Worth a Billion Words?
Library of Congress Pictures Online

Barbara Orbach Natanson

Journal of American History, 94 (June 2007) 99–111

View Larger Image

Figure 1. This photograph of Herman A. “Germany” Schaefer, taken in 1911, was the millionth image scanned from the Library of Congress, Prints and Photographs Division collections and made available online.

Courtesy Library of Congress, Prints and Photographs Division, George Grantham Bain Collection, LC-B2-2189-6, http://hdl.loc.gov/loc.pnp/ggbain.09131.

Early in the 1911 baseball season, the Washington Senators played a series of games against the New York Highlanders (then increasingly being dubbed the “Yankees”). After two games in Washington, D.C., the Senators journeyed to New York City to play the Yankees’ season opener on their home field, Hilltop Park. Patriotic decorations decked the stadium, and a band was scheduled to play. The New York–based Bain News Service was on hand to capture the scene, making photographs that could be offered to subscribing publications all over the country. The Washington Senators’s infielder Herman A. “Germany” Schaefer obliged by posing simultaneously in front of and behind the camera, playfully assuming the role of photographer for the Bain News Service’s coverage of the event. (See figure 1) ¹

Schaefer was known for antics that entertained fans and buoyed up his teammates, so it is not surprising that the news photographer focused on him to depict the high spirits that accompanied the start of the season. Although Schaefer could not have known that he was about to launch the best hitting year of his career, both players and fans were likely already in a mood to celebrate that April. The unseasonably bad weather that had marred preseason play was starting to break. Moreover, the Yankees and the New York Giants had just agreed to share Hilltop Park for the season in the wake of a disastrous fire that swept through the Polo Grounds, the home field of the Giants. An even more tragic fire had swept through the Triangle Shirtwaist Factory downtown less than a month before. The Bain News Service carried grim pictures of the aftermath, and the New York Times was still regularly covering the Triangle factory story. The pleasures of the game and attendant hoopla could certainly have offered welcome distraction from the horrifying realities of the fire and the exploitive labor practices it highlighted. ²

Just as the Germany Schaefer photograph evocatively celebrated the start of the 1911 baseball season, the image marked another celebratory occasion nearly a century later when it became the one-millionth image that the Library of Congress Prints and Photographs Division scanned, described, and made accessible through the Prints and Photographs Online Catalog. ³ The photo is one of thirty-nine thousand glass plate negatives from the George Grantham Bain Collection that can now be searched and viewed over the Internet. Prints and Photographs Division visual collections that became available online in digital form before the Bain Collection include the following:

Ansel Adams’s photographs of Japanese American internment at Manzanar (244 items)
American cartoon prints dating, 1766–1876 (500 items)
Civil War photographs, including many generated under the auspices of Mathew Brady’s studio (7,000 items) Farm Security Administration and Office of War Information black-and-white and color photographs from the depression and early World War II eras (172,000 items)
Historic American Buildings Survey, Historic American Engineering Record, and Historic American Landscapes Survey materials produced by the National Park Service, which, through photographs, measured drawings, and textual records document historic sites throughout the United States (500,000 items covering more than 37,000 structures)
photographs of the Middle East taken by the American Colony Photo Department and the Matson Photo Service between 1898 and 1946 (14,000 items)
National Child Labor Committee photographs taken by Lewis Hine between 1908 and 1924 (5,100 items)
posters, including works issued by the Work Projects Administration; World War I posters from the United States, Germany, and France; performing arts posters from the turn of the twentieth century; and propaganda posters from more than fifty countries, 1960s–1980s (9,000 items)
photographs of Russia by an early innovator in color photography technology, Sergei Mikhailovich Prokudin-Gorskii, 1905–1915 (4,000 items)
the Wright brothers negatives, which document their experiments in building and flying aircraft between 1897 and 1928 (303 items)

As one might suspect with an operation as complex as scanning and describing hundreds of thousands of images in different formats, shapes, and sizes—an enterprise that took the Prints and Photographs Division twelve years—the Germany Schaefer photo was not simply the millionth to pass through an assembly line. As we neared that landmark number, the division selected the Schaefer image, both for its visual appeal and for the many ways that it sums up the strengths of Prints and Photographs Division collections and the goals of the digitizing and cataloging enterprise.

