The oddest feature of an early photograph of the Metropolitan Museum's Armor Room may be the amount of floor assigned to the glass case at the front. It stretches across the foreground like a low table. Beyond it, individual pieces shine in tall dark recesses against the wall. A visitor standing where the photographer stood could not have read a label at the far side of the room. A reader standing where we stand can enlarge the image but cannot walk past the case. That small difference in what bodies are allowed to do opens up a history of how the museum was seen, and of what survived when a gallery was taken apart.
The photograph comes from a glass negative in the Detroit Publishing Company collection at the Library of Congress. The catalogue dates it to about 1907 and assigns it the digital identifier ending in 4a22226. Its title names the room in New York. Its record lives in Washington. The arms have their own histories at the Met, while the photographic negative has a separate life in another institution. An online image brings these two histories together with such ease that their separation is easily missed. There is a third history here as well: the successive redesigns that made it impossible for a later visitor to see exactly this room without a photograph.

The low case and the high wall
Begin in the foreground instead of looking for a celebrated object. The low display invites a standing visitor to look down into it. The wall cases ask the visitor to lift their eyes. Above the objects, architectural decoration continues across the room. Those three bands require different angles of attention. A curatorial inventory organized by object type would have little reason to record any of them. A photograph taken from the far corner could arrange the same bands differently. What the Detroit negative preserves particularly well is a relation between heights, together with the patch of open floor necessary to get from one kind of looking to the next.
The museum itself has published an unusually good account of changes in an early arms display. In his study of the Duc de Dino collection in the Metropolitan Museum Journal, Stuart W. Pyhrr discusses a 1907 photograph of the installation and a different arrangement made by Bashford Dean in 1909. Mirrored cases, closely assembled helmets and arms arranged above them appear in the account of the earlier gallery. The later photograph gives a mounted figure an emphatic position. Rather than identify every dark shape in the Detroit negative with one of the numbered pieces in the journal, put the two publications beside each other. The journal supplies information about a collection. The other photograph independently preserves the look of a room.
Even an obvious reflection should give a reader pause. A mirrored backing can let someone in the room see the far side of an exhibit. To a later reader working from a single photograph, the same convenience can make a case look more densely stocked than it was. Counting every shining silhouette in an attempt to recover an inventory is a dangerous way to proceed. Yet the silhouette need not be discarded as evidence. It shows something about the experience of abundance the visitor encountered. A photograph can be an excellent witness to the effect of an installation even where it is an unreliable source for its arithmetic.
Across the stairway
The second Detroit photograph changes the scale. The library calls it the main stairway and gives a range between 1902 and 1910. It looks toward a colonnaded opening and an upper level framed by an arch. Statues are set into the architectural composition. A title such as Main Stairway may lead a researcher who is looking for arms and armor to discard it before opening the file. There is no obvious case of Japanese blades in the middle of the frame. Yet the relative levels of the building are precisely the piece of information that an isolated photograph of a full display case does not carry.

The Met's history of Bashford Dean's Japanese collection explains why the staircase belongs in an account of sword galleries. The early display occupied thirty-eight cases on the north balcony at the head of the grand staircase. Dean recorded ideas for the arrangement in a notebook. Two pages show a contemplated rearrangement in 1907. A photograph from 1909 supplies the museum with a view of the corresponding case as installed. The Detroit stairway photograph, Dean's notebook and the museum's balcony photograph answer three different questions. One gives the building a legible scale, one gives a proposed arrangement, and one makes it possible to see a result. They should not be collapsed into three photographs of the same event.
Consider particularly the time implied by a sketch. If an object is named beside an outline on a notebook page, the note is strong evidence that its presence was contemplated when the note was made. There might have been changes between the plan and the day of photography. Even a later installed view cannot demonstrate that the case remained as shown on every day between two recorded dates. This is an elementary point about opening hours and the passage of time, but quite a consequential one for an object lent by a private collector. To establish that a loan became a museum holding calls for an acquisition record, not merely a sharper reproduction of the case.
The collector on both sides of the glass
Dean occupied several roles in this story. His interests in natural history took him to Japan. He assembled a private collection, lent objects to the Met and prepared the 1903 loan catalogue. He also designed the display. The museum bought that collection at cost in 1904; a later group came as his gift in 1914. A label that says lent by Dean and a collection entry that says gift of Dean therefore need not contradict one another. They might describe two moments in the life of one exhibit. Before joining the two, however, the researcher still needs an identification more exact than a silhouette marked simply armor in an old photograph.
