On a small Japanese sword fitting in Cleveland, rice cultivation takes up less space than a detail in a painted scroll. A dark ground sets off a group of golden forms. A visitor standing before the case might first recognize a scene, a colour or an ingenious adaptation of a design to an oblong surface. The museum record says something quite different in its field for technique. It gives a Japanese alloy name, shibuichi. The word travels easily from a collection record into a label, an exhibition checklist or a book. Much less often does the evidence on which any particular identification rests travel with it.
A published conservation study offers an opportunity to see how that evidence can be made explicit. In London, researchers took instrumental readings on three Japanese guards in the British Museum. One of the dark features they examined was an ant. Another was a leaf. A statement about the ant did not necessarily apply to the metal beneath it. This deceptively simple point explains why a good scientific report on a decorated object needs both a diagram and a table. Put the instrument reading back on the animal or leaf pictured in the diagram, and the results tell a story that neither the photograph nor a list of elements can tell alone.
Four objects and four museum labels

The Cleveland Museum of Art identifies the rice scene as a kashira, a hilt-end fitting, accession 1920.1956. The designation tells us where the little work belongs in the vocabulary of decorative metalwork. The title tells us what the museum sees in the subject. The technique field gives the material name. All three statements are valuable, but they do different jobs. Someone searching for images of agriculture might find the piece without knowing anything about metalwork. A historian of copper alloys might find the same piece through the material field and never notice that an agricultural scene is at stake. We need both readings to understand the demands the artist placed on the differences in colour.
The museum dates this fitting within a span from 1700 to 1850 and credits its acquisition to J. H. Wade. Neither statement specifies how the material name was determined. An online record need not publish an analytical report in order to make a well-informed identification. A curator may know a great deal from examination, documentary comparison and long familiarity with related work. But the reader who wants to say that a particular Cleveland object was tested by a particular method needs an object-specific source for that extra statement. An analysis of a similar fitting in another country is a point of comparison. It cannot become the missing scientific report for the Cleveland example.

Cleveland's collar with a tiger in bamboo, accession 1920.465, makes the point almost at a glance. The dark rim encloses a scene animated by contrasting passages of warmer colour. The collection calls the material shakudō. The label can identify the principal ground in a few characters, even though a viewer can distinguish more than one visible substance or surface on the ornament. A sentence declaring that every visible portion consisted of the same alloy would say considerably more. Even when a public catalogue displays only one material name, that economy of description is no reason to force the entire object into a single chemical category.
A third work, accession 1920.1963, is also classed as a shakudō kashira. Dark curved forms cross its surface and small bright passages draw the eye away from the centre. Both this object and the tiger collar carry the same broad date range and the same donor credit. Those agreements let us put two works in conversation. They do not tell us that the same individual made both or that they were once fitted together. A common category is an invitation to compare the grounds, highlights and relief. It is a poor substitute for a documented association between two individual works.
What a word for a material can mean
In museum descriptions of Japanese decorative arts, shakudō generally denotes a copper alloy containing gold that can display a dark patinated surface. Shibuichi is a copper-silver material. Bronze conventionally refers to a copper alloy associated with tin, though usage in catalogues can vary in precision. These are starting points for reading a record. Looking for the word “gold” in a material list will not tell a reader which part of an object looks golden. A small quantity of gold may belong to the metal underlying a dark area. A bright golden detail and a dark ground can therefore both call for investigation even where only one of them catches the visitor's eye as precious.
There is a particular temptation to treat a traditional name with a numerical association as though it were a result from a balance. The familiar interpretation of shibuichi refers to one quarter. Susan La Niece, Victor Harris and Hiromi Uchida give a much more interesting account of actual objects in their 2014 study. Their thirteen analyses classed as shibuichi span a range of measured silver contents from seventeen to thirty-five percent. These were historic objects, not thirteen independently selected examples meant to define a universal standard. The range provides good evidence against assigning an exact proportion to each object from its name. It leaves intact the usefulness of shibuichi as a historical material term.
The wording of the survey matters. A result counted in a table may represent an examined area on a component. The number of analyses need not equal the number of separate workshops, objects or historic owners. To ask whether one material was more common than another across a whole period would require a sampling design suited to that different question. Museum collections sometimes contain several items from the same enthusiast's cabinet and few examples from an ordinary household. In that setting, a detailed series of measurements can show that a phenomenon occurred without establishing how frequently it occurred across Japan. That is a considerable achievement in itself.