The Prints and Photographs Division’s holdings number nearly 14 million items, comprising photographs, posters, cartoons, architectural documentation, fine and popular prints, and applied graphic art such as patent medicine labels and sheet music covers. One of the largest resources in the world for exploring history visually, the holdings span the fifteenth century to the present. Photojournalism and its graphic counterparts are particular strengths of the collections, ranging from seventeenth-century satirical prints in the British Cartoon Collection and eyewitness sketches of American Civil War scenes in the documentary drawings collection to the photo archives of Look magazine and the New York World-Telegram and Sun newspaper. The George Grantham Bain Collection is a key component of the photojournalism collections, as Bain launched one of the earliest news picture agencies in the United States.

Library of Congress Mission

Not only is the Germany Schaefer image an apt indicator of one strength of Library of Congress visual holdings, but there is also telling symbolism in the nature of Bain’s enterprise and its meaning for the visual record. Bain realized the profound communication possibilities pictures offered at the very time when the introduction of halftone printing technology was making it easier to mass reproduce and distribute images. While working as a reporter in the 1890s, Bain experimented with supplying photographic illustrations for free with his articles, then (despite the resistance of editors) began requesting payment for the photographs. Drawing on his experience as a former manager for the United Press news syndication service, he adapted the practice of collecting and re-distributing news stories to photographs. He arranged for photographers to take photos and also exchanged photographs collected or made by his subscribers. Bain opened offices at 15 Park Row in New York City’s newspaper district so that he could quickly deliver photos to editors’ desks. By 1905 he had reputedly amassed one million photographs, and, as his business expanded, he moved to larger quarters in the Parker Building. The building was swept by fire in 1908, destroying all the photos he had accumulated, but he started over, and, by pooling photographs produced by a variety of sources, he created a centralized repository of images that document events in New York City and beyond. ⁴

Bain’s vision of preserving and distributing a visual record of his time parallels the mission of the Library of Congress: to preserve and make available knowledge and creativity as it has been expressed in multiple media. ⁵ Just as technological changes in Bain’s generation enabled him to make images more widely available, the digital revolution has enabled the Library of Congress to fulfill its mission more effectively. The ability to transmit information readily beyond the walls of the institution enables the library to make its resources available to Congress, the American people, and other Internet-connected researchers around the world. Future researchers will also benefit from the efforts of the Library of Congress. Digitization has helped ensure that resources that were never designed for constant handling will continue to exist into the future. Large numbers of people are now able to view historical images without having to handle the fragile originals—an indirect form of preservation.

The digitization of the Schaefer negative is a perfect example of the benefits of the library’s work. When the Library of Congress acquired the Bain collection in 1948, there were corresponding photographic prints for only a portion of the glass negatives. The library’s budget did not allow it to print the remaining images, so for more than fifty years, much of the collection could not be readily viewed. The glass negatives are too fragile to be used in the reading room, and only a trained eye is able to read a negative image well. The Prints and Photographs Division began experimenting in the 1980s with videodisc technology as an electronic surrogate to present negatives in reversed polarity, so that they appear as they would if printed. That technology proved to be an effective way to extend access to such collections, but increased access posed the threat of increased handling: users began selecting more and more negatives they wished to have printed. High-resolution scanning, however, makes it feasible to use scans to fill reproduction orders, which enables the library to retire the fragile negatives to cool storage, where they have a better chance of remaining intact for future generations to study using technologies that will likely further expand avenues for research.

Visually Rich—(Con)textually Poor: Opportunities for Collaboration

The image of Germany Schaefer is representative of the collections in the Prints and Photographs Division not only because of the circumstances of its creation and preservation but also in the amount of identification that accompanies it. What we know about each Bain photo and its subject is limited to a caption scratched on the negative or recorded on the sleeve that protected it while it was in Bain’s possession, supplemented by general information about the Bain collection and Bain’s working methods.

