Chapter XI

Cost Levels and Applications of Microcopying and Projection Reading

Word count: 27600

∣📄 p.121 There is a final limit beyond which the principle of reducing the size of the reading matter cannot be carried. This limit is reached in “filmslide” or microcopying. The name “filmslide” or “filmstat” has come to be applied to a process of copying on film and reading by projection. The word “microcopying” would be more appropriate if glass should come in as a photographic material.

Resolving Power of Emulsion

The technological limit of reduction of print is fixed by the size of the grain on a photographic emulsion.¹ The dark portions of a photograph are made up of clots of small particles of metallic silver embedded in dry gelatin. The relation between the size of these grains and the possible extent of photographic reduction of reading matter can be graphically illustrated if one imagines that he has before him a level table top and a few handfuls of sand, and wants to form letters on the table top by laying down lines of loose sand. If the sand is very fine-grained, like powder, he will be able to make smaller letters than if the sand is coarse-grained or gravelly. But no matter how fine the sand there will be a limit, a size of letter, in which the grains from the different strokes of the writing will run together, the hollows inside the “d’s” and the “e’s” fill up. Letters the size of these printed on this page could not be ∥ formed from loose sand on a table top, no matter how much care were used. The corresponding limit is found in reducing newsprint on photographic film. When the reduction is carried below twenty diameters, the letters begin to blur over; at twenty-five diameters the blurring is very noticeable; at thirty-six diameters, if the greatest care is used, the words are still sufficiently formed to permit faintly legible enlargements to be made. Beyond that, everything runs together.

Reduction Ratio in Copying

The limit set by the grain of the film to the possibility of making photographic reductions of reading matter can be called the “resolving power” of the film. The standard scientific measure for resolving power is in lines per millimeter, but this proves to be too refined a measurement for studying the practical usefulness of apparatus on the market under normal working or library conditions. Since the final term in the whole process is a legible projected image, a rough index of the legibility of the product can be used to make a comparison of films, cameras and treatments of film. Again the newsprint can be taken as a standard, and the efficiency of a microcopying process or material can be stated in terms of the number of diameters of reduction of 7-point newsprint; this number will be called the “reduction ratio.” If the reduction ratio ∣📄 p.121 for 7-point newsprint is sixteen diameters, the same material and process should copy 14-point type with approximately equal efficiency at a reduction of 25 to 52 diameters. Note, though, that fourteen-point type is not necessarily twice the size of a 7-point face.

Many conditions of a page other than the point size of the type will extend or limit the possible reduction ratio. The body of type (whether heavy or light), the spacing of the letters on the line, the whiteness of the paper, the blackness of the ink, the clearness of the edges of the letters—all these things will control the resolving power of a photographic emulsion as applied to the microcopying of a particular page. For every page, however, there is some differential which, for microphotographic purposes, will equate it with newsprint. Some specific handicap in reduction ratio can be allotted, either to the newsprint or the other page, so that newsprint reduced by x diameters is the microphotographic equivalent of the other page reduced y diameters. The reduction ratio is a quantity of fundamental importance governing all cost calculations in film copying. What is the highest practicable reduction ratio?

The limit of the reduction ratio in microcopying is set by three factors: the resolving power of the emulsion on the film or glass; the resolving power of the lens and the mechanical perfection of the camera; and the workmanship in focusing, lighting, timing the exposure, and developing the image.

Process—Positive Emulsion

The most widely used photographic emulsion for microcopying is the one that happens to be the cheapest—the “process positive” emulsion used in making positive prints in moving picture work, or in photogravure. This emulsion is manufactured to respond to the blacks and whites of a photographic negative, and is therefore well adapted to render the contrast of black and white in copying.

For a copying problem in which faded ink or stained or colored paper is involved, a panchromatic emulsion would be preferable. The emulsion used by the Recordak Corporation for microcopying is devised to take care of a wide variety of ∥ color contrasts. While not specifically a panchromatic film, it is yet more receptive to color contrasts than ordinary process positive emulsion. DuPont quarter-speed pan or Eastman micro-pan is recommended for work involving color.

There is every prospect that finer emulsions will become available, or that special processing may render the present emulsions capable of higher reduction ratios than they now accept in copying. But it remains to be seen whether the prices of these special emulsions or processings will be sufficiently close to the price of process positive film to make them a desirable substitute. In this chapter, except when otherwise stated, the kind of photographic emulsion under discussion is the process positive.

Camera and Lens System

The efficiency of a camera and lens system in affecting the reduction ratio is a product of three factors: the quality of the lens itself; the adaptation of the mechanism for accurate focusing; and the success of the film—holding mechanism in preventing a slight buckling of the film. All copying cameras save one hold the document still and make exposures one at a time through a lens which covers the whole film area on which the page is copied. The Recordak apparatus uses a different principle: the documents are copied while in motion. The light passes through a slit to the moving film, and falls in a narrow band across the film. This optical system is said to simplify the elimination of distortion near the edges of the film.

It is not easy to isolate these factors in appraising the efficiency of a camera and lens system, but the writer, after some experimentation with a number of cameras, has come to the conclusion that, taking all factors into account, it seems perfectly safe to regard eighteen diameters reduction of newsprint as well within the capacity of most film, apparatus, and workmanship. Eighteen diameters reduction of newsprint should be substantially equivalent to thirty-six diameters reduction of fourteen-point type.

Limits of Reduction Ratio

As an ultimate limit, a reduction of twenty-five diameters of newsprint might ∣📄 p.123 conceivably become practical, but at such great reductions another consideration would have to be taken into account; namely, the amount of light that can be forced through a translucent film covered with dry gelatin emulsion without damaging it by overheating.

The Eastman Company, in developing its Recordak copying machines for documents and newspapers, has allowed a certain margin of safety by setting sixteen diameters as its reduction ratio. It also has a document copying machine set to make reductions of seventeen, eighteen and four-tenths, and twenty-three diameters. The choice of certain of these ratios was probably the consequence, not so much of limitations of resolving power as of convenience in that ordinary business documents are 8” wide and will go easily on 16mm. film if reduced seventeen diameters; while the newspapers are 18” wide, and will go evenly on a 35mm. film (perforated on one side only) at the reduction ratio sixteen. The twenty-three diameter reduction is satisfactory for typescript and will put 50% more typescript on a foot of film than seventeen diameter reduction will put on.

In comparing the work of different cameras, an arbitrary scale of legibility was set up, as follows:

• A Legibility: All letters distinct.
• B Legibility: The edges of the letters begin to blur. ∥
• C Legibility: The tops of the “o’s” and “a’s” are blotted over solid, but the openings are still clear in the “o’s”, “d’s”, “p’s”, and “b’s.”
• D Legibility: Almost illegible. The openings of the “s’s”, “o’s”, “d’s”, etc., are blotted solid. Separate words can be deciphered but not read.
• E Legibility: Absolute illegibility.

Then seven film copying devices (these devices will be described in the next chapter) were tested, allowance being made in some cases for better and poorer runs, or for expert and amateur workmanship. The difference between the worst and the best workmanship was much greater than the difference between the worst and the best apparatus. The conclusions of the experiment can be summarized on the following table.

Another way of demonstrating the limits of resolving power is to microcopy a page and then enlarge it back to its original size. The loss of legibility suffered by the reducing and enlarging process will then show itself.

In the following exhibits, typescript and newsprint have been carried by various cameras to varying degrees of reduction and enlarged back to original size. These exhibits require careful explanation. The most instructive one is that prepared by Dr. L. Bendikson and Mr. E. F. Morkisch

∣📄 p.124 at the Huntington Library, San Marino, California (Figure LXIIb).

Dr. Bendikson was furnished with a page (Figure LXIIa) upon which four clippings were pasted, and he was asked to reduce this page by 16 diameters with his Leica camera and then enlarge it back to original size. But, as can be seen, two of the clippings were already reduced 50% from original size by photo-offset. Dr. Bendikson was asked to enlarge these reduced clippings back to twice the size in which they came to him. The enlargements are here reproduced as Figure LXIIb. At the top of the page are the two clippings, one of newsprint and the other of typescript, which have been passed through a reduction of 16 diameters; at the bottom of the page are the two clippings, also of newsprint and of typescript, which have been passed through a reduction of 32 diameters. The difference in legibility that appears in the exhibit registers the loss of definition in utilizing the higher reduction ratio.

Consider the number of steps, photographic and in printing, that have marked the transmission of the two clippings at the bottom of the page:

1. The original newsprint and typescript were photographed on a negative in a reduction of 50%.
2. This negative was then burnt into the photo-offset metal sheet.
3. The metal sheet was inked and printed ∥ via a rubber blanket on paper. (The appearance of newsprint in this reduction can be noted in Figure LIV, a page of the New York Times reduced 50% by photo- offset.)
4. The photo-offset copies of the two clippings were then photographed on film in an additional reduction of 16 diameters, making a total reduction of 42 diameters from original size.
5. The two clippings were then enlarged back 32 diameters on bromide paper.
6. The bromide paper was then photographed full size on a negative for the photo-offset press.
7. The negative was burnt into a metal sheet.
8. The metal sheet was then inked and the ink transmitted via rubber blanket to the page of this book.

The superb photographic craftsmanship of the Huntington Library staff may make the test unfair as a true comparison of apparatus. The clippings reproduced below were microcopied as follows, from clippings of newsprint and typescript by the Folmer Graflex Corporation, in a reduction of 16 diameters, and enlarged back to normal size.

These reproductions are all well within the zone of legibility. But it is otherwise with the clippings that have passed through microcopying from originals that were already reduced 50% (Figure LXIIc). The effective reduction from the

∣📄 p.[124a]

∣📄p.[124b] (blank) ∣📄 p.[124c]

∣📄p.[124d] (blank) ∣📄 p.[124e]

∣📄p.[124f] (blank) ∣📄 p.[124g]

∣📄p.[124h] (blank) ∣📄 p.125

original text in the case of these clippings is double that of the reduction in the clippings on Figure LXIIb. They have not been enlarged back to the original size of the texts from which they were drawn, but remain in a half-size reproduction. In most cases, the eight stages of transmission through which they have passed have obliterated all legible qualities.

The following clippings are reproduced as Figure LXIIc.

1. Clipping reduced a net 32 diameters from original and enlarged back 16 diameters, to 50% of its original size, by Dr. Bendikson with Leica equipment.
2. Clipping reduced in the same amount and enlarged back in the same way with Recordak camera and amateur enlargement photography.
3. Clipping reduced a net 32 diameters from original and enlarged back 16 diameters, to 50% of original size, by Folmer Graflex Corporation.
4. Clipping reduced a net 16 diameters from original and enlarged back 8 diameters, to 50% of original size, by the Folmer Graflex Corporation with its recording camera.

One of the most extraordinary uses of microcopying, and seemingly one of the most efficient in the history of photography, was the work of Monsieur Dagron during the siege of Paris in 1870. Dagron’s postal service into the beleagured city was set up to use carrier pigeons. To increase the number of words a pigeon could carry, messages were printed on a large sheet of paper and microphotographed on film. The film was then inserted in a quill and carried by the pigeon. A piece of this film came into the hands of Dr. Bendikson of the Huntington Library. The fragment printed herewith is an enlargement of 32 diameters from the film. The exact size of the film, which contains eleven pages of three columns each, is also shown. There is no direct evidence to establish the size of the print from which the microphotograph was made, but by counting the letters in a line, and the number of columns per inch, it is possible to estimate that the vertical dimensions of the original typography must have been about ∣📄 p.125 the same as modern newsprint, and the horizontal dimension somewhat smaller. The reduction from which this enlargement was made must have been the equivalent of some- thing between 32 and 39 diameters from newsprint.

Cost of Photosensitive Surface

The cost of the photographic surface used in microcopying varies between $.14 and $.50 per square foot. The cheapest film is the motion picture nitrate film which sells for $.01 a running foot, 45mm. wide; but this is, or ought to be, excluded from consideration. No nitrate film should ever be used for copying work, for the cellulose nitrate upon which the emulsion is spread is highly unstable in the presence of heat. In its chemical composition it is a half-finished high explosive, like a dynamite stopped halfway in the course of manufacture. It cannot be preserved with safety except in expensive vaults and it always brings a fire risk with it.

The so-called “safety film” or cellulose acetate has a fire resistance about the same as that of paper of similar density. Hold a match to a sheet of dry cellulose acetate and it will crumble but it will not sustain a flame. It has high stability—Mr. B. W. Scribner of the Bureau of Standards has stated—equal to the best-made papers. Safety film is sold ordinarily at $.015 a running foot in the 65mm. width. It can sometimes be purchased from firms dealing in odd lengths for a lower price, but the $.015 is standard, whether for perforated or unperforated film. If the film is perforated on both sides, one inch is usable between the perforations; this figures out at $.18 per square foot of usable photographic surface; if the film is perforated on one side only, 1 ⅛” out of the 35mm. are usable, and this makes the usable surface cost $.1605 per square foot. For unperforated film, with 1 ¼” usable, the price per square foot comes to $.144.

