College and Research Libraries B y W A L T E R Y U S T The Revision of Encyclopedias The editor of the "Encyclopaedia Britan- nica," gave the Reference Librarians Sub- section of the A.C.R.L.j at the Chicago midwinter conference, an inside view of the herculean task involved in revising an encyclopedia. MR. PLINY, that most ancient of en-cyclopedists, who is said to have prepared the first encyclopedia known, never, I think, revised his monumental work. I f , however, he had decided upon another edition there is no doubt that his second issue would have carried the num- ber two. It has so long been the custom to call editions by numbers. I do not know exactly why—but the term has come to connote in most books a partial change and in encyclopedias a complete over- hauling of contents. Among encyclo- pedias, new editions appeared in the eighteenth and nineteenth centuries at in- tervals of from ten years to a quarter of a century. Usually enough copies were printed in the first publication by sub- scription to last for a decade or more and a new edition appeared only when the first edition was exhausted. N o one ap- parently considered the advisability or in- deed the need of revision between print- ings, if there were any printings between editions. T h e ninth edition of the Ency- clopaedia Britannica was sold for longer than a quarter of a century without any revisions between printings. ( I must except, of course, the changes which were made in the five or more pirated editions of the ninth edition sold in this country.) T h e eleventh edition of Encyclopaedia Britannica was sold just short of a quarter of a century without changing the plates in any respect. It was commonly under- stood that it is not the office of any encyclopedia to be up to date as a year- book is. It often takes five years or a decade to know how properly to report a series of events, so that the information can be significant, conclusive, encyclopedic. Events usually cannot be conclusively re- ported immediately after they happen and users of encyclopedias generally realized that. Sometimes when a new numbered edi- tion was needed, the publisher issued not a completely revised edition, but what have been called supplementary volumes. T h i s meant that the plates of the original set of books were not changed in any respect but that additional volumes were added to the set to bring some of the material to date. A n d the entire edition, old plates and new plates, was given a new edition number. T h i s arrangement was not so satisfactory as the appearance of an entirely new edition for obvious reasons: a supplementary set of books actually reduced the physical value of the original set; it also meant that the user of an edition with supplementary volumes had in reality to refer to two sets of books in order to discover the information he wished; he also had to use two indexes to cover both the old and the new sec- tions of the work. MARCH, 1940 12 7 Chief Difficulty There were other difficulties encount- ered by the publisher because of the tra- ditional plan of issuing numbered edi- tions, or supplementary volumes, but the greatest difficulty lay in the organization of the publishing house itself. In the past, it took from two to five years to prepare supplementary volumes. During this period of preparation the information would get out that a new edition was in process and persons hesitated to buy the set of books already in existence. W h a t actually happened was that during the preparation of the new edition the publish- ing house became an editorial department entirely—the selling organization did no business at all. N o one would buy the sets on hand which represented great money value to the publisher, in anticipa- tion of the opportunity to purchase a new edition soon. T h i s meant, of course, that money was going out and none was com- ing in. Editorial and printing expenses ran high and the house enjoyed no in- come to pay for them. T h e financial history of such publish- ing houses with large encyclopedias might be represented by a cycle; half of which runs to a fair degree of prosperity because money is coming in, half to a period of outlay only. T h e business in the past has risen and fallen and the rising and falling has been sharp—with the result that every new edition of an encyclopedia has some- time or other suffered the threat of fi- nancial collapse. With the beginning of the twentieth century and under the stimulation of sub- scription book selling, subscribers to en- cyclopedias began to feel the need of up- to-dateness. They began to believe that an encyclopedia which did not carry the very latest of current events was in a sense a fraud, an attempt to "get away with something." T h e book salesman himself had a great deal to do with this change in attitude. T h e latest printing date for copyright in the book might be the current year. T h e salesman would in- sist, and I am afraid he still often does, that this date indicated how recent was all of the material in the set. He would sell an encyclopedia (and what is worse the subscriber expected to be sold on that score) much as he might a yearbook which can very easily be changed throughout with each year's printing. A Physical Impossibility T h e salesman, and often, unfortunately, the subscriber, cannot clearly understand that it would be a physical impossibility to make a complete revision of a great encyclopedia each printing year. Working at the greatest speed with the greatest number of helpers, copy for a complete revision of an encyclopedia of millions of words could not be secured, styled and prepared for the printer in twelve months. T h e editorial work done, it would then be necessary to take at least five to six months for the setting of type, the proof- reading, the printing, and the binding. T h e 1940 set, copyrighted 1940, of a great encyclopedia, completely