Japanese Sword Lamination: Comparing Maru, Kobuse, Honsanmai, and Shihozume Cross-Sections

Two ways to see through an old sword

An engineering paper published in Toyama in 1966 presented observations from fourteen pieces of a single old Japanese sword. More than half a century later, an instrument in a Japanese research facility produced a set of views through a sword signed Norimitsu while leaving the object whole. Both publications concern the inside of a blade. Only one set of sections could be placed in sample drawers. The others exist as reconstructed data. This difference in what the word section means is a good place to begin a history of Japanese steel and the names given to its structures.

Readers encounter four names especially often. Maru describes a conception of the body as a single principal material. Kobuse names a relationship between an exterior and an interior. Honsanmai belongs to the vocabulary of composite structures distinguished into several parts, and shihōzume extends the distinction around the perimeter. These are historical and descriptive categories. A neat illustration of all four helps explain why authors use different words. It supplies no evidence that all four have been identified in any collection of four surviving blades, or that a particular smith used one throughout a career.

Daihannya Nagamitsu on display at the Tokyo National Museum
Daihannya Nagamitsu on display at the Tokyo National Museum. Kakidai. CC BY-SA 4.0.

Tokyo National Museum's Daihannya Nagamitsu illustrates the gap between two kinds of knowledge unusually well. The museum can identify a distinguished named work and let the public examine a detailed photograph of its outer form. That is already a substantial achievement of cataloguing and photography. It does not place a metallographic section through the work in a reader's hands. A name, an image and a material examination are three separately identifiable records. The research on Norimitsu belongs to a different object and cannot be lent to the Nagamitsu by proximity on a gallery wall or a computer screen.

What the names describe

Kawagane and shingane are often translated as skin steel and core steel. The English words are metaphors, useful because readers can imagine an outside and an inside without a diagram. The difficulty begins when they suggest two perfectly uniform substances divided by a line as thin as ink. An author discussing a physical sample can describe variations within a region, an irregular transition or uncertainty about a boundary. The name of a zone identifies a relationship in the object. It does not predict the appearance of every small field that a microscope might reveal within it.

What a construction name can establish
Term What the label describes What could support the attribution
Maru A single body in the conventional diagram A section can show local variation even within material classified as a single body.
Kobuse An internal core and an outer body distinguished in the historical vocabulary An identified section or an imaging study with enough resolution to separate the regions.
Honsanmai A composite described using distinct side, core and edge regions Evidence from the object itself, together with the terminology used by its examiner.
Shihozume A composite diagram that distinguishes the back as well as the sides and edge The sampled position, visible interfaces and the limits of the examination.

The categories are historical models. Instrument readings and micrographs describe particular objects and sampled areas.

Sources · Jerzy Piaskowski, Historical Metallurgy, volume 27, issue 2 · Matsumoto and colleagues, neutron study of the signed Norimitsu

Hagane is used for steel, and in descriptions of certain composite arrangements a distinct term may emphasize the edge material. Terminology depends in part on what distinction an author is trying to draw. A description of honsanmai that distinguishes the body, its sides and a separately identified region is operating at a different scale from a description of a single crystal. Shihōzume draws attention to an additional distinction around the back. An interpreter should establish which system of names the author employs before deciding whether two apparently conflicting descriptions actually disagree about a piece of metal.

Maru is an especially instructive case. A statement that a structure is treated as one main body does not imply that its steel is perfectly homogeneous under magnification. Historical metal can contain small inclusions or variations of texture without ceasing to belong to that descriptive category. Conversely, a smooth, apparently uniform exterior does not establish that the whole interior has one history. An ordinary collection photograph was never intended to settle that question. The two cases demonstrate why a label for large regions of material cannot simply be scaled down into a description of microscopic features.

One also encounters tsukurikomi and kitae in accounts of Japanese blades. Broadly, these words refer to matters of formation or forging, but authors do not always divide their subjects in the same way. A gallery panel discussing outer form may emphasize the ridge or outline visible on a polished blade. A metallurgical paper may ask which internal regions are distinguishable. Two authors can therefore describe different features while using related vocabulary. A translation that says only construction can conceal the transition from a visible external form to a proposed internal arrangement.

The date on a paper and the date on a specimen

Jerzy Piaskowski's study in Historical Metallurgy appears online under a recent publication date, yet it belongs to the journal's 1993 volume. A present-day search engine can make this an unexpectedly important distinction. The paper records a physical investigation and reflects the scholarship available when it was written. The date when a journal placed or republished a file on its website is part of the history of access. Neither date automatically dates the specimen. For an account of research on Japanese swords there are thus at least three chronologies on the very first page.

Piaskowski distinguished regions of differing carbon content in the sampled object and reported a corresponding difference in microstructure. The author considered the outcome compatible with alternative traditional readings, printed there as makuri-kitae or kofuse-kitae. The small word or is the part most likely to disappear when the paper is summarized. A chemical difference is an observation about a specimen. Selecting a historical name that accounts for it involves interpretation. The microscopy made the contrast more certain without obliging the researcher to resolve every ambiguity about historical terminology.