The 16mm. film, perforated on one side only, is sold for $.0075 a foot; the usable width is .43” to .5”; therefore the price of a usable square foot (28 linear feet) is $.21 at the highest, or $.18 at the lowest. The 7Omm. film, perforated on both sides, is supplied on special order ∣📄 p.126 at $.035 a running foot, or $.185 per square foot.

A slightly different emulsion with a different weight of supporting film is manufactured for use in photoengraving work and sold by a different method. Because of these differences, the process film cut in rectangles is sold at a higher price. A dozen sheets, 5” x 7”, retail for $1.45 list price. Discounts in the trade will reduce this in large quantity orders, so that the photographic surface in this form costs $.50 per square foot in small quantities, $.325 per square foot in large quantities. The same emulsion on glass is sold for the same price in the same size.

The tabulation of these figures indicates the price range of translucent photographic surfaces, whether in the form of film or glass, of rectangles or long strips:

A strip of the 16mm. and a strip of 35mm. film are shown in the accompanying sheets.

These figures for translucent surface are to be compared with those for photostat or bromide paper. Insurance bromide paper sells for $3.60 per gross cut 8” x 10”, or $.045 per square foot. The same price, or in some cases $.005 more or less, buys most kinds of photostat paper. The difference between $.045 and the prices listed above is paid for the purpose of having the photographic image on a translucent medium. Translucency costs from $.10 to $.28 a square foot.² ∥

a razor blade set in it, will cut the paper in these strips, making the low price available.

Film or Glass Versus Paper

Why is it desirable to use a translucent medium in copying? One part of the answer to this question is found in the behavior of light. In discussing the possibilities of reading reduced-size print by means of a reading glass, it was stated that a “direct” optical system can be used in raising the ostensible size of the letters one diameter, but that it soon comes to call for more intensive light on the surface of the paper than would be required if no magnifying were done. The same principle, pressed one degree further, will permit a legible image of a page to be projected from a paper copy by reflection in a projection lantern if sufficient light is focused on the paper. As the enlarging ratio increases, it approaches and reaches the point where the surface of the paper will not reflect sufficient light to make the projected image legible. At this point the translucent medium becomes useful, for light sufficient to render legible a projected image of a reduction up to 25 diameters (and perhaps beyond that) can be forced through the film or the glass. The additional cost of the translucent photographic medium must justify itself by the increased utilization of light in the projection process, which renders possible the greater reduction ratios that are standard in microcopying.

Film, Glass or Paper? Film Strips or Rectangles?

The basic comparison of the three ∣📄 p.[126a]

∣📄p.[126b] (blank) ∣📄 p.[126c]

∣📄 p.[126d]

∣📄p.[126e] (blank) ∣📄 p.[126f]

∣📄p.[126h] (blank) ∣📄 p.127

possible reduced-scale copying methods from the standpoint of costs of surface can be established if we consider the cost of copying 256 square feet of printed matter (the equivalent of 93 pages of the New York Times, or of an octavo book of approximately 300 pages). It is evident that this amount of printed matter can be copied at different reduction ratios on areas of photographic surface at the following costs:

It is worth noting that the lower threshold for reading glass legibility, set in the preceding chapter at 7 diameters reduction from typescript, would permit the copying of this 256 square feet of paper on 5.2 square feet of photographic surface, the price of which would be $.23. If the photostatic reduction on paper were to be four diameters, the price of the paper would be $.72; a reduction of two diameters (i.e., to one-half normal size) would require photostat paper costing $2.88.

These figures indicate a very interesting problem in planning the use of photography in scholarship. The zone from 6 to 8½ diameters of reduction on paper corresponds in materials cost to the zone of 11 to 16 diameters on film, 16 to 22½ diameters on glass. Processing costs under present conditions probably give additional advantages to the film strips in the case of a long series of pages that will fill 100 feet or more of film, but this advantage in favor of film may be reversed in the case of short items, especially when a number of positive copies are required. ∥

Reading apparatus for paper in reductions of 6 to 8 diameters has not been developed as far as that for film in reductions of 16 diameters; there is no reading apparatus at all for glass, and reading machines for film positives made from rectangular glass plates with multiple exposures have not yet been brought to perfection in design. Storage and filing of paper in short items may prove less costly than storage and filing of film; glass would probably cause higher storage costs, but would have possible advantages in durability. The fundamental choice between the long strip and the rectangle, between paper, film and glass, between the low reduction ratios required by paper and the high ratios permitted by glass are not conclusively indicated by the present state of technology and experiment. But precisely because the use of moving picture film has been more highly developed as a quantity industry, it is to to this medium that most of the experiment and development has been directed, and consequently it is in terms of the use of this medium that the whole problem of microcopying can be best presented at this time.

For that reason the analysis of the possibilities of microcopying on the following pages will assume standard moving picture film in the standard widths of 16mm., 35mm., and 70mm. as the norm, and treat paper and glass or rectangular film as possible exceptions.

The Simplified Practice Division of the Department of Commerce has asked for an agreement by all those promoting microcopying to adapt all apparatus to the 16mm. and ∣📄 p.128 35mm. widths. In replying to the Division of Simplified Practise recommendation, circulated in September 1935, the writer, as chairman of the Joint Committee on Materials for Research, made the reservation that simplified practice should not exclude at this time a size of film that will permit of vertical filing, and coördinate with the use of paper and glass. But for the copying of a long fixed series of material, it seemed best to concur in this standardization of film width. A “long fixed series” is any amount of documentation that will require 100 feet or more of film for copying, and will not call for the insertion of a newly photographed item between two items already photographed.

Full Use of Area of Moving Picture Film

When 16mm., 35mm., or 7Omm. film is used for microcopying, one dimension of the photographic image is rigidly fixed; it cannot exceed the usable width of the film, nor can it be less than the width of the film without causing a wastage of photosensitive surface. The dimension along the length of the film is variable provided the photographing is done in a camera, such as the Filmograph or Recordak, which permits the operator to vary the amount of film used at each exposure. But the crosswise dimension is fixed, and therefore, attention must be given to the placing of the photographic image on the film. The same principle of maximum use of area that was found operative in photo-offset work is applicable to microphotography. Just as it is wasteful to print a 6” x 9” book on an 11” x 17” photo-offset ∥ metal sheet, so it is wasteful to make on 35mm. film an image only ¾” wide.

There are three, or possibly four positions in which book pages may be copied on moving picture film. The position chosen for photographing a book on the full width of the film determines the reduction ratio. The three positions, hereafter called first, second and third, are illustrated on the following diagram.

The relation of these book positions to the reduction ratios can be illustrated by calculating the reduction ratio of a book 6” x 9” in all three positions on the two kinds of film, 35mm. and 16mm. The corresponding variation in cost of photosensitive surface can be illustrated by calculating the footage and price of film necessary to copy 100 pages in each of the positions and on each of the widths of film.

Verneur Pratt, president of the International Filmbook Corporation, has proposed that the reading material might be arranged in a fourth way—the lines of print running across the film, making a column the full length of the film but leaving space on the film for more than one column.

If Mr. Pratt’s plan for multiple exposures on strips should be applied to 70mm. film, from 8 to 12 columns of ordinary book pages reduced 16 to 20 diameters could be run on the width of the film. There is at present no apparatus for making photographs in this way, but the price of film for 100 pages can none the less be calculated. A hundred pages would go on 4½ to 9 linear inches of film, the price of which would be $.015 to $.0275.

∣📄 p.129

Processing Costs

The processing costs for microcopies are like the costs of the photographic surface in that they apply to a footage of film strips, or, roughly, an area of film.

There are two kinds of processing involved in microcopying: first, the developing and fixing of the negative; and second, the making of contact or other prints from the negative. If a unique copy only of a book or document is desired, there is no need for the second processing.

The commercial processing of film has been organized in two kinds of business enterprise. Short strips of film are processed for the amateur photographer by services available in every drug store. Long runs of moving picture film are processed at prices that may be much less per ∣📄 p.129 film area than those charged by the ordinary service.

The short five-foot strips of film used in such cameras as the Leica and Contax can usually be developed (but not printed as positive contact prints) at the corner drug store. The charge is usually $.15 or $.20—hence $.03 or $.04 a foot. One New York specialist in developing the five-foot rolls charges $.30 or $.06 a foot. Many users of these small microcopying cameras do their own processing.

Film strips of 100 feet or more in length can be processed in the film laboratories serving the moving picture industry. Quotations from a number of laboratories, analyzed to make allowance for various special charges connected with sound track printing, establish the following scale:

∣📄 p.130 The variation in prices is not only a difference that appears from firm to firm, but also responds to differences in footage, and in number of prints to be taken from the same negative. The minimum prices appear at 10,000 feet of film where this is the basis of the differential, or at five prints from the same negative where the differential 1s based on number of prints.

The Recordak Corporation operates on a somewhat different basis; it sells the film and processing as one unit and throws in a metal reel and pasteboard box container to boot. The price of 200 feet of 16mm. film is $5.50, which includes the processing. The film is not the ordinary positive emulsion, and is not available without the processing. But if it were quotable at $.0075 a foot, the processing and servicing elements in the price would stand at $.0188 a foot. This price includes not only special processing for contrast and permanence, but also servicing in the cutting of short strips of film, packaging, and mailing.

The Recordak Corporation also offers for sale 35mm. prints of certain newspaper files. The file for the New York Times has been microcopied for the years 1914–1918 inclusive—a total of almost 90,000 pages; the set of films is on sale for $.66 per 100 pages. Figures presented by the Recordak Corporation during the planning of this project indicated that that part of the cost which is represented by the printing off of the positive copies ∣📄 p.130 (not including price of the film) is figured at $.04 per foot.

In estimating processing costs of the long rolls of film, account must be taken of the possibility that this work may be done by the individual or institution which does the photographing. The large, continuous process machines used in the moving picture industry would not be practicable for such work, but a Stineman developing reel, taking 100 feet of film, can be purchased for $32.00; the chemicals for developing 100 feet of 35mm. film cost about $.10. Automatic printing machinery for the 100-foot film strips is also available, and is not too bulky to be part of the equipment of a photographic department of a library.

Processing costs, like reduction ratios, introduce a very significant variable into microcopying. The lower priced processing may or may not be inferior to the more costly. The effect upon costs is illustrated by taking the figures on film footage required for 100 pages of a 6” x 9” book, as given above, and figuring in with them the high and low processing costs. Note that high processing costs can make it more expensive to copy at the reduction ratio of 21 diameters on 16mm. film than to copy with lot processing costs at a reduction of 12 diameters on 35mm. film. The gains in cost of materials that may be made by increasing the reduction ratio may conceivably be frittered away if the high reduction ratio renders necessary a more expensive processing. On the other hand,

∣📄 p.131 resort to low processing costs may entail a low reduction ratio, and hence the saving in processing may be sacrificed by in- creasing the material cost.

Just as in calculating typescript book costs it was necessary to assume arbitrarily an average typing charge before proceeding to further comparative studies, so in calculating microcopying costs an arbitrary average processing figure will be assumed. The difference between the price of developing a negative and the price of making a positive print from a negative will be ignored; it does not appear in most quotations. Processing will be figured regularly at $.0175 for 16mm., $.025 for 35mm. film; the joint cost of film and processing will stand then at $.025 for 1é6mm., $.04 for 35mm. film. It is under- stood that these prices are available only in large footages, running into the thousands of feet. Maximum figures for shorter runs, when necessary, will be assumed at $.03 for 16mm. film, $.05 for 35mm., to include film and processing. Very short runs of a few inches only introduce a special problem which will be separately analyzed.

First Costs and Running Costs in Microcopying

The application of microcopying to the distribution of reading matter is so new, the price levels at which it operates so revolutionary, that the fundamental elements of cost as they appear in publishing are not apparent at first glance, and yet they are present. In microcopying, as in other techniques, there is the antithesis between a cost curve that falls with each additional number of copies and a cost curve that remains constant regardless of the number of copies that are made. In other words, there is a first cost and a running cost. The first cost includes the labor of making a photograph and processing the negative film; the running cost is the cost of making each additional positive from the negative.

There may arise many situations in which there will be no positives made from the negative. The original negative will be made as a unique copy for a single purchaser. The cost of this unique copy will consist of two elements: first, labor in photography; second, film and processing. ∣📄 p.131 This situation will be studied first, and then its implications traced in those cases where a number of positive copies are to be made from a single negative for a number of purchasers. The cost of making this negative stands, however, as a “first cost,” comparable, for instance, with the cost of setting type and fitting it in a press, or preparing mimeograph stencils or photo-offset metal sheets. It is absolutely independent of the number of copies that are to be made from the negative.