The museum's account illustrates that problem with actual documents. Items identified as Dean's loans in the 1907 notebook would later form part of his gift. The pieces had already been visible to visitors before the ownership change. An ordinary visitor might have registered no difference in the space on the day that the legal or administrative status altered. In a museum ledger there is a significant difference between these two dates. The photograph is evidence that the public and administrative histories intersect. If a historian dates a gallery simply by looking up the gift year of every visible object, there is a very real risk of making the whole photograph too late.
There is another revealing date that belongs to a journey rather than a gallery. Dean arranged an exchange of material with the Imperial Museum in Tokyo. The Met dates the arrangements to 1905 and the official exchange to 1906. Its history identifies an early guard from a burial mound at Shioda as part of the exchange. The unusual feature of the episode is that it connects an object a New York visitor might associate with Japan to Egyptian antiquities traveling in the other direction. No display photograph can show both ends of that transaction. The accession credit and the museum's historical account recover a part of the institution's international life that the orderly glass case leaves out.
Why the jade photograph matters
A third Detroit negative shows the Jade Room, which the library dates to about 1903. Its decorated ceiling and reflective surfaces could easily monopolize a viewer's attention. Down at floor level, separate cases describe a route through the space. There are low horizontal surfaces near the viewer and several opportunities to turn or stop. A photograph of a single jade object against a neutral background would allow a closer examination of colour and translucence. This monochrome room view preserves the relative importance of the ceiling, the stands and the central axis. It makes those installation choices available for comparison with the Armor Room, without suggesting that the two collections once occupied one space.

Read the library's dates before treating the photographs as a walk taken one afternoon. The Jade Room record is dated approximately; the Armor Room record gives an approximate copyright year; the stairway has an interval. The nearby digital numbers can make them feel like successive frames exposed by a single photographer during a single visit. The numbers are powerful retrieval devices, and the subjects plainly belong to the same museum. Neither fact dates an exposure. An exhibition designer could arrange reproductions of these images into an intelligible walking tour provided that the display made the different dates legible. A historian who wanted to establish an actual visitor's route would need contemporary plans and some evidence of public access at the time.
A useful detail is what the glass does to the boundary between one exhibit and the next. In the Jade Room some surfaces frame rows of objects, others catch a glare that makes the contents hard to distinguish. In the Armor Room the dark backs of the taller cases give the metallic shapes greater separation. Seen side by side, the two images show how a display surface can either isolate or amalgamate things. This tells us nothing certain about the daily conditions of illumination. Exposure, reproduction, glazing and the direction from which the photograph was taken all intervene. It tells us quite a lot about the practical difficulty of recognizing objects from early gallery photographs alone.
A hall photographed with people in it
The fourth plate is Statuary Hall. The library gives a date of about 1907. Sculptures line the long axis, figures occupy the central space, and the architecture carries the eye into the depth of the image. The human figures give something like a scale to the room. They also defeat an easy claim that early museum photography always excluded visitors. A person can be present in a gallery photograph without being its intended subject. A visitor in the distance can establish that the room was occupied and still tell us very little about the crowd on a typical weekday. A camera present for a few moments is a notoriously poor witness to averages.

Look again at the Armor Room after studying this hall. The open floor now seems more significant. Its apparent emptiness could result from the photographer's choice of position, the timing of the exposure or the subject commissioned for the picture. The photograph does not record what a visitor thought of the installation. It cannot tell us whether the low case attracted more attention than the wall displays. To answer those questions one might search diaries, guidebooks or visitor accounts, asking whether each writer was describing a first encounter or a familiar room. The photograph does contribute something specific to such a search: it supplies actual distances and sightlines against which a recollection can be checked, even when exact measurements are unavailable.
The difference between an installed sculpture and a represented person is especially important in this group of pictures. A small light shape in the distance of a black-and-white image may be legible as a figure without immediately disclosing which kind of figure it is. The temptation is to supply what the argument requires, a visitor if counting an audience or an object if estimating density. Enlarging that area sometimes resolves the matter. When it does not, where the shape stands in relation to bases, neighbouring statues and the boundaries of the floor may support a more modest description. A photograph can be read carefully and still contain an unidentified figure.
