company, 58 South Meridian Street, Indianapolis. Saves both plates and negatives for six months without storage charge, allowing 50% reduction for reprints.

R. R. Donnelley and Sons Company, 350 East 22 Street, Chicago. Saves negatives without charge indefinitely. Plates can be saved on request for a short period for reruns. If plates cannot be used after this period, a charge is made for making new ones.

The importance of reprint economy is twofold. It can lower the cost of certain specific printings of books, and it can affect policy in regard to overprints. If a second printing is cheaply available, it is not necessary to risk large first printings that may remain unsold.

Photo-Offset: the Outlook for the Future

The development of the photo-offset apparatus as an additional shop equipment for ∥ letterpress printers has already had an important influence on scholarship through the production of cheap reprints. In at least one instance—the making of the new Early Modern English and Middle English Dictionaries under the sponsorship of the American Council of Learned Societies—it has resulted in an extraordinary saving of scholarly labor. The editors of these new dictionaries have selected the writings of this period to be used in the dictionaries. The manuscripts and early prints have been reproduced by photo-offset and set before the scholars in loose-leaf form—about 200 impressions of each page. The scholar takes one page and marks in it all the words to be indexed. A typist then takes one loose-leaf for each word marked by the scholar, types the marked work at the top of the leaf and files it alphabetically. This method would seem to reduce the making of concordances to an almost wholly mechanical procedure.

What kinds of technological developments of photolithography would be likely to have an important impact on scholarship? Nothing that affects the cost levels of the very long runs will have any significance.¹³ But if the time should come when the apparatus would be so far simplified and the process so far cheapened that it could become a kind of multiple photostat, a part of the equipment of any large library and serviced by ordinary library labor, turning out extra copies of library rarities for sale or exchange with other libraries and sometimes multiplying overused library materials for teaching purposes, great results in the intellectual ∣📄p.[26a] (blank) ∣📄 p.[26b] ∣📄 p.27 world would be likely to follow. How far is technology from that possibility? Certain new developments tending in this direction have come to the knowledge of the writer, but must for the time being be kept confidential.

The rotaprint was the first machine to offer itself as a piece of office equipment. The cost of a rotaprint installation is about $1000 (9″ x 14″ model, friction feed; suction feed, about $1575). The latest rotaprint model, R 50, is $660.00 complete. Compare this with the price of photostat equipment, which is $800. Then came the multilith, offering an installation for $630.00 (friction feed, in its standard size, 9½″ x 12½″ printing surface, 10″ x 14″ paper). There has been developed a combination mimeograph and offset press which will permit a changing over from the mimeograph stencil, inked from the inside, to a photo-offset metal plate, inked from the outside. There is some possibility that a system may be devised whereby a kind of photostat paper can be treated to become a printing surface. The operator would make a photostat copy of a page, treat the photostat negative, fasten it on a small press, and run off a number of copies. New developments are appearing every year. The major institutional changes in scholarship can be expected to begin with the development of a photo-offset process so simple that the printing surface is prepared at the area cost level of the photostat copy and the running costs of making copies stand at the cost level of the mimeograph,.

Reprint Processes: Dermaprint

The dermaprint is at present the process that tends toward this desideratum. It is a photographic process tied to the mimeograph. It is not planographic nor does it use the offset principle, but it does the same kind of work that the cheaper offset presses do. The relative fineness of the products of these processes can be compared if the sample dermaprint page bound herewith (Figure XIV) is compared with the sample products of the multilith and rotaprint (Figures XXV and XXVI).

The page to be dermaprinted was selected because it was yellowed with age and contained a line cut. The standard charge of the A. B. Dick Company for an ∥ 8½″ x 11″ dermaprint stencil is $3.50, which includes a small amount of retouching. If a considerable amount of retouching is necessary, there is a standard extra charge of $1.00 per hour. The cost of the materials needed to make this particular stencil was $.64, which included a film negative and a film positive. Had a paper negative and a paper positive been used, the materials cost would have been somewhat lower. One hour and forty-five minutes were needed for the preparation of the stencil.