How is this cost to be estimated in the case of long runs of material? As in so many other situations, the most expensive and specialized apparatus reduces the labor cost. The small, non-specialized cameras such as the Leica or Contax correspond to hand-set type in printing, and the large, highly-specialized cameras, such as the Recordak, Filmograph, Kennedy, and perhaps the new Graflex, correspond to the linotype in efficiency.

Experience in photographing large collections of documents has been most amply obtained by the Filmograph Corporation and by Lloyd B. Kennedy in copying land title abstract records, and by the Recordak Corporation in a variety of copying enterprises.

Filmograph and Kennedy experiences agree in establishing that one operator can copy 10,000 pages of bound records per day; Dr. T. R. Schellenberg, using a Recordak to copy the unbound typescript pages of the NRA and AAA hearings, also copied 10,000 pages a day. In all these cases there was material in hundreds of thousands of pages, uniform in size, continuous in supply. The May Company of Cleveland uses a Recordak to copy its internal records. The operators photograph 20,000 statements, 7” x 10”, per day. Speed records with the new Recordak bound book copying camera are not available. It is safe to conclude that experience with uniform and continuous matter, whether bound or unbound, permits an estimate of 10,000 pages a day as a maximum.

While these automatic machines can be operated by relatively unskilled labor, it is not too much to count the labor cost at $10.00 per day, which might cover the wages of two operators, one to do the photographing, the other to bring the material to the machine and return it to place.

This labor cost is independent of ∣📄 p.131 the reduction ratio and hence of the amount of film used in a given copying job. But with the film used at a reduction ratio of 16 to 23 diameters, and processing costs figured at their average for long runs, the cost of film and processing should stand somewhere near the labor cost—perhaps $10.00 worth of labor will photograph $12.00 to $15.00 worth of film (including processing). At this rate the first cost—i.e., the finished negative—of microcopied material in large runs can be made for $.22 to $.25 per 100 pages. This is the cost at which a unique copy of a great body of material could be made for a single purchaser.

But if a number of purchasers should wish to acquire the same material, the cost of making the positive copies for each would be exactly the film-and-processing charge for film plus a pro-rata fraction of the first cost divided among the purchasers. The behavior of costs in this situation can be illustrated from two cases: first, the actual experience of the Joint Committee on Materials for Research in distributing microcopies of the hearings held under the NRA and AAA; and second, the estimated costs of a project for copying two million pages of books printed in England before 1640—a project that would bring to subscribing libraries copies of all books listed in Pollard and Redgrave Short Title Catalogue.

The Joint Committee arranged with the Agricultural Adjustment Administration and the National Recovery Administration to make microcopies of the hearings. The total number of typescript pages was approximately 315,000. The reduction ratio was 23 diameters. The cost of making the negative, including Recordak rental, film and film developing was $455.00 for 13,400 feet of 16mm. film. The cost of making additional positives was $368.50 each. The whole project was figured to distribute the cost of making the negative among ten subscribers, making the cost of the microcopies of the 315,000 pages $413.50 or $.13 per 100 pages.

The work of microcopying this great body of material was carried to a successful conclusion by Dr. T. R. Schellenberg, in spite of many obstacles. The material consisted of unbound typescript pages, some of them carbon copies on thin paper, to∥ gether with a number of printed pamphlets. When the material was first filmed, the operative employed to feed the pages skipped some of them or fed them into the Recordak machine in an irregular way. Moreover, the machine itself was not set to take care of the photographic differences between light and dark copies, or between bond and onionskin paper. When a check up of the first film copy of the NRA records revealed imperfections, an effort was made to correct them by cutting the microcopy strips and interspersing freshly photographed strips of film in the faulty places. But this film strip, when used for making positive copies, did not prove satisfactory. Dr. Schellenberg therefore recopied the whole file of NRA hearings, with the technical help of members of the staff of the Recordak Corporation, which was very generous, not only in solving technical difficulties, but also in contributing free of charge the machine and film for the second copying. In calculating and assessing the cost of the project, the extra copying operation was not taken into account, for it was the result of conditions which need not be duplicated in the future, and is charged off to experiment and experience.

A project to make widely available copies of all early printed English books has been under discussion among librarians for several years, and has led to a number of cost calculations. The project differs from that of copying the AAA and NRA hearings, not only in its greater magnitude, but also in the greater expense that must necessarily be incurred in travel and supervision, and in setting and resetting the camera to take volumes of different sizes at different reduction ratios. A memorandum on cost was drawn up at the office of the Joint Committee in September, 1935. For purposes of calculation it was assumed that the total number of pages would be two million, and the average page size an octavo in twelve-point type. The calculation was made to cover two alternatives in reduction ratio—12 diameters and 21 diameters, and the alternative of using 45mm. or 16mm. film. The memorandum on cost, while rather elaborate, serves to illustrate the behavior of so many of the variables in microcopying that it is here quoted textually:

The first cost of making an original ∣📄 p.132 negative of the estimated two million pages is hard to calculate and not greatly dependent on reduction ratio, for so much of it will be travel, supervision, apparatus and labor of incalculable efficiency. These costs, exclusive of film cost, could be calculated for two million pages as follows:

Other costs are variable with the number of feet of film used. On 35mm. film in the third position, reduced 12 diameters, 2,000,000 pages (750,000 linear feet of book) would go on approximately 62,500 feet of film plus 10% for margin, bringing the total roughly to 70,000 feet. On 16mm. film in second position, reduced 21 diameters, 2,000,000 pages (1,000,000 feet of book) would go on 48,000 feet of film plus 10% for margin, bringing the total to 56,000 feet. The cost for the film and the processing for 35mm. film, if figured at $.04 a foot, is $2,800; for the 16mm. film, figured at $.025 a foot, it is $1,325.

Resuming our calculations, the difference in reduction ratio and film cost becomes increasingly important as the number of subscribing libraries increases. The possibility will also appear that very substantial reductions in the processing cost of the film may become available if the footage runs high enough. However, maintaining the constant calculation of $.025 for the 16mm., $.04 for the 35mm. film the costs for subscribing libraries would stand as follows:

By the time forty libraries are ∥ reached, the cost of film and processing has become dominant and any economies in these costs would be felt immediately.

The effect of differences in processing costs when the number of subscribers reaches 40 is shown by recalculating the costs at maximum and minimum processing figures ($.055 maximum and $.025 minimum for 35mm. film; $.05 maximum and $.0225 minimum for 16mm. film).

Microcopying “Reprint Costs”

Microcopying, like book production, as has been stated, gives rise to both “first cost” and “running cost.” The first cost in microcopying is compounded of two factors, the first of which is largely labor, and is dependent upon the number of pages copied, regardless of the area of film used; the second is film and processing, and is dependent upon the area of film used, regardless of the number of pages copied upon it. The labor cost of photographing is roughly analogous to the composition cost of book manufacture; it varies with the number of pages rather than the number of words. The running cost (that is to say, the cost that depends upon the number of copies) is simply the area cost of the first copy, multiplied by the size of edition.

Another kind of cost is present in book manufacturing, the “reprint cost.” This is the “first cost” of bringing out a second edition. The reprint cost is less than the first cost of the first edition of a book provided plates are made from the type, or the photo-offset metal sheets ∣📄 p.134 preserved for a second run. But the reprint cost in book manufacture, regardless of these economies, is always substantial. In microcopying large footages the reprint cost is nil. No special expense is incurred in production if, when 10,000 feet of film have been distributed to each of ten subscribing libraries, an eleventh library comes along a year later and asks to have an eleventh copy made for it. The cost of making the eleventh copy is simply the normal film-and-processing cost. For this reason microcopying offers remarkable elasticity in the distribution of material. There need never be an “overprint,” made at the risk that no one will wish to buy it. This quality of microcopying, it must be remembered, is present only to the degree that the footage involved in a single order is large enough to obtain low processing prices.

Short Footages in Unique Copies

An application of microcopying that is already developing, and promises to have great utility in the future, is the making of microcopy negatives of books or periodical articles upon the order of individual purchasers, whether research workers or libraries. The installation of efficient photographic equipment may make it possible for a large library to make its resources generally available to readers anywhere at the price level of microcopying.

If the number of pages to be copied per day in a library should be large enough to allow the continuous operation of the machine, the operating cost would rise above that of a long run of uniform material only to the extent that readjustments of focus became necessary in changing from books of one page size to books of another ‘Size. The processing costs could still stand at the level of long rum work. An additional charge would be incurred in cutting up the strips and mailing the short lengths to the various purchasers, but this type of service, in the experience of the Recordak Corporation (which cuts up long strips of bank records and mails them to individual banks) would not lift the costs to an entirely new level. Bookkeeping costs would appear, but it seems quite probable that on large-scale work the price level could be held down within ∥ striking distance of the price of long-footage orders.

Microcopying and Long-Distance Use of Libraries: the “Bibliofilm Service”

The most interesting effort to develop a service of this kind has been made by Mr. Atherton Seidell, Dr. R. H. Draeger of the U. S. Navy Department and Miss Claribel Barnett, librarian of the Department of Agriculture Library, Washington, D. C. In November 1934, Miss Barnett established the “Bibliofilm Service” in the Library of the Department of Agriculture. The date is an important one in the history of microcopying applications.

Miss Barnett has described the administrative routine of the Bibliofilm Service in an article in The Camera vol. L, no. 5, May 1935, pp. 327–328, as follows:

The procedure begins “with the suppliance of a 5” x 5” slip or card specifying the article desired—one such slip for each and every article, and each slip carrying the name and address of the customer. The slips are returned with the finished films. Attendants take the slips, search the library, send the books to the photographing room, there each article is copied on a continuous strip of film, leaving spaces between for cutting. When all the day’s orders are photographed, the film will be developed, fixed, washed and dried, then cut up and each sequence attached to its proper slip and mailed to the customer or delivered by hand as it may be. Prepayment for the service is urged when possible, this to save accounting, etc., and a deposit of funds which may be drawn upon is possible.

The film copies or “filmstats” as they are called are priced at $.10 for each article not exceeding ten pages, and $.05 for each additional ten pages. The increasing use of the Bibliofilm Service is evident from a report in Agricultural Library Notes for June, 1935; the number of pages copied per month rose from 470 in November 1934, to 6,597 in May 1935.

Miss Barnett writes that the rates during this experimental period have not covered the cost of keeping the records ∣📄 p.135 and correspondence, but that the clerical cost may be materially lessened after the procedure has been better established. Moreover, the clerical costs as they stand are not greater, she believes, than those connected with the lending of books, for such costs as charging and discharging loans, or following up loaned volumes by correspondence to secure their return, are avoided. Studies of the cost of interlibrary loans made by the University of California—though challenged by some librarians—set the cost of each loan at well over a dollar, and offer strong confirmation to the views of Miss Barnett. Scholarship stands much indebted to her for the well-conceived initiative.

Other libraries have installed equipment for making microcopies, but the cameras used do not have the efficiency of the Draeger camera used by the Department of Agriculture Library. The price for microcopies at the Library of Congress, the New York Public Library, and Yale University Library is set at $3.00 per 100 exposures; at two pages per exposure, this amounts to $1.50 per 100 pages. Huntington Library at San Marino, one of the first to install equipment, charges $5.00 per 100 exposures, or $2.50 per 100 pages. A large microcopying machine has been installed in the Preussische Staatsbibliothek, but no quotations on service have been received by the Joint Committee.

Since installation of equipment in America has already reached three of the largest libraries in the country, it is safe to say that at this very moment microcopies of several million volumes of books and periodicals are already available to research workers at a price ($1.50 per 100 pages) not much in excess of the normal price paid for books. Since the Wilson Union List of Periodicals in American Libraries will inform any reader anywhere whether a given volume of a periodical is available at one or another of the libraries offering this service, scholars are now able to arrange with speed and certainty to have copied for them periodical articles not accessible in their own institutions.

Microcopying and Copyright

While these innovations have offered a mechanical means of increasing the ∥ working power of scholars throughout the country, it has been necessary to prepare the legal ground for their use. For most of the learned journals from which scholars are likely to seek microcopies are covered by copyright. With books it is otherwise; most of the rare and inaccessible books are probably the old books, on which copyright has expired. But the whole question of photographic copying in its relation to copyright has been full of uncertainties.

Discussion with specialists in copyright problems showed that the law has been highly elaborated in court decisions to define the limits of so-called “fair use” when copyright material is republished for profit, but that no court decisions apply to the situation of the research worker who copies out notes from a copyrighted work in the course of his research. The letter of the statute would seem to indicate that the scholar with a fountain pen copying paragraphs out of an article in a learned journal was engaged in technical violation of the copyright statute, but it seemed evident that the courts would necessarily make a distinction between the kind of copying that is done in the course of research, when the copied matter merely enters a scholar’s note file, and the reprinting of material for sale or distribution. The librarians had felt for some time that the uncertainties of the law hampered them in some measure in laying down policies in the use of the photostat, and it was clear that the extension of microcopying, because of the lower price level, would increase the problem.

