University of Virginia Library

The Bibliographical Concept of Plating
by
James L. W. West III [*]

Descriptive bibliographers have been in general agreement about most of the terms they employ since the publication of Fredson Bowers' Principles of Bibliographical Description in 1949. Everyone now knows, or ought to, what edition and impression mean, for example. But in certain

areas of bibliographical taxonomy there has been disagreement over terminology. The most noteworthy recent instance was a debate over the elusive concepts of state and issue between James B. Meriwether and Joseph Katz on one side and G. Thomas Tanselle on the other.[1] This paper is addressed to another problem, however, another spot on the chart of bibliographical taxonomy—this one between edition and impression.

In Principles, Bowers worked through the history of printing from the fifteenth century to the eighteenth, and in the last hundred odd pages of the volume he covered the nineteenth and early twentieth centuries. In his opening chapters he established four basic terms for describing and classifying the forms of a published book. These were edition, impression, issue, and state. As Bowers moved into the nineteenth and twentieth centuries he encountered new practices such as American and British publication from the same typesetting, limited editions, monotype rolls, colonial editions, and multiple platings of an edition. Bowers did not introduce new terms to describe these practices; instead, he adapted the four basic terms to cover the new situations. Certainly this decision was correct: in 1949 bibliographical nomenclature was unsettled, and one of the purposes of Principles was to establish a logical and conservative taxonomy. Bibliographical terminology needed to be rescued from the imprecise and careless practices of bookdealers, collectors, and other scholars. Bowers therefore confined his thinking within the framework of edition, impression, issue, and state, creating new terms by adding qualifying words or prefixes, or by combining two terms in a phrasal description.

One of the new practices Bowers encountered was plating. He did not address the concept directly, but in his commentary on edition and impression for the nineteenth and twentieth centuries he suggested at least four ways of dealing with plating. These four methods are all more or less satisfactory, and it is not my purpose here to suggest that they be rejected. I do believe, however, that bibliographers of modern books will encounter platings more frequently than Bowers could have predicted in 1949, and that these bibliographers will need a more straightforward and efficient way of dealing with the concept than he provides. In the remarks that follow, I shall attempt to describe a workable method which is essentially an adaptation of Bowers' thinking.[2]

Stereotype plating was not much practiced until the early decades of the nineteenth century and did not become widespread until the 1840s.[3] Before

that time, editions were printed from standing type, after which the type was distributed. If an edition sold out and more copies were needed, then a new edition had to be set. During the 1830s and 1840s it became common practice to make either plaster or "wet-flong" molds from the standing type and to cast metal plates or "shells" from the resulting matrices. During the 1840s the process of electrotyping also became popular; it was slower and more expensive than the plaster or flong methods but produced a more durable plate. In the twentieth century, plates have been made from a variety of materials—metal, plastic, and even rubber.[4] With a text preserved in plates, the publisher printed a first impression, stored the plates, sold out most of that impression, and then ordered more copies printed from the original plates, thus creating successive impressions of the same edition. In this way the publisher avoided repeating the most time-consuming and expensive part of the printing process—the composition of type.

During the nineteenth century publishers found many new outlets for their books, and they began to order much longer press runs. One set of plates was often not enough. Publishers found it inefficient to print all sheets at one location from one set of plates and then to ship those sheets, bound or unbound, to various markets. And plates, especially stereos, had a tendency to wear out on long runs. The solution was to cast multiple sets of plates from the original standing type, or to keep back a set of "mother" plates for a book which was expected to sell widely and to cast sets of plates, as they were needed, from these "mother" plates. Stereos and electros, however, were quite cumbersome. A single plate was a heavy block of wood and/or metal; the plates for a 400- or 500-page book were very bulky and weighty indeed. Storage, handling, reimpression, and shipping cost a good deal and often resulted in heavy batter or damage.

The spread of offset printing in this century simplified matters. In order to reproduce a book by photo-offset, all a printer really needs is one copy of the book, though two copies are preferable. Until a few years ago, when the technology changed again, the first impression of a first edition was usually printed from offset or "relief" plates.[5] Various forms of the text—repro proofs, negatives, "blues," flat sheets, unsewn signatures, or bound copies—were later sent to other publishers or book clubs, and new offset platings were made. (After about 1930 many British "editions" of American books were produced in this way.)[6] In the late 1960s most printers stopped using

stereotype and electrotype plates altogether and went to photo-offset. Even books which had been composed in hot metal were plated and printed by offset. Repros were pulled from standing type, usually on a power-driven proof press, and these images were photographed for offset reproduction.