While the A. B. Dick Company gave their very generous cooperation in the preparation of the dermaprint stencil reproduced here, they call attention to the fact that, in general, the dermaprint process at the present time is more to be used with direct contact prints than with the use of the camera. The reader may judge for himself, by comparing the dermaprint product shown here with the cost calculations given above, what place dermaprinting can have in the reproduction of research materials.

The dermaprint stencil resembles the ordinary mimeograph stencil in appearance. It is a fibrous sheet impregnated with a substance impervious to ink. But whereas the ordinary mimeograph stencil is impregnated with a material resembling wax, which can be mechanically abraded by the typewriter stroke until the ink will penetrate through the bruised parts, the dermaprint stencil is impregnated with some kind of gelatinlike substance which is subject to tanning or toughening in the presence of the chromate ion. Whereas ordinary gelatin will dissolve in water, a gelatin that has been tanned will resist water. This is the chemical principle upon which the dermaprint works.

The raw sheet, already covered with a gelatin, is sensitized with ammonium bichromate or, more commonly, potassium bichromate. Ultra-violet light on the potassium bichromate will release the chromate ions and tan the gelatin. It is only necessary, then, to subject to ultra-violet light those parts of the sheet that are to be kept ink-resistant. The dermaprint stencil must, therefore, be prepared by using a black-on-white or positive photograph, or, as in the production of the so-called “special stencil,” from an original ∣📄 p.28 black-on-white drawing on very thin or transparent paper, tracing cloth, vellum, or celluloid. In this it differs from the ordinary photo-offset metal sheet, which is prepared as a printing surface by using a negative photograph. The light that reaches the surface of the offset metal sheet hardens upon the metal the substance that will accept ink; the light that reaches the surface of the dermaprint stencil hardens the substance that will resist ink and keep it from passing through. The positive photograph or the original drawing is pressed against the dermaprint sheet in a vacuum frame and exposed to an intense light. Then it is given a bath in warm water and dried. The warm water dissolves out that part of the gelatin which has been protected by black ink from the rays of the light, but leaves the other part intact. Thus a stencil has been prepared automatically.

∥ The cost of equipment for the dermaprint comes into the same price range with the cost of multilith apparatus unless the photostat and the mimeograph have already been purchased.¹⁴

The cost of printing surface for the dermaprint process as compared with rotaprint and multilith metal sheets is shown in Table VII below.

The supplementary photographic supplies needed for the dermaprint should normally cost more than those for photo-offset, except in the case of the direct method, ∣📄 p.29 because the dermaprint must have a positive copy, while photo-offset can use the negative. The photographer is through when he has made his negative for photo-offset, but if he is preparing for the dermaprint, he is only half through and must still make a positive from his negative.¹⁵ The cost of making this positive copy can be kept down if it is made by the filmslide process (see Chapter VI). Ordinarily the dermaprint uses just twice as much photographic paper as photo-offset. Moreover, it is not quite certain that the dermaprint can be prepared with a paper negative; if it turns out that film negatives are required, the cost of photographic materials mounts even further above the level of the photographic requirements of photo-offset.

∥ Whether the dermaprint allows genuine operating economies is still an open question. The A. B. Dick Company offers to prepare these stencils, but the prices are so high that the service has no place in competition with other methods of book production.

Customer installations make these prices less important. The A. B. Dick Company has decided definitely to sell the unsensitized stencils at $7.00 a package or $.145 each, in the legal size, as indicated in a recent letter to the Joint Committee. Dermaprinting, therefore, would seem to be a practical reproduction method for any institution possessing photographic equipment and a mimeograph.¹⁶

∣📄 p.30 The following table is a summary comparison of the factors affecting book ∥ manufacturing costs by three processes—printing, mimeoform, and photo-offset.

Notes