The caption information is sparing, indeed: “Schaefer (Washington)” is the clearest identifying text on the negative. Some illegible letters and two series of numbers also appear: “4/22/11” and “2/20/15.” ⁶ Both seemingly precise dates turn out to be misleading. Schaefer had left the Senators by 1915. The Senators and Yankees did not play on April 22, but they did play on April 21; the April 22 date may indicate that the photo was printed or added to Bain’s files the next day.

But much remains unknown about the image. In turning the tables—by showing the photo’s subject focusing a camera outward on a scene the viewer can only imagine—the photograph makes the viewer more acutely aware of the role of the photographer who is capturing that slice of time. Who photographed Schaefer in that pose? We know that Bain collected images from those who subscribed to his photo service, made images himself, and employed photographers to make newsworthy pictures for him. Bain initially tried to employ commercial studio photographers, but found them unable to work quickly and at the same time make good pictures. Instead, he trained his office boys to capture “spot” news stories and the general ambience of the city. ⁷ There is no record of who the photographer was at the Highlander-Senators game that April day, nor how arrangements were made with Schaefer. Did they plan the image, or did an opportunity simply present itself during the pre-game entertainment or some lull in the action?

The unknowns about the Schaefer image are symptomatic of the dearth of documentation available for many historical images. With nearly 14 million images to preserve and make accessible, staff are rarely able to undertake in-depth research on individual images, even when documentation can be found. Making images available digitally creates opportunities for others to assist the Library of Congress in its mission. In order to make the scans of the Bain Collection negatives available as expeditiously as possible, the library had the scanning contractor record the original captions during the scanning process. As soon as library staff finished checking a set of negatives and data for quality, we mounted them for public viewing and keyword searching, without doing further research to add information, to correct errors that the caption writer may have introduced, or to apply standardized subject headings.

In the year since the Bain Collection negatives first started becoming available, we have had numerous correspondents offer just those types of information. Historians of woman suffrage have recognized well-known participants in the suffrage campaign. Baseball historians have offered to supply identification and standardized index terms for the baseball images. As a result of the distribution of the images on the Internet, the process of documenting and improving access to the collection has become a collaborative one. ⁸

Research Value Added: Integrating Sources Virtually

The Internet creates the potential to add value to primary source materials by gathering and linking the knowledge of many people with those sources. Online presentation can also increase the research potential of historical image collections simply by bringing together pieces of historical evidence that have long been separated. The library’s experience with the National Child Labor Committee (nclc) photographs held by the Prints and Photographs Division is an example of such an opportunity.

Those photographs came to the Library of Congress with the records of the nclc in the 1950s. The library’s Manuscript Division cares for the textual records; the Prints and Photographs Division provides the specialized housing needed for the collection’s 355 glass negatives and 5,100 photographic prints. In the pre-computer era, the library used microfilm to preserve and to provide access to the photographic prints, which are grouped in albums by subject matter. As anyone who has used microfilm knows, it is neither the most comfortable nor the most conducive means for detailed analysis of photographic prints. The microfilmed photographs simply were not easy to “read.” The problem was compounded by the physical separation of the photos from their original caption cards. The captions, probably drafted by the nclc’s principal photographer, Lewis Hine, are tantalizing in their unusual amount of detail: they often included the subjects’ names, ages, wages, and working hours and comments from those depicted, as well as from the photographer. To bring together a caption and an image, however, researchers had to balance a card drawer somewhat precariously at the microfilm machines, painstakingly look up captions by photograph number, and copy the image and caption separately for later analysis.

All of that changed in 2003, when the Prints and Photographs Division began a twopronged project to produce high-resolution scans of the photographic prints and negatives and to key the captions into a database. Now, through the online catalog, researchers can retrieve images by keywords in the captions and view an image and its caption together. As we did with the Bain Collection, the Prints and Photographs Division released the documentation in phases, as it was ready, first making available the original captions and images. Subsequently, we have used the collection as a training ground for library school interns and others to learn how to analyze and index images, adding standardized subject headings that efficiently bring together images that have very disparate keywords in their captions but nevertheless share subject matter.

