Mokume-Gane and Nanako-Ji: Layered Metal Grain and Punched Roe Grounds in Japanese Sword Fittings

Two small guards and two questions about their surfaces

In one Japanese sword guard at the Metropolitan Museum of Art, bands of differently coloured metal turn across a field about the width of three fingers. On another, a gold dragon rises from clouds and waves over a dark ground crowded with tiny rounded marks. Both objects reward magnification. They also reward putting the magnifier down. Seen from across a display case, the first guard offers the rhythm of a sheet of figured wood. On the second, the quiet definition of the ground makes the dragon easier to see. The details serve a composition that can still be read by someone who has never heard the names of the techniques.

The first is a useful introduction to mokume-gane, the Japanese tradition of patterned laminated metal. The second gives us nanako-ji, the field of punched relief named for its resemblance to fish roe. Grain in the English title of a collection record can suggest either a pattern in a material or a multitude of tiny raised grains. Knowing which sense the cataloguer intends changes where to look. In one case follow a boundary between colours. In the other follow the light across a row of small protrusions. One observation concerns what materials appear at the surface, the other how that surface rises and falls.

Sword Guard (Tsuba), 14.142
Sword Guard (Tsuba), 14.142. The Metropolitan Museum of Art, Open Access.

What the Met records about a laminate

The Metropolitan Museum identifies its guard numbered 14.142 as Japanese work from the first third of the nineteenth century. The material line names a laminate along with iron, copper and shakudō, the copper and gold alloy familiar from Japanese fittings. At around seven centimetres high it is small enough for a single curve of colour to look particularly assertive in an enlarged photograph. The collection gives the weight and thickness as well as the overall outline. Those measurements return the work to the world of objects someone once chose to wear, collect or give away.

Comparing patterned metal and punched grounds
Surface Feature to examine Museum reference
Mokume-gane Colour boundaries running through the metal Sword guard, 14.142
Nanako-ji Punched dots in the dark ground Dragon Between Waves, 91.1.786
Relief decoration Raised figures above the ground Examine the dragon separately from its background.
Sword Guard ( Tsuba ) Depicting Dragon Between Waves (波龍図鐔) (91.1.786), Metropolitan Museum of Art
Sword Guard ( Tsuba ) Depicting Dragon Between Waves (波龍図鐔) (91.1.786), Metropolitan Museum of Art · The Metropolitan Museum of Art · Public Domain / Open Access

Mokume is a word borrowed from looking at wood. Timber cut in different directions can display long bands, folded shapes or small eyes apparently set within a larger field. The choice of name asks us to remember those observations when looking at metal. It does not mean that a piece of wood served as a stamp, or that the colours of a particular species were literally reproduced. The artist had access to contrasts that a joiner would never find in a board. An area that recalls growth rings in one part of a guard may read as flowing water in the next.

A useful thought experiment is to imagine the illustration printed in a single dark ink. Where areas of different metals have similar brightness, two colours might collapse into one shade of grey. The design could appear much simpler than it does in the colour plate. Change the direction of the illumination and small differences in reflection can bring some of those boundaries back. The catalogue's material description becomes especially valuable when a historical photograph has lost the very information that once made the design legible. An early monochrome plate can document the shape of the work quite accurately while underrepresenting the extent of its pattern.

Shōami Denbei and the evidence of a name

The Japan Mokumegane Research Institute locates an early chapter of this history in Akita in northern Japan. Its account places Shōami Denbei in the service of the Satake house after a period of training in Edo. A specialist fitting in its account bears his name and puts bands of contrasting metal onto the small field of a kozuka, the handle associated with an accessory knife. Such a surface is particularly revealing. It offers far less room than the broad side of a vessel, yet it gives the eye enough continuous space to follow a single band through changes of direction.

The institute also draws attention to carved lacquer as an antecedent for guribori, a related family of visual effects. A recess cut across successive colours can show a succession of thin outlines along its sides. The comparison helps explain how a craft idea might cross between workshops that handled very different materials. Lacquer and metal can both present the eye with coloured strata. Their behaviour, their surfaces and even the sensation of lifting the finished object in the hand remain quite different. A metalworker borrowing an idea about colour could therefore create something materially new while inviting a viewer to remember a treasured lacquer vessel.

In the Metropolitan Museum a separate guard, accession 29.100.1043, is catalogued with an inscription by Shōami Denbei. Its material entry gives iron and copper. Seen next to 14.142, it is an instructive counterexample to a tempting shortcut. A personal name associated with the early history of mokume does not make every object bearing that name an example of the same technique. The later guard is explicitly catalogued as laminate. The earlier work supplies evidence about a named fitting and the materials the museum records for it. One object contributes the material description we seek; another anchors a discussion of names and workshop history.