When the text of the new copyright law was under discussion in Washington in the spring of 1935, the suggestion was made that specific reference in the law should cover photographic copying for scholarly purposes; but the legal advisers of the publishers, while quite willing to have the photographic copying devices used to help scholars in research, felt that such a provision of the law would necessarily be so complex that it might either defeat its own purpose or undermine the whole structure of copyright.

The whole subject was explored in a generous and sympathetic spirit at a meeting of librarians and book publishers. Mr. H. M. Lydenberg of the New York Public ∣📄 p.136 Library, Mr. Milton Ferguson, chairman of the book-buying committee of the American Library Association, and Dr. Andrew Keogh of Yale University Library met with Mr. Frederic Melcher, chairman of the copyright committee, and the board of directors of the National Association of Book Publishers. As a result of correspondence and conference to which a number of librarians contributed, a “gentlemen’s agreement” on photographic copying for research purposes was drawn up and signed by Mr. W. W. Norton on behalf of the National Association of Book Publishers and by Dr. Robert C. Binkley on behalf of the Joint Committee on Materials for Research.

The main principles of the agreement are, first, that photographic copies of parts of books or periodical volumes may be made by a library and delivered without profit to a purchaser in lieu of loan or manual transcription. The paragraph was phrased to apply to the copying of a complete article from a periodical ∥ volume, but it would not apply to the microcopying of a whole volume or file of a periodical. The second principle is that the photographic copy is not to be a substitute for the purchase of the material; it can be taken for granted that the known practices of research scholars in the collecting of material for research will guide the intelligent application of this principle. The third’ principle is that the library which carries out the above principles in good faith is not to be held liable for violation of copyright law, but that responsibility for any possible infringement shall rest with the purchaser of the photographic copy. Subsequent correspondence has shown that some photographic copying problems arising in libraries are not explicitly covered by the agreement, but the main principles are sufficiently clear to justify the development of microcopying services in lieu of inter-library loans.

The text of the agreement is reproduced herewith:

Correspondence and Agreement on the Photographic Copying by Libraries of Copyrighted Material

May 25, 1935

Mr. W. W. Norton, President
National Association of Book Publishers
547 Fifth Avenue
New York, New York

Dear Sir:

On behalf of the Joint Committee on Materials for Research, I write to thank you for the consideration your organization has given to the increasing importance of the problem of conscientious observance of copyright that faces research libraries in connection with the growing use of photographic methods of reproduction. Not so long ago the scholar or student contented himself with pencil or ink transcripts of passages in the books and periodicals he consulted in connection with his studies. The increased use, first of the fountain pen, then of the standard typewriter, then of the portable typewriter, developed a corresponding increase in the number and amount of transcripts made for this same purpose. Within the past few years, the use of photography, the photostat, and now the 16 or 35 millimeter film cameras have added to the amount, the extent, the speed—and the problems—of this reproduction.

We realize, of course, that we can speak only on behalf of the publishers, libraries and research workers of this country. The problem is of importance also in connection with foreign copyrighted books, particularly the publications of academies or learned societies and the unofficial periodical ∣📄 p.137 press printing the results of research. As to such books and periodicals, the foreign publisher or copyright owner is the final authority. We feel, however, that the common interests of both groups would make it fair for us to assume that the practice approved by American publishers and libraries would satisfy their brothers in other countries.

We are told that from time to time instructors in various schools and colleges have reproduced in one way or another extensive portions or whole chapters from copyrighted textbooks, which have been sold or given to their classes as substitutes for the textbooks. We see how keen an injustice and how severe a harm is done to the owner of the copyright by such practices, and we participate with you in the condemnation of such obvious violations of copyright.

The indication you have given of what publishers and authors, as owners of copyright, can suggest to archives offices and museums as a guide to a code of fair practice in connection with the commendable form of public service rendered by photographic methods of reproduction is so clarifying that we suggest it be put in the form of an agreement and published.

Very sincerely yours,

(signed) Robert C. Binkley
Chairman, The Joint Committee on Materials for Research

May 27, 1935

Dr. Robert C. Binkley, Chairman
Joint Committee on Materials for Research
Western Reserve University
Cleveland, Ohio

Dear Dr. Binkley:

We deeply appreciate your desire to work out a code of fair practice which will protect the rights of authors and research workers. As publishers we naturally do not wish to impose restrictions which might hamper students in collecting research material, but on the other hand it is necessary for us all to face this problem realistically and not permit the rapid extension of photocopying to lead to a disregard of the fundamental principles of copyright.

The results of the conference in our office, attended by Mr. Lydenberg, Mr. Ferguson, and Dr. Keogh and members of our committee, have been discussed with Mr. Frederic Melcher, Chairman of the Association Copy- right Committee, and by our Board of Directors. We are happy to have the results of these conferences put in the form of an agreement and published with this correspondence.

Very sincerely yours,

(signed) W. W. Norton, President
National Association of Book Publishers

∣📄 p.138

The Joint Committee on Materials for Research and the Board of Directors of the National Association of Book Publishers, after conferring on the problem of conscientious observance of copyright that faces research libraries in connection with the growing use of photographic methods of reproduction, have agreed upon the following statement:

A library, archives office, museum, or similar institution owning books or periodical volumes in which copyright still subsists may make and deliver a single photographic reproduction or reduction of a part thereof to a scholar representing in writing that he desires such reproduction in lieu of loan of such publication or in place of manual transcription and solely for the purposes of research; provided

(1) That the person receiving it is given due notice in writing that he is not exempt from liability to the copyright proprietor for any infringement of copyright by misuse of the reproduction constituting an infringement under the copyright law;

(2) That such reproduction is made and furnished without profit to itself by the institution making it.

The exemption from liability of the library, archives office or museum herein provided for shall extend to every officer, agent or employee of such institution in the making and delivery of such reproduction when acting within the scope of his authority of employment. This exemption for the institution itself carries with it a responsibility to see that library employees caution patrons against the misuse of copyright material reproduced photographically.

Under the law of copyright, authors or their agents are assured of “the exclusive right to print, reprint, publish, copy and vend the copyrighted work,” all or any part. This means that legally no individual or institution can reproduce by photography or photomechanical means, mimeograph or other methods of reproduction a page or any part of a book without the written permission of the owner of the copyright. Society, by law, grants this exclusive right for a term of years in the belief that such exclusive control of creative work is necessary to encourage authorship and scholarship.

While the right of quotation without permission is not provided in law, the courts have recognized the right to a “fair use” of book quotations, the length of a “fair” quotation being dependent upon the type of work quoted from and the “fairness” to the author’s interest. Extensive quotation is obviously inimical to the author’s interest.

The statutes make no specific provision for a right of a research worker to make copies by hand or by typescript for his research notes, but a student has always been free to “copy” by hand; and mechanical reproductions from copyright material are presumably intended to take the place of hand transcriptions, and to be governed by the same principles governing hand transcription.

In order to guard against any possible infringement of copyright, however, libraries, archives offices and museums should require each applicant for photomechanical reproductions of material to assume full responsibility for such copying, and by his signature to a form printed for the purpose assure the institution that the duplicate being made for him is for his personal use only and is to relieve him of the task of transcription. The form should clearly indicate to the applicant that he is obligated under the law not to use the material thus copied from books for any further reproduction without the express permission of the copyright owner.

∣📄 p.139 It would not be fair to the author or publisher to make possible the substitution of the photostats for the purchase of a copy of the book itself either for an individual library or for any permanent collection in a public or research library. Orders for photocopying which, by reason of their extensiveness or for any other reasons, violate this principle should not be accepted. In case of doubt as to whether the excerpt requested complies with this condition, the safe thing to do is to defer action until the owner of the copyright has approved the reproduction.

Out-of-print books should likewise be reproduced only with permission, even if this reproduction is solely for the use of the institution making it and not for sale.

(signed) Robert C. Binkley, Chairman
Joint Committee on Materials for Research

W. W. Norton, President
National Association of Book Publishers ∥

“Publishing” Short Runs of Material by Microcopying

One of the most intricate, and at the same time the most important, of the problems encountered in the use of microcopying is its possible substitution for publication.

In its simplest form, such a substitution for publication would take place automatically if a library which offers microcopying service has in its possession unpublished material, and the probable purchasers of such material are aware of its existence and availability through the usual bibliographical guides. For instance, if a typescript copy of a monograph should be deposited in the Library of the Department of Agriculture, and listed or given notice in the learned journals, the Public Affairs Information Service, and other bibliographical guides, it would be, in effect, “published” to the degree that scholars would be willing to buy and use microcopies of it.

But in such a case each individual purchaser would require that the book be photographed anew. A new negative would be produced for each “sale” of the book.

Would it be possible to organize the distribution of microcopies of such a work in a way that would take advantage of the labor-saving that can occur when a number of positives are made from a single negative? Of course, if the material in question fills a large enough film footage, the additional copies after the first neg∥ ative are available with present technique at a price lower than the cost of the first negative. But in the case of a short strip of film—a few inches or a few feet—there is at present no way of taking advantage of the theoretically low cost of making a number of prints from a single negative in the production of positive prints to order. What is the amount and behavior of such anticipated economies? What is the difference between the cost of making a negative microcopy from a book and the cost of making a positive from a negative? And what is the cost of maintaining a catalogue, and shelving or storage system for film negatives?

It has been shown that, for long footages, the cost of making a positive from a negative is not greater than the processing cost of the same footage of negative. Even short items can be produced as negatives at the low processing cost of the long footage, because a large number of items are copied on a long roll, and the roll subsequently cut up. But if positive copies are to be made from the short scraps of negative, the. processing cost rises. Short scraps of negative give rise to high costs in making positive copies. But the cataloguing and storage expenses for short pieces are as high, per item, as for long strips. Consequently for short items it will be wasteful to save the negative, for long items wasteful not to save it. The dividing line is probably the 100-foot reel, which may mean a unit of 2,000 pages.

∣📄 p.140 For the shorter items—and this means for most scientific literature—present apparatus will not justify the expense of holding a negative for the sake of making a number of positives from it. But if a large number of microcopies of a scientific article are to be made from the same paper copy, the paper itself may wear out in the repeated photographing. “Publication” by microcopy methods would put upon a book the same amount of wear that would take place if every reader used the same copy. Librarians who have administered “reserve book rooms” for large university classes know how quickly a book will wear out under constant use.

This is the problem to which Dr. Watson Davis of Science Service has devoted a vast amount of intensive thought and planning. The solution suggested by him is an important contribution to the use of microcopying, and to the organization of the distribution of scientific literature. He proposes that the negative made from an original typescript, or indeed a negative made from a published article or book, should take the form, not of a length of film strip but of a rectangle of film, suitable for filing in a card catalogue. It is deemed technically possible to put 150 pages of microcopied material on a 3” x 5” card (a 70mm. film cut into 5” strips would measure 2 ¾” x 5”). Even if the number of pages that can be copied on such a card is materially less than 150 pages, the essential value of his suggestion lies in its answer to two of the most difficult problems in microcopying short items. For a short strip of roll film is not easy to store, it must be placed in a small compartment or container, the container labelled, and then the collection of containers and films controlled by a catalogue. If the film copy of the article or book can become itself the catalogue card, the problem of cataloguing and storing disappears. This is the first advantage of his plan. The second advantage appears in connection with processing. To make a positive print from’ a short strip of motion picture film is more difficult than to make a print from a single rectangle of film. His plan should tend to make “reprint costs” for short items behave as if the items were long, and at the same time eliminate a heavy cataloguing and storing charge. ∥

It has been suggested by several writers that the future may see whole libraries copied on film, and the negatives stored in such a way that positive copies from any of them can be run off on demand. It is sometimes thought that when a library, in responding to a demand for a microcopy of an item, makes a negative copy, it should make a positive from the negative, send the positive to the purchaser, and retain the negative to make positives for other purchasers. This is the procedure that the technique of photography suggests as normal. But how does it fit into the practical administration of library re- sources?

In the analysis of library costs, it was established that cataloguing and shelving of books are items of library expense that approach in magnitude, or even pass, the cost of the books themselves. Microcopying in lieu of inter-library lending, when organized like the U. S. Department of Agriculture Bibliofilm Service, imposes no extra cataloguing charges on the library. The film service simply uses the existing catalogue and shelving system. If an effort is to be made to use a negative more than once, the library must control its negatives by means of a catalogue system that will parallel its book catalogue, growing with the growth of the collection of negative films. This extra cost can be made up only if it results in substantial economies.