There is another reason to keep the original spelling in view. Romanization has changed between authors and languages. A database search confined to one contemporary spelling of kobuse can miss a discussion under kofuse. At the same time different spellings do not guarantee that every author used the term in precisely the same sense. An annotated museum bibliography can preserve the wording used in an older paper and add modern search terms separately. That simple arrangement makes an unfamiliar source discoverable while preventing the indexer's decision from becoming a new scientific finding.

Blade for a Sword (Katana) (2001.574), Metropolitan Museum of Art
Blade for a Sword (Katana) (2001.574), Metropolitan Museum of Art. The Metropolitan Museum of Art, Open Access. Photograph DP372017.

The Metropolitan Museum's photograph of a blade accessioned as 2001.574 belongs here as a comparison of document types. A clear image lets the reader see an object's proportions and the way a museum photographer used light to describe its surface. It is a record of this particular museum object, not a picture of Piaskowski's sample. Without that distinction an illustrated article can quietly attach the prestige of a collection object to the outcome of a completely different experiment. The titles, accession numbers and references supplied under the images keep those histories separate.

Fourteen pieces from one object

The Toyama study's fourteen pieces represent a remarkable investment in describing variation within one old object. A single specimen viewed repeatedly along its length gives the observer an opportunity to ask whether a feature continues from one position to another. If two sampled places look different, the difference can be genuine evidence about spatial variation. It need not mean that either reading is wrong. If several look much alike, their agreement strengthens the account of that object. It does not transform the fourteen pieces into fourteen independent witnesses for a regional school.

This point matters especially when collectors ask how common a particular construction was in a given century. The question requires a population of independently identified works. Fourteen locations on one blade answer a different question from fourteen blades each studied at one location. The latter group would still need a defensible account of how it was selected. A laboratory does not receive a random sample of everything ever commissioned. Access depends on what survived, who held it, its condition and what intervention the owner or custodian considered acceptable.

The Smithsonian's Museum Conservation Institute indexes an earlier investigation by Cyril Stanley Smith that concerned three older blades and one modern example. Here comparison across objects was built into the investigation at a different level. A modern specimen can help clarify which features are visible by a particular method. It cannot retroactively document the workshop history of an earlier artifact. A useful study can put them together precisely because their statuses differ. The date, identity and available documentation of each remain essential to any later use of the results.

There is a human dimension to the survival of samples. A small labeled fragment inherited by a university can outlive the researcher who prepared it. Its continued scientific value may depend on an unremarkable envelope, a handwritten position number and a photograph showing the original object. If the associations break, sophisticated new instruments might produce exquisite new measurements of the wrong location. Archives preserve relationships between things. In this field a yellowing sheet of paper with a cross-reference can be as indispensable to the next investigation as the steel in its drawer.

What a micrograph remembers

A micrograph can bring minute features into extraordinary clarity. Its field of view is correspondingly small. A scale bar says how large the visible structures are. It does not tell the reader where that field lay in the original object, which way up the blade was oriented or whether the area was representative. For those questions one needs the accompanying plan, the sequence of samples and the author's reasons for choosing particular positions. A striking illustration divorced from its caption can retain its beauty while losing much of its value as historical evidence.

Three lines that look similar on a screen can represent entirely different things. One may be a real transition observed in a sampled material. Another may be the outline of an area drawn by a scientist to indicate the field of a measurement. A third may have been added by a later illustrator to show how a named type is conventionally imagined. They could all be printed in black on white. Reading the figure legend restores the distinctions. In a museum exhibition, typography and captions can do as much as magnification to help a visitor see them.

The same care applies to the word grain. In accounts of Japanese art it may describe an effect visible in a polished surface. In crystallography it can refer to the orientation of crystals in a material. A research paper might use crystallite size for a quantity estimated by an instrument. These observations are related to the history of a material, yet they are not interchangeable units that can be counted from an ordinary photograph. The most informative comparison often begins with a simple question about the physical property each image was designed to register.

The Norimitsu investigation

A team working at J-PARC studied a signed work of roughly forty-five centimetres in blade length. The inscription identifies Bishū Osafune Norimitsu. The study used RADEN, an instrument for energy-resolved neutron imaging, and combined two approaches to investigate the material. Through reconstructed views the researchers identified an internal pattern characteristic of kobuse. Their analysis of the distribution of crystallite size supported that interpretation. The strength of the published finding lies in the relationship between the separate observations and in the ability to identify the object they concern.

In a transmission measurement the signal can carry information from a path through material. In a tomographic reconstruction observations from several directions contribute to a representation of a volume. A researcher can examine planes within that volume, rather as a reader can select different pages of a book. The analogy has a limit. A page of a book existed on paper before anybody turned it. A tomographic slice is calculated from measurements. Its appearance depends in part on how the signal was obtained and interpreted, and the researcher must explain the contrast being shown.

