Katana and European Longsword Metallurgy: Comparing Tatara Steel, Differential Quenching, and Balance

A blade enters a laboratory

A Japanese sword can arrive at a museum with an extensive object file and still pose a surprisingly elementary scientific question. What is the metal like inside? The file may name the smith, transcribe an inscription, describe the visible surface and trace a change of ownership through a shogunal household. None of those documents is a section through the steel. On the other side of the world a European sword from a river deposit may have precisely the opposite history. A damaged edge may have allowed an investigator to study its structure while the name of its maker and the date it entered the water remain unknown.

Those two objects would make a fascinating comparison, provided we remember which questions each can answer. A katana is a historical object category. A longsword is another, with a history of terminology that varies across collections. Neither is the name of an alloy. The words tatara, martensite and pommel belong to three different stages in the history of a complete object. One concerns the production of iron, another the state in which some of that iron is found under examination, and the last a component that can remain after a grip has decayed or been replaced.

The comparisons that follow begin with Japanese work and then ask what European evidence can reasonably be placed beside it. The aim is to make sense of actual catalogue measurements and published analyses. Without that discipline even an impeccably accurate number can produce a false conclusion. A number measuring the mass of a complete mounted object says little about the hardness of a few square millimetres of steel. A meticulously measured microstructure cannot tell us where the centre of mass of a lost mounting once lay. A museum needs both sorts of record and a vocabulary for keeping them apart.

What the word tatara identifies

Tatara refers to a Japanese tradition of iron production closely associated with iron sand and charcoal. Historians encounter its traces on several scales. There are production sites, records of people and supplies, historical samples of different furnace products and the finished objects that may have passed through many workshops. A scientific paper in the journal Tetsu-to-Hagané includes historical data for pig iron and steel from two late nineteenth-century furnaces. Those entries are useful precisely because the authors identify the places, the date and the products compared. A table of that kind is evidence for variation within a tradition.

Tamahagane is often translated as jewel steel. The term refers to selected material and cannot stand for every piece of metal or slag associated with a tatara. Ore, iron sand, a bloom or consolidated mass, selected steel and a completed blade represent different populations from which a sample could have been taken. Reading a composition reported for one as if it applied to another would be like assigning the average composition of a quarry to every pigment in a painting. Evidence about a source is important. The route from the source to an individual art object is a historical question of its own.

Repeated working and joining of material belong to the workshop history of many Japanese blades. They help explain why the appearance of worked steel can have artistic as well as technical significance. A visible hada is part of what connoisseurs describe when they catalogue a blade in good polish. It is not a chemical assay of the iron sand or a numerical count of the steps in the workshop. A maker’s ability to produce a surface that recalls an earlier regional tradition complicates the inference further. Similar visible results do not require identical raw material or an unbroken chain of provenance.

Blade and Mounting for a Dagger (Tantō); 91.2.54, Metropolitan Museum of Art, view 3
Blade and Mounting for a Dagger (Tantō); 91.2.54, Metropolitan Museum of Art, view 3. The Metropolitan Museum of Art, Open Access.

The dagger mounting in the illustration brings the problem back to the gallery. Its museum description identifies the blade category and shows a complete object assembled from materials with different histories. The steel could be the subject of a metallurgical investigation, while the surface of the mounting could require expertise in lacquer, wood or decorative metals. The image is evidence for the appearance and cataloguing of this particular object. It is not an illustration of a laboratory specimen unless the study actually names the same accession. That small distinction matters in publications which place an attractive polished blade next to an image of a microstructure.

Carbon as a historical witness

The Tetsu-to-Hagané paper on radiocarbon ages of swords takes advantage of a link between fuel and metal that a viewer cannot see. Carbon associated with charcoal production can preserve a signal suitable for radiocarbon measurement. The researchers compared calibrated calendar ranges with chronological evidence for the smiths identified in inscriptions. Their results illustrate both the richness and the limitations of the method. A calibrated measurement may produce more than one possible interval. The age of the wood used for charcoal and the recurrence of a smith’s name over generations also have to enter the interpretation.

This relationship between a laboratory result and documentary attribution is worth pausing over. A swordsmith worked at a moment in human history. The carbon in the object reflects processes that began earlier, when a tree was growing. A signature can provide a name, but a name may belong to successive generations in a workshop. Finding an interval consistent with the attribution is meaningful evidence. It is still a different proposition from having read the calendar date of manufacture directly from the metal. Good experimental accounts preserve that distinction because it tells the next investigator which uncertainties remain.

The same caution applies to claims of exceptional purity. Analyses of particular Japanese blades have reported low concentrations of several elements other than carbon. To use the word pure in a comparison, an author has to specify the element, the material measured and the method’s detection limits. A steel containing carbon is not chemically pure iron. Slag inclusions tell a different story again. A small region selected for analysis might contain an inclusion that a nearby region lacks. The question of what can be said about a whole blade begins with the location and size of the area actually studied.