Reading the ant on a London guard
Agnese Benzonelli, Ian Freestone and Marcos Martinón-Torres published a later study in which observations on historical guards could be compared with controlled work on reference material. The authors identified the historical objects by British Museum registration numbers and illustrated the positions studied. One guard is TS.244. A second is 1952,0211.31. A third is 1958,0730.56.d. Writing those numbers out is more than a bibliographical courtesy. The public records have photographs and descriptions; the scientific publication has a drawing of the chosen areas. The numbers allow a reader to compare the evidence across those different documents without having to recognize a guard from memory.
On the second guard the dark animal and the surrounding field posed separate questions. The researchers identified the ground as patinated copper, while the small ant was among the areas classed as shakudō. A short label saying simply that this was a dark copper-alloy guard could be unobjectionable at the scale of a general display. A description of the comparative artistry of the dark passages would require finer distinctions. The artist had brought together areas that could read as dark to the eye yet had different material identities. Being able to see that choice can deepen appreciation of a tiny image that the average gallery visitor might otherwise walk straight past.
On the third guard the base was identified as shibuichi and the dark leaf as shakudō. The pattern of contrast is thus different from the ant example. The evidence does not support a simple rule that identifies everything of one apparent colour on a guard with one alloy. Nor is comparison limited to visibly dissimilar areas. There can be a meaningful contrast between two dark passages that only becomes apparent through changes in light, surface texture or the position from which the object is viewed. The paper's combination of material analysis, microscopy and colour measurement is particularly well suited to this problem of close looking.
The size of an observation introduces a second complication. In their table for the historical guards, the authors describe a portable X-ray fluorescence measurement area approximately five millimetres across. Imagine that area superimposed on a very small ornament. Some of the neighbouring material may fall inside it. The authors explicitly flag this issue for the inlays. The number displayed by the instrument can be perfectly real, yet attributing every part of it to the little creature at the centre would ignore what else lay within the observation. The method has disclosed something about a defined area; the boundaries of the historical motif were not necessarily the boundaries of that area.
Where the colour lives

A colour term in an art historian's description can refer to the overall appearance of a work in a gallery. Instrumental colour measurements refer to particular observations made under specified conditions. An elemental analysis describes the response of materials contributing to the measurement. These statements can inform each other while retaining their separate meanings. A dark metal ground curves away from the light at the edges of a small fitting. A polished high point reflects the environment differently from a recessed passage. The Cleveland kashira illustrates how readily a single photograph can contain both dark brown and nearly black areas on what the catalogue describes with a single material term.
A comparison within the paper makes these distinctions more concrete. The researchers studied both the colour of patinated surfaces and the underlying compositions of their historical examples. They also used optical microscopy to examine very local features. The colour recorded for a sufficiently large region need not be the colour of every minute feature visible under magnification. Think of looking first at the whole tiger collar and then only at the small reflections on the animal. Both views are genuine. Only when the observer supplies the scale does it become clear whether a proposed comparison is between whole objects, grounds, individual motifs or bright points within a motif.
A digital photograph introduces its own translation of these appearances. The lighting used to show shallow relief, the background, the angle of the piece and a viewer's screen all affect what appears dark. A consistent photograph from a museum is invaluable for finding the same element and seeing how it relates to the rest of the design. It is much less informative about whether the very dark ant on one museum's guard has exactly the same material composition as a dark surface in another museum's photograph. Instrumental results can make such a question more specific. They do not eliminate the need to know which surface each photograph actually depicts.
The earlier survey also considered how much the presence of a patina might influence a particular X-ray fluorescence test. For the authors' experimental shakudō comparison, its presence made no significant difference to their result under the conditions tested. This is useful because the researchers asked the question explicitly and had material suited to the comparison. It would be unwarranted to turn that finding into a guarantee for every corrosion product, coating, inlay and historical repair on any copper object. A finding travels most reliably when its conditions travel with it. That is especially true when methods of examination offer the great benefit of leaving an object intact.
What a bronze mirror adds to the comparison

Cleveland's mirror with paulownia crests, accession 1917.660, belongs to a different class of object. The institution dates it to the Muromachi period and lists its technique as bronze. It has a round head and an extended handle. A photograph of its decorated side offers a good change of scale after the minute hilt fittings. Whatever reflections may once have interested a user of the mirror, it is the design on its back that this photograph makes available for study. A reader can connect the designation of bronze with a specific collection object while keeping clear that the photograph presents one face in one state of preservation.
This mirror also complicates the way an exhibition might arrange its material categories. Place it beside the Cleveland fittings, label all four as copper-based metalwork, and a broad family relationship becomes visible. Identify the crisper divisions supplied in the individual records, and one starts to see how a collection distinguishes an older bronze object from particular Japanese coloured alloys. A comparison interested in the apparent darkness of metal ought to retain both levels of description. Otherwise different works may either be collapsed into a single vague category or separated so thoroughly that the reason for putting them side by side is lost.