Dr. Davis has received from the Chemical Foundation a grant which he will use to develop the apparatus necessary to put his plan into effect. When the apparatus is ready, the day will have arrived when microcopying is technically suited to do for the intellectual world what publishing has hitherto done, with immeasurably greater elasticity and at a fraction of the cost. The long roll of film will probably remain as the best medium for the microcopying of long items, because it will do for the microcopies what binding does for a book—keep the series in its fixed order. The card negative will serve for the short items, arranged in an open order, with new items placed between items already in the file.

The Problem of Standardization of Equipment and Reduction Ratio

A memorandum by Mr. Verneur E. Pratt ∣📄 p.141 of the International Filmbook Corporation, and a series of letters and memoranda by Dr. Watson Davis, call attention to the vital importance of standardization in the development of apparatus for microcopying. Mr. Pratt writes:

If certain rumors and actual reports are to be considered authentic, there is an alarming disparity in the use of films. Some are planning microcopies to be made on 16mm. film; others upon 35mm. film, still others upon widths not usually considered standard. The minifications vary from 12 to 24. Some projectors and cameras now being designed and readied for the market are adopting 35mm. film with the customary double row of sprocket holes. Others, with a view of salvaging more of the available area, are planning on the use of film on which but one row of sprocket holes is punched. A camera concern is developing a reading machine which uses no sprocket holes ….

Furthermore, there is as yet no standardization of titling, of separation of dockets, folios or pages. There is no uniformity of treatment in locating pages on the film—some photographing newspaper pages vertically but book pages horizontally. Some are ignoring frame lines and others attempting to squeeze a certain number of pages into the definite standard film size.

Unless this situation is immediately controlled through action of a standardization committee, this lack of uniformity is going to cost millions of dollars to correct. in the early days of electric lamps each manufacturer brought out his own specifications of the base. The result was that electric lamps had to be especially made to fit each manufacturer’s sockets. No one has been able to calculate the millions which this foolish lack of standardization cost.

The problem of standardization brings to the fore the important fact that microcopying involves not only the manufac∥ ture of the copies themselves, but also the distribution of reading equipment. The cheaper the reading equipment, the more easily will it be distributed to a number of scholars sufficient to render practical the large-scale utilization of microcopying in research. But at the same time, the reduction in the price of reading equipment means an increase in the cost of microcopies, because reading machines that will handle the high reduction ratios that go with low-cost micro-films will be more expensive than reading machines that handle only low reduction ratios. The manufacturing cost of reading equipment increases with the reduction ratio of the microcopy it is prepared to handle, and with the area of reading matter laid before the reader at a time. The magnifying power, field of vision, and the light intensity (which determines the amount of contrast in the projected image) are the three optical elements of the reading equipment which seem to have the controlling position in the determination of costs of equipment. There are other important points to be noted in reading equipment, such as ease of loading the film, or ease in passing from one page to another, but these can be ignored for the moment. They are independent of the choice of magnifying power, field of vision and light intensity.

As soon as the cost of reading machines is introduced into the analysis of microcopy costs, the remarkable fact emerges that some of the fundamental elements of publishing accountancy are still present, but with the situations of the publisher and reader reversed. If a reader buys for $100 a reading machine which he uses to read 1,000 pages, he is paying $.10 a page for reading equipment; if he gets 100,000 pages of reading matter, the cost per page of his reading equipment is only ⅒ of a cent. His reading machine is a “first cost,” and the footage of film he acquires or uses in it a “running cost.” If he cuts this first cost by purchasing a projector that will not handle microcopies of 23 diameters reduction, but will accept only microcopies of 9 or 12 diameters reduction, his running cost of reading matter will rise because he must pay more per page for the microcopies. If he cuts his first cost by contenting himself with a machine that will show only a small area of projected image ∣📄 p.142 at a time, or give a lower light intensity, he may retain the low running cost of his reading matter, but must accept a sacrifice in the legibility factor.

The situation may be clarified by an illustration taken from transportation. The railway (like publishing in printed form) imposes a heavy first cost in making every trip. But this cost per passenger will fall with the number of passengers, exactly as the cost per reader of printed books falls with the number of copies sold. The railway company makes the initial investment, the passenger pays only for what he uses, just as the publisher carries the initial cost and risk of publishing, and the book buyer limits his investment to what he can actually use. The system will only be economically justifiable if a large number of people wish to go to the same place at the same time, just as publishing is only economically feasible if the saleable edition is large.

The quantitative relationship between the higher cost of projectors to read film at high reduction ratios and the lower cost of the film itself is easily established from the calculations made above of the cost of microcopying the English books printed before 1640. Basing this calculation on the estimates of the cost to the libraries of film when forty libraries purchase copies, two quantities can be plotted against each other on a graph. On one axis is the cost differential between high power and low power projectors; on the other axis is the cost differential between high reduction and low reduction film. The figures for the film are expressed in thousands of pages, the figures for the projector in dollar differences between alternative reading machine prices. Every 30,000 pages will justify an expenditure of an additional $25.00 for a reading machine.

The passenger has an alternative: he may buy a car. In that case he will have a fixed cost per mile for gas and oil, like the cost per page of books purchased in microcopy form. This cost will be constant, no matter how much he travels. He can go where he will, whether anyone else is going there at the same time or not, just as the microcopy reader can get any item in microcopy form, whether other readers want it or not. He is limited only by ∥

the road network, as the reader of micro- copies is limited only by the resources of libraries offering microcopying service; and in this respect both car owner and microcopy reader are users of public property. And the microcopy, like the automobile, if used on a large scale, may change the whole design of a compartment of life. But the development of microcopying in lieu of, or as a supplement to, printed book publication, like the development of the automobile in its relation to railway travel, will attain its technological possibilities only to the degree that a wide market cuts the first cost of the reading equipment, the individual owners of it distribute this first cost over a large footage of film, land the libraries offer the services that render both of these conditions possible.

This comparison brings) to light at once the tremendous difficulty and the supreme importance of standardization at an early stage in the development of the new devices. Let us imagine, as a kind of a ∣📄 p.143 fantasy, that the automobile in its early stages had been some kind of a machine that required for its operation a special kind of road. Then let us suppose that in one part of the country roads were under construction that would serve for one make of automobile, while in another part of the country roads were built for another make of car, with road and automobile so related that the machine built to go on one road would not go on another! That is the kind of problem facing the intellectual world at this moment in the standardization of microcopying. Shall we build a projector that will read any microcopied film, whether card or film strip, whether broad or narrow, whether a high or low reduction ratio? Conceivably such projectors can be designed, but will not their cost be so high that wide distribution will be forestalled? Or shall we have several makes of projector, each adapted to a particular kind of microcopy film, in which case the utility of each to the reader will be limited even though the price be less. Or shall libraries and others standardize their microcopying in such a way that one simple projector will serve for all purposes, in which case there is a risk that the chosen standard may not be the best, and almost a certainty that if it is well adapted to one type of material (as a long file of newspapers) it will be ill adapted to another type of material (as a short typescript monograph).

There are at present too many unknowns to permit the solution of this intricate equation. But the facts and principles already set forth will serve at least to eliminate some of the unknowns, and to simplify in some measure the problem of standardization.

Essentials of Design for Reading Equipment: (a) Length of Line

Into this labyrinth there are many approaches. The analysis might begin with estimates of buying power of various categories of institutions and readers; with estimates of the number of pages likely to be copied and hence the comparison of the relative importance to be attached to film prices and equipment prices; with a calcu∥ lation of the probability that the essence of the matter lies in the short items on the one hand, long footages on the other; or with an anticipatory measurement of the proportions which various classes of reading matter, such as newspapers, books, manuscripts, etc., will hold in the corpus of microcopied material. But there is one seemingly trivial dimension which appears to the writer to occupy a crucial position in determining policy, and that is the length of the line of print, typescript or manuscript. This unit, it has been noted, always bears some relation to size of print of type, and also to size of page. Bad typography or writing may make the line too long for the print, but there will always be some limits to the variation. Small type is almost invariably set in shorter lines than large type, and if a large page is to be used with small type, the type appears in the form of columns on the page. The resolving power of a photographic emulsion and process is exhibited in the reproduction of a letter, the labor cost of making an original negative microcopy is a quantity that varies with the number of pages, the processing costs of film vary with footage, but the technological meeting point of microcopying and microreading devices is found in the line of reading matter.

Copying and reading devices are not incompatible though they differ greatly in the way in which they render the size of print or of page. A given piece of film, when put in one projector, may yield an image equivalent to 8-point type, in another projector an image equivalent to 16-point type. The same film might show in one projector a spread of two pages laid before the reader’s eye at once, in another projector one page only, and in a third projector only the upper half of a page. These differences would not render the projectors incompatible with each other, nor with the microcopying camera. But if a camera makes a microcopy in such dimensions that the projector will display from it an image that does not show full lines, the two are absolutely incompatible.

Already there have appeared in the development of microcopying apparatus several outstanding examples of this incompatibility. The Recordak newspaper projector projects an image from a full frame of ∣📄 p.144 a 16mm. film or from one quarter of the “double frame” of a 35mm. film. A newspaper page, copied on the Eastman Recordak machine, is projected and read, a quarter-page at a time. There is no incompatibility here, for the quarter-page of the newspaper holds several columns, and no lines of reading are broken off in the middle. But the Library of Congress microcopies of manuscripts from foreign archives have been made to the number of over a million pages on 35mm. film, and most of them in the “first position” noted above, i.e., one page to a double frame, with the lines running full across the film. These film copies are incompatible with the Recordak projector. The Carnegie Endowment for International Peace, acting on the advice of the writer, made copies of the Paris Peace Conference papers of Colonel House at the Yale University Library. These were copied in the same position as the Library of Congress documents from foreign archives, and are equally incompatible with the Recordak projector. The NRA hearings were copied on 16mm. film in the same position, i.e., with lines running full across the film, but since the full width of the film is only half the width of the 35mm. film, the Recordak projector will display a full page of the typescript, not cutting any lines.

The present tentative design of a projection apparatus to be made by the International Filmbook Company seems to show that less than the full area of a book page is to be projected. About a dozen lines of print appear on the screen at a time. Then the film is moved to bring the remainder of the page into view. The projector is compatible only with film that has the lines running transversely across the film strip, as in positions 1 and 3—or indeed in the multiple arrangement suggested by Mr. Pratt. (see above.) To conform to this type of reading apparatus, Mr. Pratt recommends

locating all pages vertically on the film, with the lines of the printed matter at right angles to the length of the film. This will permit of simple projectors without revolving objective lenses and gates.

Another projector designed by the International Filmbook Corporation will ∥ permit the reading of lines running in either direction, and will expose a variable film area by substituting lenses and condensers.

The Bibliofilm Service photographers, using the Draeger camera, copy their pages in the second position on 35mm. film. If they should copy in the first position, their film could not be read in a Recordak projector; the copies they are making in the second position cannot be read in the International Filmbook Teledex, but can be read in the International Filmbook Optigraph if the rolls are long enough; if they should copy in the third position their film would be readable in either the Recordak or Filmbook projector.

Essentials of Design for Reading Equipment. (b) The Page Rectangle

Another given quantity, independent of any choice made by the designers of microcopying or reading equipment, is the general proportions of the normal page of reading matter. Ordinarily the page, whether of typescript, manuscript or print, sea rectangle of which the vertical dimension is greater than the horizontal by a factor of 25% to 50%. 8½” x 11”, 6” x 9”, and other standard sizes from folio to duodecimo carry these proportions. The proportions are independent of the size of the page, and equally independent of reduction ratio in photographing or enlargement ratio in reading. They are related to length of line, because the line always corresponds to the short side of the page rectangle (except for pages set in more than one column).

Taken in conjunction with two other fixed elements of the problem, namely, (1) the standardization of the width of the film and (2) the advantage of using the full area of the film the page rectangle, the proportions of the page rectangle become important determinants of design. This fact becomes especially noticeable when a fixed frame camera is used for copying.

The ordinary moving picture frame on 35mm. film is a rectangle 1” x ¾”, with the long side across the film; the “double frame” used in the Leica, Contax, and Ludwig cameras exposes an area 1” x 1 ⅜”, or 1” x 1½”, with the short side across the film. Edwards Brothers of Ann Arbor is prepared to furnish microcopies of ∣📄 p.145 European books made with a single frame camera; the equipment at Yale, New York Public Library, and Huntington Library, is of the double-frame type. With a single-frame camera, pages can be copied in the second or third position without loss of film area; with a double-frame camera pages can be copied in the first and second position without loss. But if a single-frame camera is used to copy pages in the first position, or a double-frame camera to copy them in the third position, from one-third to one-half of the area of the film is wasted.

These considerations do not apply to copying cameras with changeable frames, but so much copying equipment is already set up with fixed frames, and especially with the fixed double frame, that any standardization of projectors that would force users of double-frame cameras to waste from one-third to one-half of their film in order to render their copying compatible with reading equipment would incur disadvantages in distribution. This fact, ∥ translated into terms of design, seems to call for projectors that can be adjusted to read lines of print that run across the film as well as lines that run along the film.