In general, American trade publishers before about 1970 seem to have used relief plates until a book had exhausted its initial popular sale. Trade copies, sold mostly through bookshops at retail price, had to be well printed; and relief plates make a clearer, darker impression than offset plates. After initial trade sales slowed, the relief plates were usually destroyed, and subsequent impressions were manufactured by offset in order to avoid storage costs for relief plates. This was the case, for example, with The Thurber Carnival, which went through six impressions, from relief plates, from January to April of 1945. The stock of the sixth printing lasted until June 1959, when Harper & Brothers produced a seventh "impression" by offset replating.[7] I have found this same pattern in William Styron's Set This House on Fire (1960) and The Confessions of Nat Turner (1967), and in Reynolds Price's A Long and Happy Life (1962).

During the 1970s type was often composed on a tape-driven, computer-driven, or "memory" typewriter, and the images of the letterforms were frequently created by type elements similar to those on IBM typewriters. The technology of type composition has continued to develop quite rapidly, and today most shops use a process that is entirely computerized and photographic. Images of the letterforms are created by exposing film or special paper to light, and these images are transferred to offset plates. Stereos and electros have gone the way of hot metal composition, and there is probably no major printing plant in this country which still casts them. This is not surprising, as the advantages of offset are obvious. Major problems of storage and shipping are avoided, and the text can easily be enlarged or reduced to fit a different format—that for the Modern Library, for instance, or for a "quality" paperback series.

The recent development of word processors and optical character readers has created bewildering possibilities. The emphasis today is on capturing the initial key-stroke; the publisher attempts to make the author or his typist double as the compositor. The goal is to typeset or "key" the text only once.

Much energy is currently being put into programs which will "interface" the author's word processor with the publisher's composing machine. When that technology becomes generally available, it may have some good side effects. If, for example, a first novelist can present a publisher with camera-ready copy, he can make the publication of his initial effort more feasible commercially, and he may be able to avoid editorial cutting and meddling in the bargain. The possibilities are dizzying, and the bibliographer of modern books will have to deal with them. I believe that he can do so by beginning with the basic methods in Principles and applying his own curiosity and ingenuity.

In this paper I shall suggest some ways of dealing with the concept of plating. This discussion will cover the years from about 1840 to 1970. I am not yet prepared to deal with the period after 1970, but the approaches suggested here should be adaptable to the technology currently under development. The main difficulty is simply to keep up. The field is moving ahead so rapidly that nearly all equipment today is outdated by the time it is installed.

Bowers suggests four descriptive approaches to the concept of plating. On the first page of his discussion of edition and impression for nineteenth-and twentieth century books, he notes:

Bowers's approach is sensible and practical. But what of those cases in which it is possible to differentiate sheets printed from different sets of plates? One can certainly treat these as separate impressions, no matter what the publisher does. Elsewhere Bowers is reluctant to follow the common terminology of publishers, as in refusing to use edition loosely to mean impression, but here he simply notes that "there is nothing for it but to follow [the publisher's practice]." Bowers' compromise term for cases in which the sheets are distinguishable is states of a general impression. But in every sense of the term that I understand, these are separate impressions—if one can identify them.

Offset replating is handled by Bowers in a different way. On p. 382 he suggests facsimile impression as a working taxon, but here he seems to have in mind a modern offset reproduction of an old text—say, a 1965 facsimile of the 1798 text of Lyrical Ballads. In more complicated situations Bowers suggests, for the sake of clarity, that the replating be called an impression but

be qualified by a term within parentheses in such a description as second (offset) impression. This term—again an adaptation of one of the four basic words—is admittedly imprecise because the actual impression of ink onto paper is not made in both cases by the same physical "object," whether it be a forme of type, a relief plate, or an offset blanket. The term (offset) impression still conveys the message, however, and functions satisfactorily for a book with a simple history of offset replating.

The most useful taxon for dealing with the concept of plating is Bowers' term subsidiary edition or sub-edition. Bowers uses the term most often to record a difference in publishing auspices. He recognizes that "it is necessary to conceive not only of impressions with their issues but also of a family of subsidiary editions stemming from the parent edition type-setting, some having a direct line of descent, others a collateral. The editions in both these latter branches are identified by suitable descriptive titles which broaden the old terms to take account of the complexities attending machine-printing. Each member edition of this family tree may have its own impressions" (pp. 382-383). Bowers uses sub-edition in a great many circumstances, including "cheap editions," colonial editions, editions by other publishers, revised editions, enlarged editions, and limited editions (those printed from the parent typesetting and bound up before the "trade edition"). He emphasizes that the scholar must be "precise in a bibliography about the origin of a subsidiary edition" (p. 384), and it is therefore in the notes that the bibliographer would record the occurence of plating and replating. This approach works satisfactorily for books with a simple history of plating but is not as efficient in more complicated situations. Too, the root word of sub-edition is edition, a term which implies that type has been reset. With replating no type is reset unless corrections or other small alterations are made. Indeed the main reason for replating is to avoid the resetting of type.