For the contextual information that accompanies the online collection, the Prints and Photographs Division partnered with the Manuscript Division to illustrate the research potential in coordinating the images with the textual records of the nclc. We scanned one of Lewis Hine’s reports, which is physically held in the Manuscript Division in the Records of the National Child Labor Committee, and created links to the photographs Hine referred to in the report. ⁹

View Larger Images

Figure 2. Lewis Hine provided the following caption for this photograph that he took in August 1911, in Eastport, Maine: “In center of the picture is Phoebe Thomas, 8 year old Syrian girl, running home from the factory all alone, her hand and arm bathed with blood, crying at the top of her voice. She had cut the end of her thumb nearly off, cutting sardines in the factory, and was sent home alone, her mother being busy.” Prints of this photograph did little to convey the power of her story, but a high-resolution scan of the image and zooming tools brought Hine’s visual and textual accounts of her experience into better focus.

Courtesy Library of Congress, Prints and Photographs Division, National Child Labor Committee Collection, lot 7476, no. 2444, http://hdl.loc.gov/loc.pnp/nclc.00966.

Research Value Added: Detail Revealed

Examining high-resolution scans of the nclc photographs demonstrates yet another way in which digital technology adds value and expands research potential. While keying captions one day, I took note of the ominous tone of the caption for photo 2443:

Eight year old Syrian girl, Pheobe [i.e. Phoebe] Thomas, going to work at 6 a.m., August 14, with great butcher knife, to cut sardines in Seacoast Canning Co. Factory #4, Eastport Me. Said she was a cutter, and I saw her working later. (See photos of her accident, #2444, #2445, #2449.)

The story unfolded in the caption for photo 2444:

In center of the picture is Phoebe Thomas, 8 year old Syrian girl, running home from the factory all alone, her hand and arm bathed with blood, crying at the top of her voice. She had cut the end of her thumb nearly off, cutting sardines in the factory, and was sent home alone, her mother being busy. The loss of blood was considerable, and might have been serious. (See succeeding photos.)

But that second photo puzzled me. Phoebe did not appear to be the central focus of the image, and I saw no trace of the distress discussed in the caption. I initially attributed this to the limits of the photographic technology: In his haste to take the picture, Hine apparently did not have time to get near enough or to compose the image well enough to convey what was happening. Although images are regularly credited with being worth a thousand words, sometimes words communicate more than can easily be seen. When I looked at the high-resolution scan of the image on my computer screen, however, I realized that, whatever the deficiencies in Hine’s composition, he had captured much more of the scene than was evident from the approximately four-by-six-inch photo. The expression on Phoebe’s face did, in fact, mirror Hine’s description of her agony. Together, image and text brought to life the horror of the scene and the plight of child workers and their families that Hine strove to convey (See figure 2).¹⁰

In some cases, the twenty-first-century scanning of Hine’s images accomplishes his early-twentieth-century propaganda objective more successfully than earlier reproduction technologies permitted. In other cases, the scanned images bring to light details not evident to the naked eye and, perhaps, not consciously recorded by the photographer. This technology enables researchers to raise new questions about what they see and about the role of the photographer in composing the scenes he or she captured.

New Research Potential: Conventions, Comparisons, and Context

Historians continue to note the pedagogical potential in using visual materials of all types to teach about the elements of historical analysis. ¹¹ The intricate process of analyzing visual sources can hardly be broken down to a formula, but some of the general elements include: a close reading of the content; a comparison to like and unlike items; and an awareness of visual conventions and context (including the creator’s purpose, the intended audience, and the technology used to produce the item).