revised, would be fortunate in having material in it as recent as two years before the date of publication. Y o u understand that I speak in general terms. A s a matter of fact, certain items of most recent occur- rence might very well be added even dur- ing the period the revised encyclopedia is on the presses, but such recent material would be of a very small quantity. And such minimum up-to-date insertions can be of value only to the unscrupulous sales- man. 148 COLLEGE AND RESEARCH LIBRARIES It is, I believe, a most unfortunate development in the distribution of ency- clopedias that the question of constant timeliness should enter so much into the sale and purchase. Not many pocketbooks could meet the cost of the thorough annual revision sometimes suggested by salesmen and expected by purchasers even if it were physically possible to make one. Continuous Revision It would seem to be understood why the modern tendency is toward continuous revision and against new numbered edi- tions. In the first place the publishing organization remains intact during the annual partial revisions for yearly print- ings. T h e sales are made with no disas- trous effect of a new edition outmoding the remaining copies of the current set. Each year's partial revision is paid for by the current sale. With this plan of continuous revision the set is in a state of flux. T h e extent of each revision for each separate printing may be great or it may be small. Whether it be great or small depends upon the character of the publishing house, and by that I mean, whether the publisher is deeply concerned about keeping his prod- uct in value all of the time; it also de- pends on the amount of selling that is done. If few sets are sold during the year it is impossible for the publisher to make many expensive revisions for the next printing. Of course, ideally, the publisher, if a suf- ficient number of books were sold, should be able to revise fairly extensively each year, provided his original capital invest- ment had been paid off. T h e percentage cost for revision, set up in the selling price, is an important item in the ideal situation. With the capital investment paid, the entire amortization percentage can be and should be devoted toward the problem of keeping the volumes in value. In substance the Britannica s method, I think, is carried out by all reputable pub- lishing houses, the extent to which it is carried out depending upon the amount of money available for the editorial work. When the eleventh edition was prepared, the original pages of the old ninth with the supplementary volumes of the tenth were cut up, pasted down on sheets of paper and each subject sent to a known authority in the field for revision. A group of con- tributing editors advised the editor as to necessary new inclusions—new entries to cover interests which had developed in the interval from the date of the ninth edition to that of the eleventh. T h e authors were asked to read over the old material and revise it to date. In some instances quanti- ties of the old material were salvaged or slightly modified and new material added. T h e same method was used in the prepara- tion of the fourteenth edition: the old eleventh with the supplementary volumes of the twelfth and thirteenth were cut up, pasted down, and the articles sent to the proper authorities for revision. T h e very same method is used in continuous revi- sion. First of all, all of the entries in the encyclopedia are divided into general classes of subjects. There are in all thirty rough divisions into which the 40,000 articles of the present edition of the Bri- tannica are divided. T h i s classification of entries occupies a card file, half of which is devoted to a classified carding, half to an alphabetical listing and each card bears the history of the article to which it re- fers. A s I have said, the classification is a rough one. Physics, chemistry, and the industries overlap considerably for obvious reasons, but the aim of the editor is to classify for authorship, so that related MARCH, 1940 12 7 articles, whatever the classification, may be handled the same year. I am reminded that in our experience some of the classi- fication was too rough. The young wom- an, for example, who classified under geology the subject "gall-stones" is no longer with us, nor the one who classified " J o b " under occupations. The thirty classifications are scheduled for revision over a ten-year period in such a manner that each classification is over- hauled at least twice during that time. Living men and the current history of countries are reviewed each year; popula- tion figures every fifth and tenth year and so on. It might seem necessary to revise all statistics annually. Ideally they should be revised annually, but there are statistics in so many classifications that it would be physically impossible to make anything like so extensive a revision. Per cent of Material Revisable Scheduling the classifications over a period of ten years does not, of course, mean that we expect each article in the classification to be revised. As a matter of fact, a hundred or more years of ex- perience have indicated that less than 20 per cent of the material in an encyclopedia is revisable. Approximately 80 per cent is "frozen." It is true that great political, economic, and industrial changes may bring about a change in point of view to- ward the examination of history, or a new discovery change the approach to the vari- ous sciences. When such changes of points of view are apparent, then of course, some of this 80 per cent of so- called unrevisable material requires scru- tiny. But that does not often occur. When I say 20 per cent of the material in the encyclopedia is revisable, unfortunately I do not mean that 20 per cent of the pages in the encyclopedia are revisable. The percentage of page space is higher— much