The same study describes a pattern associated with hardening. This is a second result about the material's history. It would be a mistake to label every contrast in the data as a joining boundary merely because the paper also discusses kobuse. Several processes can leave a record in one object. Phase, orientation and region of material answer different questions, even when they contribute to neighboring areas of the same illustration. A good figure caption identifies which of these is under discussion and makes it possible to follow the reasoning back into the results.

Koryū Kagemitsu in a historical catalogue photograph
Koryū Kagemitsu in a historical catalogue photograph. 文化財協会. Public domain.

Koryū Kagemitsu, pictured from Tokyo National Museum's catalogue, provides a useful check on a reader's growing confidence. A familiar regional name or a visual affinity might suggest a worthwhile future comparison with a signed Osafune work. That is an argument for posing a question about a second object. It does not authorize copying the Norimitsu result into its catalogue record. The moment a museum is able to attach the name of a tested object to a published observation, it also acquires a duty to avoid extending that observation to a neighbor by implication.

A history of comparison rather than a ladder of quality

The time-of-flight neutron diffraction study published in the Journal of Analytical Atomic Spectrometry asked how quantitative findings might be compared across Japanese blades. Its authors also called for further work on reference standards. Such a qualification reveals what makes comparison scientifically useful. If the behavior of a known material under a given measurement is documented, the reading of an unknown can be better constrained. An increase in the precision of the instrument does not by itself provide the missing historical identity of the reference object.

An energy-resolved imaging paper in the same journal subsequently used reconstructed slices and volumes to investigate characteristics of different works. Read beside the diffraction study, it illustrates how a field accumulates different forms of evidence without retiring the old ones. A small sampled area might preserve information available to optical microscopy that a whole-object method describes more indirectly. A scan can place a contrast within the larger form in a way that an isolated specimen cannot. Researchers gain by comparing the claims each method can sustain at its own scale.

Blade and Mounting for a Dagger (Tantō) (91.2.32), Metropolitan Museum of Art
Blade and Mounting for a Dagger (Tantō) (91.2.32), Metropolitan Museum of Art. The Metropolitan Museum of Art, Open Access. Photograph LC-91_2_32-006.

A mounted dagger in the Metropolitan Museum makes visible a more ordinary aspect of the same problem. A complete ensemble has a visual identity formed by metal, decorative furniture and its surrounding materials. An investigator interested in steel structure and a curator interested in the ensemble may choose quite different photographs of it. Both choices can be faithful to the work. But if a label concerns only the mounting date, transferring that date to a claim about the steel creates a conclusion neither specialist intended. Museum database fields have boundaries for good historical reasons.

A comparison of maru, kobuse, honsanmai and shihōzume therefore ends unevenly. The Norimitsu investigation provides positive evidence for one named interpretation in one signed object. Piaskowski's sampled specimen preserves a material distinction with more than one historical reading. The group of studies reviewed here does not supply an equally secure, independently documented example of every category from every period. Nor do the papers test the quality or construction of a contemporary object whose name happens to contain sanmai. A distinction that supported a historical interpretation has no automatic commercial successor.

Consider how a museum might exhibit the material without concealing that uneven evidence. It could give a physical sample its magnified image, its recorded original position and a label identifying the study. A neighboring screen could present a digital slice of the Norimitsu with the instrument and measured quantity named. Conventional diagrams could appear on another panel, clearly dated and credited to their authors. The visitor would see real objects, observations and explanatory models together, with enough information to compare them without treating them as equivalent illustrations.

The reward for that care is a reversal of the usual picture of expertise. The most satisfying answer to the question of what lies inside an old blade is sometimes quite limited in scope. It names the particular object, the evidence recovered from it, the reading supported by that evidence and the remaining uncertainty. Such an answer can survive a new measurement because it tells future researchers exactly what was done and what remains to be investigated. An immaculate diagram without an object may be remembered more easily. A small and traceable observation is a better place from which to write history.

Frequently asked questions

What is san mai steel in a history of Japanese swords?

San mai is a term for a three-part composite conception. Its appearance in a catalogue or diagram does not by itself demonstrate the construction of an individual historic object.

What is laminated steel in a museum description?

The term indicates a material interpreted as having combined portions of steel. It can refer to features at different scales, so the author’s definition and evidence matter.

What is metallography?

It is the study of the internal structure of metals. Published metallographic images can document local differences in a surviving specimen.

What is a cross-section in a sword study?

It is a view through a defined plane in an object. A physical sample and a digitally reconstructed slice can show corresponding positions while recording different material properties.

What is neutron diffraction used for in heritage research?

It provides information about crystalline materials. The phases and orientation represented in a measurement can help researchers interpret an object while retaining its exterior.

What is a micrograph, and can it identify an entire blade?

A micrograph records a small field at a stated scale. Assigning a construction to the whole object requires the location of that field and supporting evidence elsewhere.

Films and talks

PSI Paul Scherrer Institut — Neutron Imaging at SINQ

The Met — Conserving Japanese Works of Art in Foreign Collections

Collection records and further reading

Related Articles