Any term in a system of bibliographical nomenclature will work at a given spot so long as everyone agrees on a reasonably strict definition of that term and thereafter uses it consistently. We could therefore in theory all agree to use states of a general impression or facsimile impression or (offset) impression or sub-edition to describe platings and replatings, so long as we always specified carefully the circumstances under which publication occurred. But bibliographical terms, if they are to gain general acceptance, should if possible describe what actually happens at the printing shop. The current terminology does not do so. The problem is that the old terms—states of a general impression, facsimile impression, (offset) impression, and sub-edition—all use as their root or base words either edition or impression, and those two words carry mental associations for a bibliographer which are inappropriate to what the printer has actually done.

A new term would be helpful. I should like to propose that bibliographers experiment with the term plating as a step between edition and impression.[8]

The word is accurate because it describes what the printer has done: he has made a set of plates. A plating should be defined as "all copies of a book printed from a single set of relief or offset plates." Once the concept is grasped it becomes easy to distinguish the first plating of the first edition from the third plating of the first edition. Similarly one can differentiate between the first impression of the first plating and the third impression of that same plating. It would also be useful to distinguish whether platings were relief or offset. If one has the information, he can even record whether the plates were stereotype, electrotype, or flexographic. If the bibliographer treats all platings in a single chronological line of descent, he can insert the descriptive term within parentheses. He might, for example, distinguish the second plating of the second edition (offset) from the third plating of the second edition (electrotype). On the other hand the bibliographer might want to trace a separate line of descent for each method of plating. He then might dispense with parentheses and speak of the first stereotype plating of the first edition as distinct from the first offset plating of the first edition. More likely he will wish to separate the American and English publication histories. He can then distinguish the first American plating (relief) from the first English plating (offset), being careful to note that the English publication is a sub-edition of the original American typesetting, and that the image of the British plating derives from the American text. The printing history of a given book will dictate a bibliographer's decisions on these matters. Each plating can have its own chain of impressions under any arrangement.

The bibliographer of books printed in the eighteenth century or before will have no reason to use plating; he will continue to employ edition, impression, issue, and state in the old way. But the bibliographer of nineteenth- and twentieth-century books can insert plating as a step between edition and impression for those books which were plated and replated during their publication histories. If plating did not occur (or he cannot prove that it occurred), then the bibliographer will simply use edition, impression, issue, and state as he always has. If he does use plating, he need not adjust his definitions of the other four terms in any way. If he suspects that replating has occurred but cannot prove it, he should use the four basic terms and record his suspicions in a note.

How does one identify a plating or replating? First, one should attempt to consult the publisher's records. If these do not survive or are off limits, however, one can still detect platings from physical evidence (and one should always check the physical evidence in any case to be sure that it confirms what the publisher's records say). First one must learn to differentiate offset printing from letterpress. In books printed from standing type or relief plates, one can usually discern slight indentations where the inked letters have struck the paper. The indentations will often be visible if the pages of the book

are held at an angle to a strong light.[9] There are no such indentations with offset printing, and an experienced pressman can tell the difference by running his fingers across the pages of a book. Offset printing sometimes leaves a distinguishing mark called "offset slur," a blurring of the letters which results from slippage between the offset blanket and the paper.[10] If, by using these methods, one discovers that early impressions are from letterpress or relief plates and later impressions are from offset plates, then one has evidence of offset replating.