The digitization of historical images promotes both a broader and deeper analysis of the images. In Photographs: Archival Care and Management, Helena Zinkham suggests a systematic approach to the close “reading” of images that requires spending minutes examining each element of the picture—a procedure enhanced by the greater detail made available through digitization and zooming capabilities. She also discusses how the mak ers of the images may have deployed the “vocabulary” of visual expression to accomplish their purpose. For instance, Lewis Hine often posed his subjects beside adults or recognizable stationary objects to emphasize the small size of the child workers, and he varied the depth of field and perspective, sometimes providing close-up portraits and sometimes widening the view to highlight the risky surroundings. Making an entire corpus of Hine’s child labor images readily available enables comparisons among the images, from which an awareness of such conventions emerges. The availability of other images, made by people with different objectives, invites investigation of how intended purposes may have affected composition and presentation. Researchers might, for example, compare Hine’s images of newsboys with those appearing in Bain News Service photos or in Detroit Publishing Company photos made for use as postcards and souvenir images. ¹²

The value for historical research of digitizing the entire surviving output of a particular photographer or photographic enterprise increases with the size of the collection. For about fifty years, researchers studying the photographs made by the photo unit operating out of the Farm Security Administration (fsa) and, later, the Office of War Information (owi), had to formulate their conclusions based on intensive looking through approximately 107,000 photographic prints. Researchers had two options for examining the material: (1) in an arrangement by geographic region and, within each region, by a subject classification scheme devised while the photographs were at the owi; or (2) in an arrangement based on what Library of Congress staff deduced the photographer’s assignment must have been. Viewing the photos in the first arrangement entailed either the physically taxing task of looking through file cabinets or examining them on microfiche produced by a commercial publisher. The second arrangement (based on the photographer’s assignment) exists only on microfilm not widely held beyond the Library of Congress.

The sheer size of the collection and the time required to view it did not encourage systematic looking. Furthermore, the photographic prints represent only part of the story. The collection came to the library with the unit’s photographic negatives, not all of which had been printed when the collection was a working file. More than sixty thousand images contained in the collection had been virtually unseen since the 1940s.

Research opportunities expanded in the mid-1990s when, for preservation reasons, the Library of Congress arranged to duplicate the negatives onto film stock and then digitized the resulting intermediary reproductions. Although the resulting digital images are low resolution by today’s standards, systematic examination of the images is now possible. Researchers can view unprinted and printed images alongside each other, can rapidly view wide swaths of the collection, and can retrieve images by photographer name, place name, and keyword. Moreover, the same body of images is available to everyone who has access to the Internet, enabling all researchers to share and evaluate the same evidence.

Context: Lost and Found

View Larger Images

Figure 3. “Bound for Glory: America in Color, 1939–1943,” an exhibition of early color photographs made by Farm Security Administration (fsa) photographers on the Library of Congress Web site in 2005, excited speculation from bloggers about the nature and authenticity of the images. Among them was “Chopping cotton on rented land near White Plains, Greene County, Ga.” Photograph by Jack Delano, June 1941.

Courtesy Library of Congress, Prints and Photographs Division, Farm Security Administration Collection, LC-USF35-599, http://hdl.loc.gov/loc.pnp/fsac.1a33894.

The ease with which the Internet permits people to share, copy, and reuse images poses a hazard as well as an opportunity for interpreting images and their context. A recent episode regarding a subset of the fsa/owi images drove that point home.

In 1939, members of the fsa photo unit began shooting with Kodachrome color transparency film, which the Eastman Kodak Company had introduced to the public in 1936. About sixteen hundred of the resulting color photographs came to the Library of Congress with the rest of the collection. Because of the rarity of color photographic images from that period and growing research interest in them, that set of images was among the Prints and Photographs Division’s first candidates for reproduction onto videodiscs, when libraries began to use that technology in the early 1980s. The images were made available on the Internet in 1998, along with contextual information that explained how the images came to be created and how they had been digitized. The digitization information was updated when the images were re-digitized at a higher resolution in 2004, in conjunction with the publication of the book Bound for Glory and a subsequent exhibition of the same name (See figure 3).¹³

The online version of the exhibition did a wonderful job of raising awareness of the images. Because the links to the photos’ contextual information were less evident, however, viewers maintained a healthy speculation not only about the content of the images but also about the technology that was able to depict in color a period usually seen only in black-and-white. A lively discussion ensued on a blog devoted to photography in the media. Many of the participants in the discussion showed a talent for close reading of individual images and raised questions that history teachers would welcome. But it was also clear that displaying the images separate from their contextual information invited misinterpretation and confusion about what was being presented and how. Some thought the images had been faked; others thought the Library of Congress had enhanced the color. ¹⁴