This element of the design of a projector is not a part of the optical system, but of the mechanical structure of the apparatus.

So far as the optical system is concerned, the important fact to notice is that pages always appear as rectangles, of which the short side is the line of print. Two-page rectangles may be photographed on one film rectangle, but the initial proportions of the page rectangle will still control the proportions of the rectangle of film used in copying them. If the camera has a fixed frame area, either single-frame or double-frame, the proportions of the film rectangle are determined by the camera. But even if the camera is adjustable and can expose a larger or smaller length of film at a time, the proportions of the page will still control the proportions of the film

∣📄 p.146 image, which will always be a rectangle with the long side one-third to one-half longer than the short side.

Essentials of Design for Reading Equipment (c) the Aperture

The aperture of a lens is usually measured as a diameter of the lens circle: but in connection with the problem of projecting pages it must be given another meaning. The aperture dimension that is effective for reading microcopied pages is the length of the long side of a rectangle with the proportions of a page (roughly 1” x 1½”), that can be inscribed in the lens circle. The word “aperture” as used below is a lens area covering a rectangle of film measured along its long side.³ The aperture thus defined need not correspond in size to the page rectangle on the film, nor to the film rectangle which holds the image of one or two pages, nor the single frame or double frame that has its dimensions set by the camera. But in the ratio of its long side to its short side it will still follow the general proportions of a page. The aperture is to the projector what the frame of film is to the camera.

side of the aperture rectangle (a) and the diameter of the lens circle (d) will always be so related that d = 1.2 x a. An aperture of 1” goes with a lens diameter of 1.2”; an aperture of ½” with a lens diameter of .6”, etc.

It is possible to read a large frame of film with a small aperture if the film frame is moved in front of the aperture, so that the different parts are projected successively. This is true whether the frame of film has been used to copy one or two or even more pages. If the long side of an aperture equals the long side of a page rectangle, the page can be read without moving the film; if the long side of the aperture equals the short side of the page rectangle the page can be read by moving the film in front of the aperture; if the long side of the aperture is less than the short side of the page rectangle, the film and projector are incompatible, and the page cannot be read at all, (unless, indeed, the page happens to be set up on columns, and the long side of the aperture is equal to the width of the microcopied column).

To consider an extreme case; two pages can be photographed (in the third position) on a single frame of film. One page of the original print will correspond ∥ to half a frame of film. Its long side
will be ¾”, its short side ½”; the length of line will be ½”. The aperture required for reading this film need be only ½” (½” x ⅜”) but in reading this film with a ½” aperture, the film must be moved successively into four different positions: first, to read the top of the first page; then to read the bottom of the first page; then to read the top of the second page; and finally to read the bottom of the second page. An aperture of 1” (1” x ¾”) would make it possible to read both pages without moving the film; an aperture of 1½” (1½” x 1”) would not only permit the reading of both pages at once but would also lay before the reader’s eyes, at the same time, two frames of film, with four pages of reading matter.

These principles are used in the Recordak newspaper projector. A newspaper is reduced sixteen diameters on 35mm. film. The aperture of the lens projects from the newspaper page rectangle (1” x 1½”) an area roughly ½” x ¾”. The film is moved in front of the aperture in such a way that the reader has before him at any time a projection of about a quarter of the newspaper page. Since the newspaper print is set in columns, this arrangement does not violate the principle that the aperture must be sufficient to exhibit the full length of a line. But in the case of the streamer headlines, running full across the newspaper page, this principle is violated, for the reader cannot have before his eye at one time a full streamer headline. This is an admitted inconvenience. It is overlooked in the Recordak design because a very small proportion of the reading matter of a newspaper consists of streamer headlines.

The dimensions of length of line and the proportions of the page rectangle, when considered in connection with the standardized 35mm. and 16mm. film define certain specific dimensions in aperture as fixed alternatives in design.

An aperture 1½” will display the long side of a page rectangle taken on a double frame of 35mm. film. An aperture of ½” will display the short side of a page rectangle taken on 16mm. film. The ∣📄 p.147 intermediate steps in the series are show on the chart on the following page.

These dimensions for aperture are final in the sense that there is no reason for building apparatus that will use an aperture intermediate between them. There is no use for an aperture of more than 1½” for 35mm. film; there is no use in making the aperture larger than ¾” unless it is made a full inch; there is no stopping point between ½” and ¾”. An aperture of ⅝” is wasteful; it can accomplish nothing that is not equally well accomplished with an aperture of ½”. These conclusions follow from the fact that the standard dimensions of the film and the normal proportions of a page rectangle are fixed quantities.

Essentials of Design for Reading Equipment: (d) the Enlargement Ratio

With these apertures accepted as fixed alternatives for reading equipment, it remains not only to decide between them, but also to decide upon an enlargement ratio.

The enlargement ratio has only two fixed points that can be deduced from other fixed elements of the situation. It should hardly be less than six diameters, for then it would overlap the possibilities of direct optical magnifying by reading glass equipment. It is shown above that the cost of photosensitive surface on film is about four times that of the paper surface. If a 6” x 9” page is to be copied in the first position on 35mm. film, with a reduction of 6 diameters, it could be copied at no greater materials cost on paper in a reduction of 3 diameters. The paper copy, legible with a reading glass, would offer marked convenience in storage and filing. Equipment for making such paper copies is in sight. It would seem, therefore, that no enlargements of less than six diameters need be considered. Possibly the same argument in favor of paper could be extended to cover enlargements of eight or even of ten diameters, but for the moment it can be concluded that there is somewhere a minimum enlargement below which projection equipment is unnecessary, and this minimum is certainly six, and possibly 10 diameters.

The upper limit of enlargement is dependent upon the range of the high reduc∥ tion ratios. If they are set at 16 for newsprint and 24 for typescript, then the reading equipment will follow by disregarding the possibility of enlargement in a higher degree. If the reduction ratio should be pushed up—as some experimenters propose—to 40 diameters for typescript, the projector will have to follow it. These considerations fix the upper and lower limits for enlarging power in reading equipment. Between them there are no fixed stations.

The enlargement ratio is, however, related in an important way to aperture. For instance, to take an extreme case, it is evident that a ¼” aperture in a machine that would make enlargements of only 10 diameters would have a very restricted use; the octavo book pages or typescript pages that would have their length of line brought down to ¼” would have to pass through a reduction of 24 or more diameters. The projected image of such a microcopy, if enlarged only 10 diameters, would be less than half the original size and might therefore be illegible. On the other hand, a machine with a 1½” aperture set to make enlargements of 20 diameters, while it would be useful in projecting a full newspaper page at once, would be wasteful in projecting microcopied books that had passed through a reduction ratio of 20 because it would show more pages than necessary, and would be equally wasteful in projecting book pages that had been enlarged only 10 diameters, because it would raise the size of the print to poster size.

There is another and equally significant relationship between aperture and enlargement ratio. The cost of a lens and condenser system increases with the enlargement ratio, but diminishes with the aperture. Thus an optical system with large aperture and low enlargement ratio will cost the same as an optical system with small aperture and high enlargement ratio. If the present design of a certain reading equipment renders it too costly, the cost can be reduced by redesigning it either for smaller aperture or for lower enlarging power. If reading equipment at present available has insufficient enlarging power, the design can be altered to give a higher enlarging power without. altering the cost, but only by diminishing. the aperture.

It might be argued that enlargement ∣📄 p.148

∣📄 p.149

• proofed to here

TABLE XLV

Aperture 6 diameters 9 diameters

Aperture 6 diameters 9 diameters

Aperture 9 diam, 12 diam.

∣📄 p.150 ratio can always be increased by bringing the lens nearer to the film, and setting the screen further from the lens. Thus a very cheap projector will make enlargements of 40 or more diameters by throwing an image on a distant wall. It does not follow, however, that a lens system devised to make enlargements of 10 diameters can be readjusted to make enlargements of 20 diameters, because the low power lens will not give the requisite definition, and the lighting system will not give the requisite contrast. The manufacturers of projectors will constantly be driven back to the alternative of sacrificing aperture for enlarging power, or enlarging power for aperture, or of sacrificing market if they standardize at a cost level that gives high enlarging power and large aperture.

The terms of the equation to be solved by the industry are now set. It remains only to discover exactly which alternative aperture-enlargement combinations are available at various cost levels. What are the combinations of aperture and enlargement that can be sold for $40, for $60, for $100? The figures arranged above, showing the immense effect of the high reduction ratios in cutting the cost of film will exert a pressure toward the adoption of high enlarging power with small aperture. The argument for low enlarging power and large aperture will turn on several points, of which the most important probably are:

1. That much equipment now in the hands of scholars is of the double-frame type, using 35mm. film, and with ordinary page copying this operates at the lower reduction ratios.
2. That the high reduction ratios in microcopying require a degree of skill that the scholar, an amateur in photography, often cannot supply. He can safely make mistakes in focusing, exposing, and processing at 10 diameters of reduction, through the same errors would be fatal if he were making reductions of 20 diameters.
3. That in short items made in unique copies, labor cost is so high in proportion to materials cost that the price advantages of high reduction ratios are illusory. ∥
4. That the large-scale ventures in copying, such as the NRA and AAA code hearings, or the two million pages of early English books, which make the high reduction ratios so advantageous, involve the expenditure of such large sums that they could sustain the manufacture of special equipment for reading the films made in such quantity.

It is certain that there is a clear need for some reading equipment, such as the Recordak, that will handle reductions of 16 or more diameters by making enlargements of 16 to 20 diameters. Should all reading equipment be adapted to this use, or should there be an alternative line of projectors suited only to the reading of material copied in the zone of 10 diameters? If all reading equipment is to be adapted to the reading of material copied in the high reduction ratios, should it be designed to meet a low price level by sacrificing aperture? The analysis given above may clarify the problem and thus assist in its solution, but it cannot establish a proved answer.

The four companies now most active in developing reading equipment are the Recordak Corporation, the Spencer Lens Company, the Folmer Graflex Company, the International Filmbook Corporation, and Dr. Draeger of the U. S. Navy. Leitz, Bausch and Lomb, and Spencer Lens have projectors that can be used for projection reading, though not specifically designed for it. Leitz has a combination enlarger and projector. The Graflex is working on a combination camera and projector. Spencer Lens Company and Dr. Draeger are using the single-frame aperture of one inch; the Recordak an aperture of ¾”; the International Filmbook Corporation is trying to keep the cost of apparatus down by keeping the aperture small, while Dr. Draeger and the Spencer Lens Company are working to keep the price down by contenting themselves with a low reduction ratio. The librarians in general, and especially Mr. Metcalf of the New York Public Library, and the promoters of microcopying in Washington, especially Dr. Draeger and Dr. Watson Davis, are working with the Bureau of Standards and with other agencies toward standardization. The conclusion is not yet clear, but the observer must testify to the clear ∣📄 p.151 intelligence with which those working in the field have envisaged the difficulty of the problem of standardization and have undertaken to grapple with it.

Conferences held in Washington, December 11 and 12, 1935, by various persons interested in the development of microcopying resulted in the following tentative proposals which were then circulated for review, and approved on behalf of the Joint Committee on Materials for Research by the Chairman and Executive Secretary.

I

There are three media of microcopying to be considered: paper, 35mm. film and 16mm. film. Since the use of paper does not involve the need for a projector, it can be considered separately, but the great convenience in filing paper microcopies as compared with film, and the relative cheapness of paper as against film, indicate that paper must not be ignored in the microcopying problem. Dr. Bendikson of the Huntington Library, Dr. Dallas Irvine of the National Archives, and the Folmer Graflex Corporation in the development of their new copying camera are among those who have most carefully thought out the possibilities of microcopying on paper.

II

Related to the distinction between the 16mm. and the 35mm. film, but not identical with the problem of the choice of film width, is the problem of reduction ratio on film for projection.

The International Filmbook Corporation and the Recordak Corporation and the Recordak newspaper copying enterprises contemplate the use of 35mm. film with a high reduction ratio (16 to 23 diameters, or over). Such high reduction ratios are also to be used on the 16mm. film. Lower reduction ratios are developed by users of the Leica camera and by the Bibliofilm Service of the Department of Agriculture, which uses the Draeger camera, especially in connection with the microcopying of short items. Any reading equipment can easily be equipped with reel-holding devices that will take either 16mm. or 35mm. film, and it should be a rule that designers are not encouraged to develop apparatus that will not permit the use of either width of film. In connection with the problem of enlargement ratio, a different situation exists.