Stereotype plating also leaves distinctive physical evidence. If either the plaster or the "wet-flong" method was used, a slight reduction in the dimensions of the type page will have been caused by shrinkage of the cast. This difference can usually be detected by measuring multiple copies of a book. For example, the type page of a first printing from standing type will be slightly larger than the type page of later printings from stereos. Shrinkage is proportional to the size of the type page: approximately one to two millimeters of shrinkage in width will have resulted for a page that is 90 mm wide, for example, or two to four millimeters in height for a page that is 170 mm high. A larger page will show greater shrinkage, a smaller page less shrinkage.[11] When all plates were cast by the same method from one unused set of "mother" plates, however, or when multiple sets of plates were cast from the original standing type (itself never used for printing), then no significant variation will be found. Too, neither the "dry-flong" method (in general use by 1910) nor the electrotyping process created any discernible reduction in the size of the type page; multiple platings produced by these methods are therefore impossible to differentiate one from the other, unless one finds evidence of a different kind.[12]

Fortunately, such evidence is often present. Sometimes one will assemble a run of impressions—the first through the tenth, let us say—and will note the usual increase in type wear and batter from impression to impression. Then, surprisingly, the text of the eleventh printing will be crisp and batter-free. Or in similar fashion, one will discover textual variants between the first and tenth printings and then will find, in the eleventh printing, a reversion to the text of the first printing.[13] Such evidence probably indicates that

a new set of plates has been cast, or was cast at the same time as the first set and is now being put into use.[14]

Sometimes the stereotype or electrotype plates were cast with two or more pages to a single plate, and these plates were themselves signed for a particular format. In such cases, resigning or other evidence of reimposition may indicate replating. This evidence is not always reliable, however, because printers would often cast (and sign) these plates for more than one imposition pattern. Also, it was a simple matter for a printer to knock off the old signature markings and mortise in new ones on an old set of plates.

The bibliographer can also detect the presence of plating by examining the repairs made to damaged type. Peter L. Shillingsburg has described what one should look for:

Offset replating can be detected in several different ways. If the size of the type page has been significantly enlarged or reduced between offset impressions, then replating must necessarily have taken place. Variations of a millimeter or less are insignificant, but larger differences usually indicate replating. If there is evidence of reimposition between two books printed by offset methods, then replating must have occurred. If collation on a Hinman or Lindstrand machine turns up variants, then replating has occurred, at least for the plates on which the variants appear. Any change in gutter or register measurement between books printed by offset is also evidence of replating.[16] The distance between pages on an offset plate (which normally reproduces an entire forme on a single sheet of metal) cannot be altered once the plate is created. Differences in gutter and register measurements in books

printed from relief plates, however, do not indicate replating. Measurements can be altered between these plates when two or more of them are on the press at once. Even in the middle of a run, a pressman can add furniture or tighten the chase, thus altering the gutter or register measurements in the gathered and bound books.

All of these methods of identification are fallible. I have only tried to suggest the general kinds and combinations of evidence the bibliographer should look for.

Offset replating creates problems for bibliographers because it gives the author and his publisher a chance easily to make changes in the text of a book. Textual alterations were often made in stereotype or electrotype plates: old type was chiseled off and new type was soldered in. Publishers and authors for years had readings changed in this way to correct typos, to perform small-scale alterations in the text of a book, or to censor or expurgate it after publication. The same thing can be done with an offset replating. In fact, it is a simpler process with offset repro copy than with relief plates. All the publisher need do is identify the type font used for setting the book and find a shop which has that font, or one nearly like it. Then the publisher can have the erroneous lines reset with correct spellings or the offensive lines reset with inoffensive words. The reset lines must be the same length as the old ones if the type page is to remain justified; otherwise, large blocks of text must be reset. But extensive resetting is sometimes done if the publisher wants badly enough to change the text. Once the lines have been reset, the printer need only take scissors and paste and strip in the new lines over the old ones on his repro copy. He can then shoot negatives, make new offset plates, and go to press.[17]

This is what happened to the British edition of William Styron's Set This House on Fire. The printing history of this novel is worth examining in some detail, because it shows how pervasive the practice of offset replating has become in modern publishing, and because it affords an opportunity to test the workability of the plating concept. The first two trade printings of Set This House on Fire were published by Random House in 1960, from relief plates. The book was then replated four times by photo-offset: once in 1960 by the Book Find Club, again in 1961 by Hamish Hamilton in London, again in 1970 by Jonathan Cape in London, and a fourth time in 1971 by Random House itself (which had melted down its relief plates after the second trade printing). The 1970 Cape plating is of particular interest because Cape did not go back to one of the Random House impressions for repro copy. Instead Cape used copies of the Hamish Hamilton text. This created a problem, because back in 1961 Hamilton had used cut-and-paste methods to censor its text of Set This House on Fire.