When an exhibition learned to move
The Met opened a new series of arms and armor galleries in 1915. Two galleries were outfitted for the Japanese material. The museum's published photograph of the new installation provides a valuable check on any impression that Dean's balcony remained the permanent setting for these objects. It also introduces a problem familiar to anyone who has tried to match an old postcard to a room today. A museum name can endure while the rooms occupied by a department change. A description indexed by museum and city alone is rarely sufficient to reconstruct the history of a display. Gallery names, levels in the building and changes of date are part of the identification.
Two films give the reader further ways of testing the still photographs. The Met's 1924 film about its armor galleries moves the viewer through a presentation shaped for a camera. Scene follows scene in a chosen order. The camera can emphasize a detail that a person in the gallery might have passed by. The 1928 film about the working side of the museum opens up an entirely different subject, the employees and processes that sustain an apparently effortless public display. Both survive in the museum's film archive. Playing the two together makes visible the distinction between offering a view of an exhibited collection and making the institution's work into the subject of a film.
A moving picture does not automatically supply the missing truth of a still one. The 1924 production had its own audience and an arrangement of scenes quite different from a continuous record of an ordinary day. As a historical source it can document what the museum wanted viewers of that production to attend to. It can reveal relations between exhibits and spaces where the sequence actually shows them. The 1928 production, in turn, can put conservation and cleaning into a history too often told entirely through spectacular acquisitions. Their dates are part of their attraction. Both films postdate the balcony pictures and neither can serve as moving footage of the unrecorded moment when Dean drew his notebook sketch.
Bringing a lost room into focus
For a small research exhibition the four Library of Congress pictures could be mounted beside reproductions of the design notebook and the museum's own dated gallery views. Under each image one could give the institution holding the negative, the building depicted, the type of date supplied by the catalogue and one simple claim the photograph supports. Doing that work in advance prevents surprisingly large errors in wall text. The room with jade belongs in the comparison because it shows a contemporary alternative for displaying small objects. The stairway belongs because the history of the Japanese display explicitly mentions a balcony at its head. Neither picture requires an invented identification of a sword to earn its place.
The most affecting part of the Armor Room negative is still the foreground. The display table occupies real space, its near edge cutting across the visitor's approach. At a museum today a visitor can ordinarily change position to clear a reflection, step closer to a caption, or look back from the doorway. All those small decisions have been taken from us in this one view. In compensation, we can put it beside an earlier plan, a later photograph, a dated acquisition entry and two films in a way few original visitors could have done. The case has survived as a historical relationship among those records. Looking after that relationship is a worthwhile form of museum work in its own right.
Questions about the archive
Who holds the four historic gallery photographs reproduced here?
The Library of Congress holds the Detroit Publishing Company negatives. The building depicted is the Metropolitan Museum of Art in New York. The digital identifiers appear in the captions so the four photographic records can be retrieved separately.
Were these four photographs taken on the same day?
The catalogue gives different date evidence. The Jade Room is dated about 1903, the main stairway between 1902 and 1910, and the Armor Room and Statuary Hall to about 1907. Nearby digital numbers do not establish a single photographic session.
Where did Dean’s early Japanese installation stand?
The Met places its thirty-eight cases on the north balcony at the head of the grand staircase. The published notebook and balcony photograph document the Japanese display more directly than the general Detroit stairway negative does.
Why can a loan label precede the museum accession date?
An object can be on public view while still privately owned. The Met identifies loans noted in Dean’s 1907 notebook among objects later donated in 1914. A dated view and a museum gift record may therefore describe different events.
When did the Japanese display move into the new galleries?
The Met dates the move to the new series of galleries to 1915. Two galleries were fitted out for the Japanese collection. Its photograph of the later installation should be distinguished from the earlier balcony view.
What do the two early museum films add?
The 1924 production presents the armor galleries through a sequence designed for a viewer. The 1928 film records the working side of the museum, including aspects of caring for the displays. Neither is contemporary footage of Dean drawing the 1907 plan.
Museum films
The Met — A Visit to the Armor Galleries, 1924 | From the Vaults
The Met — Behind the Scenes: The Working Side of the Museum, 1928 | From the Vaults
Collection records and further reading
- Bashford Dean and Japanese Arms and Armor
- Stuart W. Pyhrr, Armor for America: The Duc de Dino Collection
- Smithsonian Libraries: Dean’s 1916 Notes on Arms and Armor
- The Met: exhibition on Dean and the creation of the department
- The Met Film Archive: The Working Side of the Museum
- Reading a Japanese Sword Exhibition Catalogue: Dates, Eras and Corrections