higher. An article itself may be three and a half pages long but it may touch five pages in the volumes and this necessitates a handling of five pages when the article of three-page length is revised. As a result, in any revision, there are these points an editor must consider: the length of the article itself in pages and the num- ber of pages involved in the mechanical change. He should know the first because he must determine the pay he will give the author; he should know the second because he must keep the mechanical changes nec- essary for the revision within the limits of the editorial budget for the year. Suppose, for example, we have scheduled for revision this year the following groups: industry, engineering, living biography, and current history, which we included under the classification, geography. Our card index gives the name and location of each of the articles in these classifications in the encyclopedia. Tear sheets are se- cured from the printer and the text of each entry is pasted on a large white sheet of paper. If there is a contributing editor in charge of the classification "industry," he is given an opportunity to decide which of the articles in this classification requires revision and to decide the name of the new author, if the original author is not avail- able, to make the revision. The paste-ups, including whatever pictures go with the article, are then sent to the author. He is instructed to revise the material to date, to keep within the same space if possible, and to salvage as much of the original copy as possible. He is asked to examine the pictures for timeliness. If it seems to him that an adequate report on his subject has expanded out of all proportion to the space originally allotted to it, he com-' 150 COLLEGE AND RESEARCH LIBRARIES municates with the editor and a new space arrangement is decided upon. Fitting Copy Since we are working on plates of metal you can understand that problems of fitting immediately arise. Although we have wished frequently that type might be made of rubber, adjustments can usually be made successfully, even in metal. T h i s year, for example, the article on wire was completely rewritten. T h e new article is thirty-nine lines shorter than the space occupied by the old. T h e author did not believe the article could be expanded. In order to make up the additional lineage an article on William W i r t was added to the pages—an important figure in early American history. He should have been in the books long ago. Or, take another instance, the original article in the four- teenth edition on petroleum occupied per- haps as much space as it should have in 1929, the date of the original printing of the edition, but today it is out of propor- tion to the material, for example, on coal. Consequently, I asked D r . Fanning to write sixteen new pages in addition to the amount of material occupied by the original article on petroleum. These will occupy A and B pages added to the book. (And forever after, because we have placed A and B pages in the book, the manufac- turing cost will be increased each year.) In some instances it is necessary to re- duce certain articles in order that another expanded article can be made to fit. T h i s type of adjustment is made only when the material does not extend beyond one page; that is to say, when the adjustments can be localized. If the author cannot expand his article or reduce it, then the expansion or reduction is made in another article on the page. I may say that this reduction is never done without careful considera- tion. If an article will not be weakened by necessary excision, the excision is made. Librarians are sometimes troubled by this method, which, as a matter of fact, is not used excessively; but whether it is used often or not, it seems to some to be a rather highhanded treatment of original copy. I have no such reverence for origi- nal copy. I have had twenty-six years of experience in the handling of copy and in the writing of it. I have never seen an article the worse for wise condensation. With the possible exception of the Gettys- burg Address, the Lord's Prayer, and maybe one or two others, there are few pieces which cannot be condensed to im- prove them. T h a t this is true may be understood if you remember that authors themselves are quite willing to reduce in order to expand, paradoxical though the statement may be. Indeed, this method of localized adjustment is no different from that used in making an entirely new edition in a complete resetting. T h e finan- cial set-up for a new edition allows the editor a fixed sum of money to make the books. He calculates the size of his books and scales his articles to fit the space avail- able. If he is building a new edition out of an old one, he uses precisely the same methods I have just been describing, send- ing out pasted-up copies to be cut or ex- panded. T h e only difference is that in making the brand-new edition he resets all of his material—old and new. Repair of Plates Another item of expense to the pub- lisher is the repair of plates. Ordinarily about $ 1 5 0 0 a year is budgeted for the repair of plates alone. Repeated printing over the same plates wears down the type in spots. T h e printer, making ready for MARCH, 1940 12 7 the next printing, carefully examines those plates which are not to be patched or reset. T h e plates which show wear are then taken out and repaired for the current printing. Index With each change in an article the in- dex must be examined to see if there is a corresponding modification required there. T h i s is one of the most exacting of edi- torial tasks. Sometimes a card file of the index is used to control this w o r k — 500,000 cards for the Encyclopaedia Bri- tannica. It has been found, however, to be considerably more difficult and has taken longer to make changes by way of the cards than directly from a checking