A pattern of treatment is different from a recipe

The term hamon directs a visitor to the boundary associated with the hardened region along a Japanese blade. The appearance of that boundary can be extraordinarily diverse. In metallography martensite refers to a structure that can arise through transformation of austenite during sufficiently rapid cooling. Pearlite names a different arrangement involving ferrite and cementite. These are names for observations at an appropriate scale. A viewer who sees a bright line along a polished blade is seeing light interacting with a surface that has its own history of finishing as well as the underlying material.

The historical use of a differential treatment helps explain why two places on the same object can yield different observations. Accounts of Japanese craft associate this practice with a coating and with water cooling. It would take more evidence to identify the precise treatment of an individual museum piece solely from its finished appearance. A modern process description can demonstrate one way a contrast arises. It cannot establish which choice a medieval worker made for a particular surviving blade. This is one reason laboratory researchers carefully distinguish measurements of their own specimens from the accounts that motivated the investigation.

Koryū Kagemitsu, Tokyo National Museum, via ColBase
Koryū Kagemitsu, Tokyo National Museum, via ColBase. Tokyo National Museum, via ColBase.

Yoshiaki Kiyanagi’s published investigation uses neutrons and muons to study a Muromachi-period sword. The methods offer complementary information on crystallographic variation and the distribution of carbon. The report describes inhomogeneity within the work and differences between the edge region and the back. The tang also showed a different pattern from parts of the blade. These findings have an advantage over a generalized diagram of a katana. They describe observations made in specified parts of one physical work. They also have a limitation that a reader should welcome. One examined work provides no national average.

The photograph of Koryū Kagemitsu illustrates why choosing a specimen is an important step in any investigation. A well-known Japanese blade in an accessible collection is a convenient reference for readers who want to study a polished surface. Its availability as a photograph does not make it the Muromachi sword in Kiyanagi’s study. Naming the photographed work and the examined object separately prevents a subtle error that is surprisingly easy to repeat. A public can appreciate a famous object and understand an experimental result without being asked to believe that the two have the same history.

An archaeological comparison from Europe

In their study of eight early medieval swords David Edge and Alan Williams found more than one material arrangement. Some pieces were made from a single piece of steel, others had a more complex construction. That result alone is enough to show the weakness of a simple contrast between a composite Japanese tradition and a uniform European one. The passage from the early medieval evidence to the fifteenth-century longsword also spans a substantial part of European history. A finding about an early medieval sword remains valuable on its own terms without turning it into a measurement of a later object.

Williams’s book The Sword and the Crucible follows European sword metallurgy across time and includes evidence for bloomery iron, more complex constructions and material associated with crucible steel. In a bloomery, reduction produces a mass of iron and entrapped material that must undergo further work. A blast furnace belongs to another production history, in which liquid iron and later refining become important. The existence of blast furnaces in medieval Europe does not identify the feedstock of any particular European sword. Neither the age of an object nor the current location of the museum that owns it supplies the missing link.

A study by Williams and Tomaž Lazar of material in the National Museum of Slovenia makes the question of scale particularly clear. Some examinations used surface hardness testing. Metallographic investigations involved small samples taken from specified portions such as the edge or core. The latter method can reveal local phases and inclusions that a silhouette or mass measurement will never show. Removing a sample is also an intervention into an archaeological object. The decision about which pieces can be sampled affects which kinds of historical work become best represented in the scientific literature.

The surviving European evidence includes different combinations of ferrite, pearlite and hardened structures, varying with the object and the part sampled. Tempered martensite is a possible finding with a specific metallurgical meaning. It is quite a different matter to infer an oil quench for an entire continent from such a finding. Condition, subsequent heating, restoration and the limited size of the sample all affect the history an analyst can recover. On both sides of the comparison, authors should identify which observations narrow the possible histories and which details remain open to more than one interpretation.

Hand-and-a-Half Sword, 2006.564
Hand-and-a-Half Sword, 2006.564. The Metropolitan Museum of Art, Open Access.

Hardness, stiffness and energy each have units

An often repeated claim assigns a figure of sixty HRC to a Japanese edge and proceeds as though this explains how an entire blade behaves when bent. HRC is a Rockwell hardness scale. A measured value records resistance to indentation under the conditions of that test. It does not provide an elastic modulus or a measure of energy absorbed by the assembled object. A researcher who wants to report elastic stiffness must specify what was loaded and what deflection was observed. Energy and fracture behaviour involve further quantities. The words sound related in ordinary English because materials can be described loosely as hard or tough. The experiments answer distinct questions.

Microhardness studies on historic steel frequently report Vickers values instead. The authors may specify a load and an array of positions across a polished section. A succession of values can reveal a gradient within the specimen. Comparing that sequence with a single Rockwell number of unknown origin discards most of its information. Nor does the highest reading in a section characterize every point on the blade. A scientifically informative caption includes the scale, the position and the identity of the sample. A caption that gives only a country and a number has lost the conditions that make the measurement interpretable.