It is possible to make this exhibition comparison without assigning a precise tin content to the bronze mirror. The Cleveland entry cited here supplies a catalogue designation. The London studies supply analytical results for named works in their respective samples. No row in either paper is presented here as the result for the Cleveland mirror. This restraint pays a dividend. The reader can now see exactly what an additional conservation report might contribute. A dated examination tied to the mirror, with a record of the areas studied and a discussion of any deposits on the surface, would extend rather than merely repeat the information already available on the label.
Following the evidence back to the work
One of the most instructive illustrations in the earlier paper brings together two seventeenth-century fittings under British Museum number 1939,0519.48. The authors distinguish the principal shibuichi pieces from inlaid decoration and from a copper plate associated with the collar. Even an observer comfortable with the names of Japanese alloys can learn from the way that caption attaches a material to a part. It supplies enough detail for another scholar to understand that seeing copper in one component would be quite consistent with the name given to the rest. A carefully written caption can prevent an apparent scientific disagreement before it ever arises.
An exhibition translator faces a comparable exercise. Translating the material field of 1920.1956 as shibuichi retains a historical category; adding a short explanation that it is a copper-silver alloy makes the word accessible. Replacing the name entirely with an exact figure taken from a survey of other objects would lose the category and gain a spurious precision. One could instead put the survey in a separate sentence and identify the examples it examined. A visitor would then understand both what the catalogue calls the fitting and why historical materials that share a name deserve individual study. It is a small change in syntax with a large effect on the account.
Another useful editorial practice is to preserve the notation used for small or undetected signals. In the later paper a table states the reporting threshold for certain elements and distinguishes entries below detection from positive numerical readings. An empty spreadsheet cell in a subsequent transcription could stand for either of these or for something never examined. Treating every such cell as a measured zero would generate results the researchers never claimed. The distinction can seem technical until a historian asks whether the apparent absence of an element unites a set of pieces. Then the original table heading and footnote become as important as the rows themselves.
Returning to the little rice scene, a viewer now has a more precise reason to look slowly. The material named by the museum, the arrangement of contrasting forms and the play of reflected light can all be appreciated without pretending that a photograph is a laboratory report. The studies in London place a new sort of question within reach. Which areas would a conservator have to distinguish if the material relationships in this particular scene were at issue? The answer begins with the composition of the image and the limits of the surviving evidence. The work's most engaging quality, its ability to gather a whole scene into a small dark ground, remains in plain view.
Questions about museum alloy identifications
Can a black surface alone identify shakudō?
No. A British Museum study distinguishes a patinated copper ground from a shakudō ant on one guard. Another guard combines a shibuichi ground with a shakudō leaf.
Does every shibuichi object contain exactly one quarter silver?
The 2014 survey reported seventeen to thirty-five percent silver across thirteen shibuichi analyses. The historical term alone cannot establish the measured proportion in an individual fitting.
Why does a report mark the point where a reading was taken?
The measured area can contain more than one material. The 2023 paper specifies an approximately five-millimetre spot for its historical objects and notes that surrounding material could contribute when a small inlay was examined.
Which Cleveland object depicts rice cultivation?
Accession 1920.1956 is a kashira with a rice cultivation scene. The museum lists shibuichi, and the image and material designation refer to this specific object.
Were the Cleveland pieces tested in the London study discussed here?
The three historical guards in the 2023 comparison have British Museum registration numbers. The Cleveland pieces illustrate their own catalogue descriptions; those London results cannot be assigned to them.
What does a value below the detection limit mean?
Under the conditions of that observation the element could not be reliably detected at the stated level. It should be distinguished from an area never examined and from an assertion of complete absence.
Museum films
JASA: Japanese Art Society of America — Samurai Splendor: Sword Fittings from Edo Japan (Lecture by Markus Sesko)
Birmingham Museums Trust — Hoard Conservation Blog: X-Ray Fluorescence (XRF) Machine
Collection records and further reading
- Benzonelli, Freestone and Martinón-Torres: historical guards, figure 1 and table 2
- La Niece, Harris and Uchida: Alloys of Japanese Patinated Metalwork
- British Museum: guard TS.244
- British Museum: guard 1952,0211.31
- British Museum: guard 1958,0730.56.d
- Victoria and Albert Museum: conservation science methods
- Rabbits on the Moon: Ishiguro Masayoshi and Edo Metalwork