The novel contains a fair number of four-letter words, but curiously

these were not what the Hamilton editor expunged. Rather, he was offended by the word "bleeding." Cass Kinsolving, hero of Set This House on Fire, is windy and bombastic and often speaks in overblown language. "Triple bleeding God," and variants thereof, are his favorite oaths. Most Americans would not find such oaths offensive; in fact, there are other off-color words in the book, curses that Hamilton left uncensored, which would probably bother a conservative American reader more. But Hamilton concentrated on "bleeding" and "triple bleeding," perhaps because England is a heavily Anglican

country and surely because "bloody" is a strong word there, or at least still was in 1961. So in the Hamilton plating, "Triple bleeding God" became "God Almighty," as did "Bleeding Savior" and "Bleeding God." "Thrice-punctured Christ" was entirely removed, as was "triple bleeding horseshit." And "Sweet bleeding Father" became simply "Sweet Father." A few other strong lines were toned down—the old Southern expression "grinning like a shit-eating dog" became "grinning like a big hairy dog," for instance—but for the most part the expurgations were aimed at modification or removal of the word "bleeding." In 1970 Jonathan Cape, Styron's second British publisher, made a new offset replating of Set This House on Fire from the censored Hamish Hamilton text, after first pasting in a few more misguided corrections. The Cape book is thus a second-generation, expurgated, twice-altered, offset replating.

The diagram in Figure I shows how the text of Set This House on Fire was transmitted from plating to plating. This diagram, however, records the transmission of the image of the text; it does not properly trace its publication history. In a descriptive bibliography, it would be much better to separate the American publication history distinctly from the British. The entry for Set This House on Fire would then be arranged in this fashion:

                     

FIRST AMERICAN EDITION
First American Plating (Relief): First Impression, March 1960. Second Impression, Sept. 1960,	Random House, 1960. (Plated from standing type)
Second American Plating (Offset): with corrections.	Book Find Club, 1960. (Sub-edition of Random House type-setting, derived from the first American plating, first impression)
First Impression, July 1960.
Third American Plating (Offset):	Random House, 1971. (Derived from the first American plating, corrected second impression)
First Impression, March 1971.
FIRST ENGLISH "EDITION"
First English Plating (Offset):	Hamish Hamilton, 1961. (Sub-edition of Random House type-setting, derived from the first American plating, first impression, with expurgations by Hamilton)
First Impression, February 1961.
Second English Plating (Offset):	Jonathan Cape, 1970. (Sub-edition of Random House type-setting, but derived from the Hamish Hamilton plating, first impression, and with further textual alterations by Cape)
First Impression, January 1970.

Increasingly in modern books, the American and British manufacturing histories overlap or derive from one another, and the bibliographer must record these interrelationships. A good example is Cape's offset replating of the 1967 Random House edition of The Confessions of Nat Turner. This Cape book is a bibliographical curio. Before shooting negatives, Cape made twenty-seven changes in the Random House text, some justified but others of questionable value, such as those designed to "correct" Nat's Negro slave dialect. These alterations came to light when the first printing of the Cape plating was machine-collated against the first impression of the first Random House plating. But numerous other variants also surfaced, and the pattern of variation was strange. I had assumed that the Cape plating would derive from a single Random House trade printing, and I was curious to know

which one because Random House had introduced some thirty-eight changes in its own plates between the first and sixth trade impressions. After analyzing the variants, I concluded that Cape had used two copies of the Random House text, one a first printing and the other a fourth printing (which varied from the first printing at fourteen points). I believed that the two copies had been disbound and that pages had been selected at random from one or the other for pasting up camera copy.

I was wrong, however: the variants actually broke down according to inner and outer formes.[18] Random House therefore probably did not send two copies, but two sets of sheets to Cape for photoreproduction. This would have been an advantage for the British printers; they would not have had to decide on an imposition pattern or arrange and align the separate pages for shooting. One set sent by Random House was apparently from the first impression, the other from the fourth impression. One suspects that a single set of sheets from the first impression (September 1967) was sent to Cape shortly after Cape contracted to publish the book in England. The Cape printers, however, must have wanted a second set. By that time Nat Turner, a bestseller in America, was probably into its fourth trade impression (23 October 1967), and a set of sheets was pulled from that impression and sent to Cape. As a result Cape produced an odd text which includes some of the corrections that had been made in the American relief plates, prints some independent changes which are found only in its text, and preserves from the first impression some errors that Random House had already corrected.[19]

There are an almost infinite number of variations one can play on the plating theme. One can devise, in the abstract, some truly curious bibliographical situations involving wet-flong molds, stereotype plates, censoring, and offset replating. I believe that the concept of plating can be used for something besides bibliographical parlor games, however, and I suggest that scholars experiment seriously with the term. For the description of modern books it is a helpful taxon; I hope that its acceptance and use will spread.

Notes