There is a similar distinction between a transverse section and the profile of an object along its length. The first records a slice at one position. The second shows changes of width and depth over a much greater distance. Neither in isolation determines the distribution of mass in a complete mounted object. Surface and section also vary with what has happened since manufacture. A long history of polishing removes some material, corrosion can remove more, and repairs can introduce materials with quite different properties. A present-day measurement can be entirely accurate while leaving its original value unknown.

A pommel inscribed MARIA

The Metropolitan Museum’s hand-and-a-half sword 2006.564, dated about 1400 to 1430, is a particularly good object with which to test assumptions about a European hilt. Its pommel carries the word MARIA, an invocation of the Virgin Mary. The museum also draws attention to the unusual shape of the pommel and the reverse curvature of the guard. A caption that calls the pommel simply a counterweight would miss an inscription someone expected another person to read. A Japanese tsuba can likewise be important to the distribution of material within an ensemble while remaining a work of religious, narrative or decorative art.

The Met gives the mounted European object’s overall length as 124.8 centimetres, its blade length as 97.8 centimetres and its mass as 1,560 grams. The three quantities are each associated with a stated object or component. None of them, alone or in combination, locates its centre of mass. The Wallace Collection provides an instructive contrast in its record for A460, an arming sword rather than a hand-and-a-half sword. It explicitly places the balance point 6.5 centimetres forward of the guard block. That is an example of a properly referenced museum measurement, not a value that can be transferred to the Met sword or to an unspecified katana.

Hand and a Half Sword, 1988.26
Hand and a Half Sword, 1988.26. The Metropolitan Museum of Art, Open Access.

The Wallace Collection’s A476 has yet another useful description. It is called a longsword, dated about 1525 to 1550, recorded at 1.16 kilograms and described as single-edged. Even the name longsword, then, does not guarantee the blade profile a reader may have pictured. If a comparison drew its mass from this entry, its balance point from A460 and its photograph from the Met, the result would be a convincing description of an object that never existed. This is why an accession number alongside a value is more useful than a promise that the number represents a typical national style.

What a fair comparison leaves open

A researcher could assemble a useful comparison table for one Japanese katana and one European longsword. Each row would begin with its accession number and specify whether the piece is bare or mounted. A laboratory row would identify the location examined and the scale of the test. A balance entry would name a reference point and the state of the grip, guard and pommel or other fittings. Empty cells would be valuable information too. When a public catalogue gives dimensions without a measured balance point, the honest entry is that the value has not been published there.

An especially instructive empty cell concerns toughness. One might have an image of a beautiful hamon, a radiocarbon interval and a securely dated ownership document for a Japanese blade and still possess no fracture measurement for it. The museum has compelling reasons to prefer that state of affairs to a new destructive experiment. For an archaeological fragment there may instead be a carefully studied section and little or no knowledge of its original mounting. Each object gives something different to the history of metallurgy. Recognizing what each one can spare is part of treating it responsibly as a cultural object.

The most productive comparison between Japan and Europe ultimately concerns the way people learn from the metal that has survived. In Japan, named blades and a long tradition of connoisseurship can guide a researcher toward features worth investigating. Across European collections, excavated objects and published analytical series provide other opportunities for testing assumptions about material and construction. Inscribed steel, museum records and laboratory work can illuminate one another. They do so most clearly when a crescent on a polished surface, a grain under a microscope and a figure measured from a guard are each allowed to mean exactly what was observed.

Frequently asked questions

What is tatara steel?

Tatara refers to a Japanese ironmaking tradition associated with iron sand and charcoal. Its products varied, and tamahagane names selected steel rather than every material from the furnace.

Were all medieval European swords made from the same type of steel?

Metallographic studies document several material arrangements and different degrees of hardening. Individual specimens and their dates matter.

Is a hamon a hardness measurement?

No. It is a visible feature associated with the hardened region. Numerical hardness requires a documented test, scale and position.

Does 60 HRC describe how a whole blade bends?

Rockwell C hardness measures resistance to indentation at a test location. Stiffness, toughness and energy absorbed during bending are different quantities.

Can a museum infer a quenching liquid from a photograph?

A photograph can document visible appearance. Identifying a historical treatment requires additional physical or documentary evidence, and the exact liquid may remain uncertain.

What does the point of balance in a museum catalogue mean?

It locates the centre of mass of the object in the condition measured, relative to a stated reference. Measurements from a bare blade and a mounted sword are not directly interchangeable.

Films and talks

PSI Paul Scherrer Institut — Neutron Imaging at SINQ

Birmingham Museums Trust — Hoard Conservation Blog: X-Ray Fluorescence (XRF) Machine

Collection records and further reading

Related Articles