The dragon and its field of fish roe

The Met's guard with a dragon between waves presents a quite different exercise for the eye. Its record locates the punched ground between the images of waves and clouds and on the rim. That last detail is worth following. A viewer who goes straight to the golden creature at the centre could easily miss the way a comparatively quiet texture continues at the outer boundary. Walking the eye around the edge reveals a continuity between widely separated parts of the composition. The dragon's visual prominence depends in part on this unobtrusive work elsewhere.

The Ashmolean Museum's catalogue of the Church collection describes the traditional rounded mark in terms of a punch with a hollow end. The tool leaves a small relief in the existing metal. In photographs the result can look like a field made by attaching innumerable specks. The distinction matters whenever nanako is translated as granulation. In other metalworking traditions that word can refer quite specifically to applying separate little spheres. For an English label on a Japanese fitting, punched fish-roe ground conveys both the appearance and enough of the physical distinction to prevent a genuine misunderstanding.

Consider what happens to the texture when the sun or a display lamp moves. Each small rounded surface carries a bright side and a shaded side. Thousands of those pairs of light and shadow alter together. A field too fine to resolve at an ordinary viewing distance can consequently take on a gentle lustre. On the raised dragon highlights read as distinctive features of its body. On the ground they merge over the area around it. The artist has given light two roles within a single composition, without demanding that the viewer count one mark or even know the name of the background technique.

Sword Guard (Tsuba) Depicting Dragon Between Waves (波龍図鐔), 91.1.786
Sword Guard (Tsuba) Depicting Dragon Between Waves (波龍図鐔), 91.1.786. The Metropolitan Museum of Art, Open Access.

How many grains will fit in a square centimetre?

It is quite reasonable to wonder how dense the tiny marks are. A number becomes meaningful only if another observer can tell which area was counted. A photograph cropped from the rim and one cropped from a broad flat part of the guard do not necessarily describe the same density. Part of the rim will be viewed obliquely. Its apparent spacing will shrink in one direction, just as windows near the far end of a photographed building seem to stand closer together. Count directly on the uncorrected image and the rim can appear much more densely marked than it actually is.

The physical dimensions of the object recorded by a museum can supply a rough visual scale when the whole guard is shown facing the camera. They cannot automatically calibrate a second image made with a different lens and a different crop. A defensible count would identify a defined patch of ground in a calibrated view, indicate how a mark intersecting the boundary was treated, and give some account of the variation across the piece. A single maximum from a particularly fine patch makes a poor description of an entire workshop's production.

A close view can still yield useful observations in the absence of a numerical scale. Follow a row around the perimeter of a decorative relief. Does it preserve the same direction, change course, or disappear into the edge of the design? Compare the interval between rows with the spacing of marks along a row. These are relative observations that can be tested on another image of the same region. They often tell the viewer more about the organization of the field than a striking but undocumented number of marks per square centimetre.

A fan, a vine and a family of fittings

The Walters Art Museum describes a nineteenth-century kozuka ornamented with an open fan, flowers and vines. Fine fish-roe work occupies the ground. The long narrow shape offers a comparison with the Met dragon that almost anyone can make. In the relatively broad guard, attention passes around a central image and along a rim. On the smaller fitting the eye travels lengthwise. The even texture links intervals of pattern along that longer journey. A term learned from looking at a guard has become useful on a different object without implying that both belonged to the same set or were made by the same hands.

The Royal Collection Trust illustrates a different relationship. Several fittings on one mounted sword bear dragons and share a fish-roe ground. Here the fact that the components are recorded together supplies evidence that two unrelated museum entries cannot offer. A repeated ground helps bind items of different sizes into a coherent ensemble. The designers had to maintain that visual relationship across surfaces a viewer might see in different orientations. A pattern legible on a guard had to retain something of its character on a smaller accessory partly enclosed by the mounting.

These two examples also help explain why catalogues alternate between naming objects by their subjects and naming them by their materials. Search for dragon and a relief may appear whose most instructive feature for a metalworker is actually its ground. Search for fan and vine and one can find an exceptional passage of nanako under a title that says nothing about punches. A search confined to a technique written into the title would lose both. Reading the descriptive paragraph and the medium line alongside the title restores an aspect of the work that iconographic indexing alone cannot capture.

Sword Guard (Tsuba), 29.100.1043
Sword Guard (Tsuba), 29.100.1043. The Metropolitan Museum of Art, Open Access.

A material that could travel off a sword

Another object in the Metropolitan Museum leaves the category of sword fitting altogether. Its mid-nineteenth-century inrō is built from four cases. The catalogue distinguishes the wood-grain mixed-metal body from gold and silver decoration and identifies a separate bead and toggle. Flower medallions give the surface emphases of their own. Beside the small flat field of the laminate guard, the gently curving face of the case presents a different opportunity for looking. A long coloured band can turn away from the viewer at an edge and reappear only when the object is seen from the other side.