∥

III

Projecting equipment, granted that it is adapted to both 16mm. and 35mm. film, will vary in the following essential particulars:

a. It may be fixed focus or variable focus. The Recordak equipment has a fixed enlarging ratio of 22 diameters. The projective Folmer Graflex equipment will have a varying enlargement ratio that can be set at will. Each of these methods has its advantages and disadvantages.

b. Some reading equipment throws the image on a screen from which it is reflected back to the eye; other equipment throws the image through a semi-translucent screen. The Recordak and Folmer Graflex will be of the first type. The translucent reading machines will be the International Filmbook apparatus, the Draeger projection reader, and possibly the machine put on sale by the Leitz Company, and possibly the machine to be developed by the Spencer Lens Company.

c. Reading machines differ in the size of aperture; that is to say, in the amount of film they will enlarge for reading at one time. The Leica projector takes a double frame of 35mm, film, 1” x 1½”. The aperture of the Spencer Lens Company projector now contemplated will be a single frame, ¾” 1”; that of the Recordak is one-quarter of a frame; and that of the International Filmbook Corporation may be even less. In general, there is a relationship between size of aperture and enlargement power, because if the aperture is only one-quarter of a frame it will not permit the reading of lines of print that have been microcopied to extend across the full width of 35mm. film.

IV

Material that has been microcopied on 16mm. film at a high reduction ratio can be printed off on 35mm. film at a lower reduction ratio; material that has been microcopied on 35mm. film at a low reduction ratio can be copied off on 16mm. film at a high reduction ratio. This fact is of vital importance because it permits the purchaser of a low power projecting apparatus to order copies of any long run of 16mm., high reduction ratio film (as the NRA and AAA records) in a form that he can read. Conversely, the purchaser of a small ∣📄 p.152 aperture, high reduction ratio projector can utilize material that was originally copied with a low reduction ratio on 35mm. film, if he has it printed off on the reduced size.

V

Taking into account the foregoing facts and possibilities, it would appear that the following developments in projection apparatus can be integrated into a common system and so arranged that the products of any microcopying enterprise could be made available to the projectors of any apparatus, with a few exceptions that will be noted:

a. Double frame aperture, low enlarging power—Leica (now available but costly).

b. Single-frame aperture, fixed focus, 12 diameter enlarging power, reading by transmitted light—under development by Dr. Draeger.

c. One-half frame aperture, 22 diameter enlarging power, fixed focus, reading by reflected light—Recordak apparatus.

d. Variable aperture, variable enlarging power, reading by reflected light—Folmer Graflex machine.

e. One-quarter frame aperture, very high enlarging power, fixed focus, transmitted light—International Filmbook Corporation development.

VI

Large-scale copying enterprises, such as the proposed microcopying of books published prior to 1640, can be made on either 35mm. or 16mm. film, either at a high reduction ratio (as 16, 20, 23) or a low reduction ratio (as 9 to 14 diameters). If the master copy is made in a low reduction ratio, the possessors of Recordak or International Filmbook apparatus would merely purchase their film in a reduction on 16mm. film. If the original is made at a high reduction ratio on 16mm. film, the possessors of the Draeger, Leica, and presumably Folmer Graflex apparatus would purchase it in the enlarged form on 35mm. film. The newspaper material now being copied by the Recordak Corporation in a reduction of 16 diameters can be read presumably on any one of four reading ma∥ chines: the Draeger projector will enlarge it back 14 diameters, at some loss of legibility, but it will show the full width of the newspaper page; the Recordak projector will show only one-quarter of a page, but enlarge it above original size; the International Filmbook machines will presumably show less than one-quarter of the area of the newspaper page, but may give it even more enlargement than the Recordak; Folmer Graflex can presumably be adapted to the reading of this type of film. Microcopies produced like the NRA and AAA series can be read presumably directly and without change on the Recordak, International Filmbook and Folmer Graflex equipment. For reading on the Draeger, Leica and Spencer Lens projectors, they would be reproduced for the purchaser in 35mm. width, making the reduction equivalent to 11½ diameters rather than 23 diameters. The short strips of film now being distributed by the Bibliofilm Service in reductions of 9 or 10 diameters can presumably be read on all projection apparatus except that of the International Filmbook Corporation, and doubtless if special instructions are given for the filming of such strips by the Bibliofilm Service they can be adapted even to the International Filmbook projector.

VII

Conclusion: it is submitted that this array of equipment will provide for the reading of any film now being used, or likely to be used, with present: techniques, except such copies made at very high reduction on multiple exposure on 35mm. film. Such copies would be legible only in the Recordak and International Filmbook apparatus.

Film Storage and Durability

To the extent that microphotography is used merely as a substitute for interlibrary lending, the question of the permanence of the film is not important. There is no doubt that its stability is sufficient for a span of years in which a scholar is carrying out a specific research project. His collection of material, in microcopy form, will not go to pieces on his hands.

But if microcopying is to take the ∣📄 p.153 place of publication, or if libraries are to alter their general acquisition policies to give a large place to film instead of paper, the fundamental question of durability of film must be met. “Dr. A. F. Kuhlman, in an article on the preservation of newspapers on film, has stated the issue comprehensively:

Before we are ready to preserve newspapers in film form, dependable and favorable answers must be found to a series of unsolved problems. First, the Bureau of Standards must settle the question of the life expectancy of the film record. The Bureau must determine: (1) the chemical and physical properties of the cellulose acetate film and of the emulsion and image produced upon it in filming records; (2) specifications for producing and processing the film to obtain the most usable and durable production; (3) the effect upon the film copy, while in use or storage, of various environmental conditions—temperature, humidity, light, human contact, etc.; and (4) specifications for storage and use. These points determined, it should be possible to predict, on the strength of verified data, the life expectancy of the film if produced, processed, stored and used under clearly specified conditions.”⁴

The Carnegie Foundation has made a grant of $5,000.00 to the Bureau of Standards to study film durability, and the research was under way in the year 1935. The research had progressed sufficiently by January, 1936 to warrant the conclusion that cellulose acetate film, properly processed, offers a satisfactory prospect of durability.

Moreover, the extraordinary evidence on durability that comes to light with the preservation of microcopied material from the siege of Paris must not be ignored. The reduction ratio, it has been established, was 32 to 39 diameters from newsprint. The film was probably a nitrate collodion base, less durable in all probability than the acetate safety film. This ∥ discussion of durability has been based on the assumption that photographers will use a film base more durable, and a reduction ratio less exacting, than that which has survived for 65 years from the siege of Paris.

Correspondence with film manufacturers and motion picture engineers is inconclusive in establishing rules for the storage of safety film. Some advice special measures to insure controlled conditions of humidity and temperature. All are agreed that there is no special fire risk in the case of safety film. The Recordak Corporation does not recommend any special control of atmospheric conditions in the storage of film. The Bureau of Standards reports a probability that film will dry out if stored in atmospheres of deficient relative humidity, but dried film can be reconditioned and restored. The practical plan that was agreed upon as satisfactory at the Richmond meeting of a group interested in microcopying (May, 1936) is to store film in a metal cabinet, keep a cheap hair hydrometer in the cabinet, and maintain humidity by putting a sponge in the cabinet if it gets too dry. The storage of film will not impose upon librarians any significant special expenses.

Assuming the permanence of film and the suitability of ordinary storage space for its preservation, the microcopying of documents for the sake of preserving them will be paid for by savings in storage space under certain conditions. Figures prepared by the Recordak Corporation indicate that it will pay to microcopy documents that are to be retained for fifty years if storage space costs are more than $.014 per cubic foot per year, and if storage space costs $.028 per cubic foot per year it will pay to microcopy documents that are to be retained for only twenty-five years. The costs of Recordak copying are less than the costs of the filing cabinets required to hold the original material. The microcopying fixes all the documents in an unbreakable series. This is an advantage as a protection against loss of individual documents through theft or carelessness; it is a disadvantage if a file is still active in the sense that new material ought to be interspersed between material already filed. ∣📄 p.153 Since library space costs $.40 to $.50 a cubic foot for construction (see Chapter II), and gives rise to further costs for lighting and heating, it would appear that the interest charges and maintenance costs of library space would pay for transferring material from paper to film. It does not follow from this fact that such a transfer of material is desirable, but only that it is economically feasible. The reduction ratio in microcopying that would bring about this cost relationship is of course a high one—sixteen to twenty-three diameters.⁵

The four methods in use for the storage of film are: (a) storage in cans; (b) storage on reels; (c) storage in a cartridge; (d) storage in partitioned boxes. The cans, such as those sold by the Jam Handy Company of Chicago at $3.00 per ∥ 100, are small tin pill boxes. They keep each short length of film in place, but they offer no protection to the ends, edges or surface of the film. The reels provided by the Recordak Corporation are suitable for long strips, just as the pill boxes are for the small strips; they protect sides and surface. The International Filmbook Corporation has a foolproof cartridge that can be inserted in the projection reading machine; it protects ends, sides and surfaces. The Leitz Company sells small metal boxes divided into twenty-five chambers each of which holds a five-foot roll of Leica film.

The preservation of the film involves not only the intrinsic stability of base and emulsion, but also requires special care in the processing, to the end ∣📄 p.155 that no traces of the developing or fixing solutions remain on the film; and it may call for the protection of the surface from abrasion in use. It is a question whether the film surface should not be given a special hardening treatment to protect it from abrasion. The International Filmbook Corporation claims that its projection apparatus is devised to prevent abrasion; the film is not in contact with glass while it is in focus in the projector.

If the cellulose acetate film base and emulsion should ever prove to be insufficiently stable, there is a chance that a microphotographic image on aluminum could be used. The firm of Siemens and Halske has developed a process which, as nearly as the writer can gather from available information, puts a photographic image on metal without a gelatine emulsion. The light, sensitive substance is an inorganic colloid, aluminum oxide. This is the “Seo Process.” Negotiations for its introduction in America are under way.

Microcopying of Newspapers⁶

The use of microcopying in the preservation of newspaper files is of special interest to librarians for several reasons.

a. The library’s investment in storage space is greater in the case of newspapers than in the case of books.

b. Newspapers printed since 1880 are perishable. Mechanical wood pulp paper began to be used about that time. A few newspapers today publish a special edition on permanent paper, and several suggested processes are in the field for preserving the wood pulp medium. The most tested process is that of the New York Public Library; a sheet of transparent Japanese tissue paper is pasted with rice paste over both sides of each sheet of newsprint. The cost of this process is $.04 a page.

c. There is no market supply of old newspaper files. Acquisition possibilities as well as prices of old files are a matter of accident. A library cannot plan an equipment of newspaper files with any assurance that it will be able to complete its plan.

From another point of view, newspapers would seem to be among the least ∥ suited for microcopying, because the price of original files of newspapers is low in terms of printed area of page. In copying the NRA records, for instance, microcopying distributed the copies at $.12 per 100 pages, when the only alternative was hectographing or mimeographing at $2.00 per 100 pages. The Recordak Corporation and the New York Times Company have arranged the microcopying of nearly 90,000 pages of the New York Times for the five years 1914–1918 inclusive. The positive copies of this file are offered to purchasers at $.66 per 100 pages. Since the bare subscription price of a year of the Times is only $18.00, or $.10 per 100 pages, it would seem at first glance that microcopying must be more expensive than original purchase.

In both cases the cost of binding and storage must be added to the price of the paper copies, and the cost of a projector or projectors financed into the cost of the film copies. But when considerations of storage, perishability, and availibility are taken into account, the case for microcopying of newspapers is much better than the above figures would suggest.

A comparison of the different prices at which newspaper files may be preserved brings into a comparative scale the three alternatives now confronting libraries, namely:

1. Preservation by purchase and storage of original files which may be on perishable paper in most cases, on permanent paper in a few cases.
2. Preservation by photo-offset in a reduced scale on permanent paper.
3. Preservation by microcopying on film.

The figures on microcopying in items 17, 18, and 19 below do not take into account the important factor of “first cost” in making the original negative film. The Recordak project is financed on a risk basis like ordinary book publishing. Early estimates of the cost of the first negative set this cost figure at $.10 a foot, or $284.00 for the original negative. The ∣📄 p.156

TABLE XLVI

COSTS OF ACQUIRING OR MAINTAINING ONE YEAR OF A NEWSPAPER FILE OF APPROXIMATELY 18,000 PAGES PER YEAR, BOUND IN 24 VOLUMES

(Based on dimensions and typography of the New York Times)

I. Preservation of current file. (Not applicable to back files)

II. Preservation in photo-offset copy, on good paper, reduced to one-half size (Applicable to back files)

III. Preservation in microcopy form.

cost of this negative is of course distributed among the subscribers to the sets of film, who buy their positive sets at $120.00. The price of the processing is set at a figure well below the maximum for this kind of work, and the project has been set on foot in such a way that the sponsors take the risk that they may not sell enough positive copies to distribute their “first cost,” and to expect, to gain a profit if the number of buyers more ∥ than distributes this first cost. It is probable that in this instance the actual first cost is much in excess of $.10 a foot because of the labor of collating the file. A special condition that confronts those who would microcopy newspaper pages is set by the state of present equipment. The newspaper Recordak copies loose pages only. Each page of the newspaper is passed through the machine, first on one side, then on the other. In this way ∣📄 p.157 newspaper pages can be copied in sequence in the “first position” on the film, in a reduction of 16 diameters. If bound newspaper files are to be microcopied without disturbing the binding, two obstacles would have to be met. First, the binding of the papers often leaves the inner column obscured because of the thickness of the volume and the narrowness of the binding margin. Second, a great amount of moving of the newspaper would be necessary if it were to be copied in the “first position” on the film. The camera would first take page one, then the whole bound newspaper volume would have to be moved the full width of a page to permit the copying of page two, then back again for page three, etc. If the bound newspaper volumes were to be copied in the “second position,” the reduction ratio would be greater because the whole height of the newspaper—22” in the case of the New York Times—would have to go on the same width of film that otherwise takes 18”, The New York Times microcopied in the first position is reduced only 16 diameters, in the second position it would have to be reduced 20 diameters.