of the new and old text and a simultaneous examination of the index for the key words. If the entries or key words orig- inally in the article do not appear in the revised article, the index reference must be omitted, and the type of the index page extended to fit and patched. If the entry words or key words are moved to another quarter of the page or to another page altogether, the index correspondingly must be modified. Maps Maps are continually undergoing change at the hand of the cartographer. Our maps are made in Scotland. With each printing of the maps, proofs are sent to us with corrections indicated on press proofs. From these, changes to correspond are made in the index. Proofreading Professional proofreaders are perma- nently on the editorial staff. A s copy is received, it is examined by the editor, O . K . ' d for payment, and handed to the proofreader who styles the original copy. It then goes to the typist who types the copy on copy-fitting paper. When that is done the newly typed copy is read by the proofreader against the original copy. T h e copy-fitted paper is then returned to the editor for his final examination. If the copy is too long or too short, it is re- turned to the author for condensation or expansion. If the copy is only a little too long, an unnecessary word or two is struck out by the editor; or, if it is short by a few lines, it is expanded by carefully distributing a few paragraphs. Copy, so far as proofreading can make it, is correct and fits into the page line for line, before it goes to the printer. W e do not see copy in galley proof unless the copy does not fit. T h e new copy if cor- rectly fitted is placed into the page and two sets of page proofs returned. T h e author receives one, the staff proofreader the other. T h e proofreader again reads against the original copy and the pages are then held against the receipt of the author's corrected proof. When the proof is entirely cleared, it is re- turned to the printer and the page is plated. T h i s method of expertly proofreading copy on copyfitting paper before the printer receives it, has reduced alteration costs from as high as 40 per cent of the original manufacturing cost to less than 2 per cent. It also offers editorial ad- vantages: it is possible actually to limit copy to the size of the excision, it is pos- sible to index copy before the printer sees it and sets it in type, and it is possible to have the copy checked against related copy for consistency. It gives the editor a greater opportunity to scrutinize his ma- terial as a whole before the printer fixes it in hard lead type. 152 COLLEGE AND RESEARCH LIBRARIES Not Always Easy to Find Exact Truth Of course, all this preparation does not mean that errors of fact will not creep into the book. Since the human element enters into the making of books, errors are bound to creep in ; and they will. But apart from that, as I have suggested, it is not always easy to find the exact truth. I will give you an instance. Some time ago—and I forget the exact name of the river—a reader pointed out that he found in the Britannica conflicting statements concerning the length of an artificial river somewhere in Missouri. It was quite true. T h e length of the river differed in two articles. I attempted to run down this apparent inconsistency by writing to five authoritative sources, na- tional, state, and local. Believe it or not, I received five different lengths for that particular river, all of them authentic. N o w what is an encyclopedia to do? O r , another instance: D r . A prepared for Britannica Junior a revision of the article on Austria. D r . B prepared a similar article for the Encyclopaedia Britannica. D r . B believes that the date of the Aus- trian Anschluss is the date of Hitler's en- trance into Austria. D r . A believes that the date of the Anschluss is the date of the proclamation. T h i s difference of opinion is probably a difference of opinion held generally among historians. T h i n k of the arguments there will be over this date down the years! In discussing the problems of editing an encyclopedia I have said little about the contributors, how we find them, corre- spond with them and pay them. I have talked only of the less colorful aspects of encyclopedia making, the manufacturing and editing routine that can be accom- plished only by a great expenditure of money. I have talked much about costs and little about authority. And, of course, the reason is that the chief problem of revising is a problem of fittings and costs. T h e authorities are many, gracious, and their services gladly given. T h e costs and the fittings are not nearly so tractable, if I may use the word. They are the edi- tors' and the publishers' difficulty—and when a publisher devotes even a small por- tion of his income—when there is a n y — to a revision (provided his books are hon- estly prepared and honestly kept in value) the reader should be a little grateful. Be- cause—and it is indeed a commentary on the folk who buy books—the publisher knows that his books will be bought whether they are worth buying or not, whether they are revised or not, and that there are many intelligent folk who will think them good books even though they may not be. He knows, too, that 75 per cent of the experienced book salesmen do not care what is inside a set of books— and can sell one set as well as another— whether they are useful books or not. I say the conscientious publisher knows this as well as the unscrupulous one. T h e conscientious publisher, however, attempts to keep his books in value—and at a considerable reduction of his possible profit—year after year—by allowing his editor an annual amount of money (and every cent of it is needed) to do the best possible job he can. T h e business of re- vising books needs the business man even before it needs the editor and the scholar. MARCH, 1940 12 7