The separable cases introduce a further visual interruption. A museum photograph can show the complete object suspended as it would appear assembled, or it can lay the components out for inspection. A pattern that seems continuous in the first view may be distributed across more than one physical piece. To describe an individual case, it is helpful to start from its position in the ensemble. Otherwise an entirely accurate observation about a few millimetres of a surface can be attributed to the bead or toggle merely because the catalogue puts the materials for the whole ensemble in one paragraph.

Mokume also circulated far beyond Edo-period fittings. The Met has examined the technique in the history of international design, including the different ways American metalworkers interpreted Japanese surfaces. Such comparisons expose the difficulty of translating appearance into technical identity. A designer might admire the swirling pattern of a laminate and reproduce a similar visual movement in metal assembled by a different process. An object can be compelling evidence for the reception of a pattern even when it is unsuitable as a specimen of the earlier workshop technique. A good comparative exhibition allows both histories to emerge.

Metal Inrō with Flower Medallions, 91.1.1062
Metal Inrō with Flower Medallions, 91.1.1062. The Metropolitan Museum of Art, Open Access.

What to ask of a colour photograph

An object's colours on a screen are the result of a chain of decisions. Illumination, camera response, white balance, image compression and the screen itself can shift a deep blue-black toward brown or a warm copper tone toward red. Collections that identify shakudō provide material information independent of such shifts. But a catalogue may list several materials for a whole guard without assigning each one to a particular line of colour. The difference matters most when a reader tries to use the image to identify a very small patch under magnification.

The two techniques react differently to the limitations of a digital image. In a laminate, compression can blur a boundary between neighbouring colours until a narrow stratum disappears. In a punched ground, resizing a file can bring rows of dots into a false broad pattern resembling watered silk. The latter effect arises from the interaction of the pattern and the image grid. It need not exist on the metal. A second published view at a different size can help settle the question without handling or altering the object. Comparing the obverse and reverse is especially useful when the institution supplies both.

Wear also deserves a place in the account. On a punched ground, a raised area can catch the light differently wherever its top has been flattened. On a laminate, a small scratch crossing several colours can interrupt the eye's ability to follow a band. Neither observation by itself identifies the event that caused it. A reader who sees a change in a photograph can record its location and compare an older photograph before suggesting a history of the damage. Photographic reproduction, collection handling and earlier use occupy different periods of an object's biography.

Four catalogue entries to keep separate

The laminate guard and the guard inscribed by Denbei share a curious numerical coincidence. The Met records the same weight for both, a little over a hundred and thirty grams, though their dimensions and material entries differ. If a spreadsheet preserved only the category and weight, these might appear to be duplicate records. Their separate accession numbers, dates and descriptions put the two works back in view. The dragon between waves is a third guard, distinguishable at once by its account of gold spray and the textured rim. The medallion case supplies a fourth example with a completely different construction and purpose.

For someone assembling a teaching slide, the cleanest comparison is often the least showy. Place the laminate guard beside the dragon guard at approximately the same physical scale. Identify a small area in each where the foreground motif does not distract. Let viewers describe what would remain visible if the lighting were evenly diffused, then what might be revealed by illumination from the side. Finish by reading the respective material lines aloud. The shift from an observation to a documented identification becomes audible. A person encountering the two Japanese words for the first time can now attach each to a visual experience and to a particular museum object.

The cultural connection between these surfaces survives the technical distinction. Japanese metalworkers asked people to recognize a substance or living form in something that belonged to a completely different material world. Lines in an alloy recall timber. Tiny changes in a punched field call to mind roe. On a collection label the metaphors are short, almost casual. In the presence of the actual objects, each requires a remarkable act of translation by both maker and viewer. The fact that the terms survive so readily in languages far from their place of origin is a measure of how well that invitation still works.

Frequently asked questions

What is mokume-gane?

It is patterned laminated metal. Differently coloured layers become visible at the surface in figures often compared with wood grain.

How does nanako-ji differ from mokume-gane?

Nanako-ji is a ground of tiny punched reliefs. Its pattern is visible in the surface topography, while mokume-gane depends on layers of metal.

Does nanako mean that loose metal beads were soldered on?

The classic fish-roe ground on fittings is punched into the surface. The rounded marks can resemble beads when photographed.

Which museum guard shows both waves and a nanako ground?

The Metropolitan Museum records fish-roe texture between waves and clouds and along the rim of its Dragon Between Waves guard.

Who is associated with early Japanese wood-grain metalwork?

The Japan Mokumegane Research Institute places Shoami Denbei of Akita at the beginning of the Edo-period tradition and discusses signed fittings.

Is there a standard number of nanako grains per square centimetre?

Density varies with the object and region sampled. A count needs a calibrated image, a defined area and a stated counting method.

Films and talks

AKT秋田テレビ(公式 ニュース) — 「金銀銅杢目金」の世界 秋田市で作品展 (19/09/11 19:00)

美術展ナビ — 根津美術館「甲冑・刀・刀装具展 光村コレクション・ダイジェスト」学芸員の見どころ解説

Collection records and further reading

Related Articles