There are some newspapers with larger sheets than the New York Times, and some with smaller print. This is especially true of old files. The experts of the Recordak Corporation testify that even these pages can be handled on 35mm. film, but some interested persons—notably Dr. Thomas Martin of the Library of Congress—feel that it is safer to count on using 70mm. film for particularly difficult copying operations. The newspaper file may be “bad copy” not only because of the large size of the page or the small size of the type, but also because of blurring or discoloration. Some of the Russian newspapers from the years of the Revolution, now stored in the Hoover War Library, are almost illegible in their present form because of this blurring of ink on a very poor quality of paper. However, with the example of M. Dagron and his microcopies in 32–39 diameters of reduction preserved from the year 1870, the technicians can probably be trusted to solve the micro- copying of bad copy without breaking too far from the cost levels indicated by the experiment with the New York Times.

Another important element in the question of microcopying has to do with ∥ the attitude of the readers and research workers toward this form. In 1934 the New York Public Library did a great service to scholarship by utilizing microcopies of current months of the New York Times in its newspaper reading room. The weight of opinion on the part of the readers was favorable to the use of films in preference to the newspapers themselves, largely because of the greater convenience in handling. The experiment was sufficient to justify the conclusion that readers will accept microcopied newspapers in lieu of the original files. The new reading equipment being devised will increase the advantage of the process, because it will make it possible for a reader to make his own photostatic copy of a part of a newspaper page whenever he wishes. This will reduce the cost of photostat copies of parts of newspapers to a materials plus processing cost, for it will make the reader his own photographer, eliminating expensive labor in the photostat room.

The present state of our knowledge regarding the advisability of depending on microcopying for the collection and preservation of newspapers can be summarized in a few statements:

1. No library should at present embark on a large scale plan for the purchase, binding and storage of newspaper files without taking into account the revolutionary implications of microcopying.
2. If, as now seems probable, an increasing number of newspapers offer their current issues to libraries in microcopy form at a reasonable price, the chances are all in favor of preferring the microcopies to a wood pulp paper edition on grounds of permanence.

Microcopying and Library Cataloguing

Another special use of the microcopying process can be found in the copying of library catalogues. It has been stated above that the leading library problem of the present lies in the field of interlibrary coordination. It was the service of the last generation of librarians to bring the technique of collecting, accessioning, and cataloguing books to such a high point that a great library becomes in effect a running index, constantly brought up to date, of the world’s knowledge. But as the demands ∣📄 p.158 of scholarship increase beyond the powers of any library to supply them, the logical unit for the next stage in organization becomes the interlibrary system.

One of the instruments of interlibrary cooperation is undoubtedly the union catalogue. The largest project of this kind is “Project B” of the Library of Congress, under the direction of Mr. Ernest Kletsch. The Library of Congress is filing in a great filing system the author cards of the holdings of all American libraries, especially the holdings of rare books. Already eight million cards are filed. But another use of the union list is in the field of library coöperation. If the libraries of a given region can centralize information about their holdings, they can not only save the scholar a great deal of time in discovering whether or not a given title is accessible, but they can make their accession money go further by avoiding unnecessary duplication in purchasing.

The need for “white collar projects” under the Works Projects Administration in 1935 led to the development of a technique for using microcopying in the preparation of a union list. Projects drawn up and submitted to Public Works authorities in the autumn of 1935 in Cleveland and in Philadelphia provided for the microcopying of all card catalogues of the libraries of each of these regions. The making of the cards for the union catalogue was then to be given to an army of typists, who would read the film copies of the catalogues. The reasons for copying indirectly, through film, rather than directly from the card catalogues of the participating libraries were, first, that the film copy could be made quickly, by one operator, using one drawer of cards at a time, and thus without disorganizing the functioning of a catalogue room; and second, that the typing and checking operations could be centralized in one room, with adequate supervision. Of course the cards would be more accurately and quickly copied if the film were enlarged by the available automatic enlarging machines to 3” x 5” card size, but the importance of using as large a proportion as possible of Works Progress Administration money for wages made it necessary to resort to typing.

However, a better technique for ∥ using microcopying in the compilation of union lists was worked out by Dr. T. R. Schellenberg and Mr. Paul Vanderbilt in connection with the Philadelphia union list project. They proposed to microcopy a catalogue, compare the microcopy film with a foundation catalogue of Library of Congress cards, designate the holdings indicated on the film copy by marking the Library of Congress card whenever the title was already present in the foundation catalogue, and punch the film whenever there was no card present in the foundation catalogue. The Recordak Corporation machinery would then pick out automatically and reproduce photographically on 3” x 5” cards the film items that were punched for reproduction.

Under some circumstances it may be advisable to reproduce a whole library catalogue on cards. Three photographic techniques are available for this purpose. A number of cards can be laid out on a flat surface and copied on a large sheet of photostat paper with the ordinary photostat machine. The sheet can then be cut up into 3” x 5” cards, and the cards filed. Or, the Dexigraph or photostat can be used to reproduce the cards one by one on 3” x 5” cards of photostat paper. This is the process used by the Library of Congress in its Project B; it was also used by Yale University Library in duplicating its great catalogue for internal administrative purposes. Then the Recordak Corporation can take a 16mm. microcopy of a card catalogue and enlarge it automatically on 3” x 5” cards. The photostat and Dexigraph methods yield a negative (white on black) card; the Recordak method a positive (black on white) card. The costs run from $20,000 to $30,000 per million cards.

There is a possible substitute for a union catalogue that may make its way in the future in library administration. A great card catalogue may be copied on the 16mm. film, and the catalogue itself held by another library for reference and checking purposes. A group of such catalogues might be called an “assembled catalogue” rather than a union catalogue. The cards would be consulted in the reading machine, like any other microcopied material. The cost of making a microcopy of a card catalogue, for this kind of use, would be only $500.00 per million cards, including labor costs. While it would be less useful than ∣📄 p.159 a union list in one respect—one would have to consult each library catalogue separately to determine whether or not a given book were available—it would in another way offer an advantage over the ordinary union list, for the whole subject entry classifications of a library catalogue would be reproduced. If, for instance, the film copy of the catalogues of the Library of Congress, the New York Public Library, and the Widener Library of Harvard University were on the floor of Columbia University Library, a scholar could check up quickly the holdings that each of these libraries had catalogued under a given subject head. This kind of use could not even be made of the great Project B union catalogue. The place of the “assembled catalogue” in library administration is untested. Another weakness in the use of the microcopy of a library catalogue as a basis for the organization of interlibrary coöperation is the impossibility of filing new entries between old ones. Perhaps the idea will prove most serviceable if it is combined with the union list idea. Libraries might procure microcopies of each other’s catalogues, and then arrange for the exchange of catalogue cards on certain classes of future entries, or supplement the “assembled catalogue” (a number of copies of separate library catalogues assembled in one room), with a true union list of current acquisitions, made by filing one card for each new title acquired by each of the coöperating institutions.

Photography as the Scholar’s Amanuensis

The analysis of costs and problems in the preceding pages has been devoted principally to the place of microphotography as a part of a library system. But the technique offers a variety of adaptations to the needs of individual scholars who may do their own work. Some scholars will find it useful to possess a camera that can be used for pictorial and document photography alike. For them the Leica or Contax, with its numerous accessory fixtures, is the best implement for microcopying. The selection of pages for copying and the copying itself are merged as one operation, so that labor costs in photographing do not appear; the photographing is part of the research. The scholars who have taken microcopies of documents ∥ have often found it necessary afterwards to enlarge them on photostat paper. If the material is such that enlargements will be necessary, a step can be saved by making the original photograph on a narrow strip of photostat paper in the Graflex camera, soon to be placed on the market. Or again, reflection copying without a camera may facilitate the gathering of records by a scholar.

Just as the scholars of the last generation found in general that it was desirable to be able to use the typewriter, so the scholars of the next generation will find it necessary to use photography. Just as each scholar works out his own note system, combining typescript and longhand notes, so the scholar of the future will make his combinations of typescript, longhand, photographs on paper, and photographs on film. Microcopying, as a technique in the hands of a man who does his own work of photographing and processing, is reduced to its bare materials costs. It offers the possibility that a scholar, by purchasing microcopies from libraries and by making his own microcopies of excerpts from books, may build up organized accumulations of data that will resemble a private library in extensiveness, and a note system in its internal organization. Pictorial and textual material can be fitted into the same file. That which scholars in the past have been able to do with the help of an amanuensis, the scholar of the future may be able to accomplish with photographic equipment.

It is possible that the development of this use of photography—whether on paper or film—will have an effect upon the intellectual standards of scholarship, for the technique will automatically tend to distinguish between that kind of intellectual labor that is merely the gathering together of documents and that which is the creative analysis of evidence,

The Merging of Collecting and Publishing

An even more sweeping change in the pattern of intellectual organization is to be anticipated from the full development of the possibilities of microphotography. The change that can be expected is one that, in a way, would restore a condition that passed out of existence with the invention of printing. For it was the technique and ∣📄 p.160 accountancy of printing that created the sharp distinction between the operations of collecting material and publishing. In the days of the manuscript, copies of a book were made to order for a collector; a possessor of a book was in a position to have it copied when a copy of it was desired. There was hardly any distinction between “published” and “unpublished” material. If a monastery maintained a scriptorium, and if its policy was to accede to requests for copies of its books, its whole library was, in a sense, “published” material.

Precisely this situation returns to the degree that microcopying becomes generally available. A library with a collection of theatre programs, or pamphlets of the revolutionary war, or historical manuscripts may adopt the policy of making microcopies of its materials upon request. If it adopts this policy, its whole collection becomes available to the world as if published. If the policy becomes widespread, so that one institution can expect to have its requests honored at other institutions, its collection can be perfected and rounded out according to a logical scheme. In that way, the making of a collection comes to resemble more nearly the preparation of a work of scholarship than a mere mercantile operation in the book market. Before the eyes of scholarship there looms a new unit of intellectual activity—the making of collections which will be “complete” within their predetermined limits to a degree that has been hitherto impossible. And this new unit, the collection, ceases to be functionally distinct from a publication.

The more clearly this merging of collecting with publishing is achieved, the greater will be the place that scholarship must give to the bibliographical and abstracting services that will guide the research worker through the embarrassing array of materials that will become accessible to him. The fact that a given item is printed as a book has hitherto served to give it recognition in bibliographical guides from which unprinted items are ex∥ cluded. If this mechanical basis of distinction loses its significance, the emphasis upon intellectual bases of distinction must be greater than ever.

Some librarians have raised the question whether it would not be wise to make up special bibliographies of microcopies that are available. The problem they raise is important, but it leads beyond the point they have in mind. If a library has installed apparatus for microcopying and is willing to microphotograph its holdings on request, does not its entire holding become, in effect, available as microcopied material? If manuscript material is microcopied in Europe and brought to America, it becomes a new item, a new resource of American scholarship. But the same is true of a body of American manuscripts that may be discovered in an attic and deposited in a library that offers microcopying service. It is premature at this stage to do more than suggest how wide the new fields are in the science of librarianship that are now open for exploration. There is however, one conclusion that can be drawn, even in the present state of the technique. This conclusion is that we may expect that the future will make the small libraries more important than they have hitherto been: in the intellectual organization of the country. They will be able, not only to draw more freely upon the large libraries in the service of their scholars, but also to shoulder their share of the burden of maintaining the great archive of our civilization. Each unit in the vast library system can perform a task that will be proportionate to its resources, and yet fitted into the general scheme in a unique and useful way. The library may come to be, not only a depository of printed material, not only a collector of existing records, but even a maker of new records. It may come to be a part of the function of the library of the future to reduce to writing information that would otherwise go unrecorded. There is no community so small that it does not offer to a sensitive mind aspects of human life that are worthy of record, and facts that should be entered in the dossiers of scholarship.