That experience highlighted the need, when presenting digital images, to link to data such as collection names, descriptions of the visual items, and, whenever possible, information about the circumstances in which the items were made and the technology used to reproduce them. This is an obligation the Library of Congress takes seriously. At a minimum, in the catalog records (or “metadata”) that accompany single images or groups of related images, we supply original captions associated with the images or brief descriptions, if they lack captions. When mounting entire collections online, we attempt to supply information about the scope of the collection (including any portions not digitized), how the images were produced and acquired, and how the library digitized them. We have also worked to develop a sustainable means by which those who use images can cite the image and link to it in the library’s database, where it can be viewed with the cataloging data and contextual information we supply. Researchers can cite a persistent uniform resource locator (url) that enables others to follow the citation to the resource in question. ¹⁵

Technologies for maintaining the link between digital objects and their metadata are still under development, as are technologies for presenting digital images in ways that convey qualities of the original. For instance, digital images are generally presented in standard sizes that are convenient for viewing on monitors but that are less effective for conveying the size of the original. In the online display of the library’s Yanker Poster Collection, an assemblage of international political posters and propaganda gathered by Gary Yanker between the 1960s and the 1980s, a scale bar helps distinguish small handbills that probably were designed for reading at close range from much larger posters that were designed to attract attention from a distance.¹⁶ Characteristics such as weight, texture, odor, and fragility are even less translatable to the digital medium. Nevertheless, digital images alert researchers to image collections ripe for research and can lead potential users to institutions where related nondigitized material may be available. Digital images also allow individuals to be selective in deciding which originals to view, thereby reducing wear and tear on rare items.

A wide distribution of images not only permits collaborative information sharing by repositories, researchers, and image creators, but, for images made within living memory, it also allows us to gain insights from image subjects more easily. Individuals depicted in fsa/owi photos, for example, have viewed those pieces of their personal histories, either onsite or remotely, and offered context and commentary on how the experience of being photographed affected them or how they affected the making of the photographs. ¹⁷ Such interactions underline how the photographer, subject, and viewer collaborate in interpreting images and investing them with meaning. What other insights might be gained through widespread access to the images and follow-up research in the communities where the pictures were made?

Sustaining and Transforming

In 2005, the Prints and Photographs Online Catalog processed 8 million searches, more than 95 percent of them launched remotely by users accessing the images from their homes, offices, and libraries. With each new user comes the promise of new questions and new intellectual connections. Making available large bodies of digital material that are in a variety of formats will help researchers deepen their understanding of images. They can view not only the images but also related information, whether in texts, moving images, maps, or audio materials, for the library’s digitization program has included all those media. New questions and research methodologies may well emerge as a result. The Library of Congress is experimenting with a “federated search” system that should make it easier to locate visual information contained in images held by the Prints and Photographs Division alongside information from the rich array of the library’s other media. For instance, digitized materials relating to Germany Schaefer that are available on the Library of Congress Web site include the following:

additional images in the Bain Collection, including one showing Schaefer playing for Cleveland, the team with which he ended his career in 1918
a photograph from the Chicago Historical Society’s Chicago Daily News negatives collection that portrays Schaefer in action at the start of his career, playing for the National League’s Chicago Orphans
baseball cards that used his image to market tobacco products
the Spalding Base Ball Guides (1889–1939), through which one can track portions of Schaefer’s career and the evolution of the game
an article in a 1918 issue of the U.S. Army newspaper, Stars and Stripes, that records that Schaefer (the “well-known major league comedian”) dropped his “Germany” moniker, to become “Liberty” Schaefer ¹⁸

The Library of Congress plans to continue its digital conversion of historical resources, while expanding its programs to acquire information that is “born-digital” and collaborating with other institutions whenever possible to promote collection building and standards for the presentation and preservation of digital data. ¹⁹ Digital technology has become an important, even transformative, tool the library can use to accomplish its mission to sustain a universal collection of knowledge and creativity. Ideally, digital technology will also play a transformative role in opening new avenues for research, resulting in new bodies of knowledge and creativity.