Aka Fudō, Ao Fudō, Ki Fudō: reading the colour-type recensions

Three named Fudō Myōō paintings are distinguished not by attribute but by body colour and the lineage of the cult image: the Aka (Red) Fudō at Myōō-in on Kōyasan, the Ao (Blue) Fudō of Shōren-in, and the Ki (Yellow) Fudō of Onjō-ji. The colour is diagnostic of transmission, not decoration. Because all three are temple-bound hibutsu with no open image, they are read here through scholarship against a clean public-domain standard Fudō bronze used strictly as the baseline.

What this resolves, and the object in front of you

There is a recurring confusion in catalogue shorthand and gallery labels: “the Red Fudō,” “the Blue Fudō,” as if these were colour-coded sub-types of one figure the way Bato and Jūichimen are sub-forms of Kannon. They are not. Aka Fudō, Ao Fudō, and Ki Fudō are not iconographic variants in the sense of differing attributes. Each is a single, named, temple-bound painting, a particular venerated icon whose authority derives from the lineage of its transmission. The colour names the cult image, not a new programme. All three hold the same sword and noose the standard Fudō holds.

The honest difficulty in writing this is that none of the three can be looked at. The Onjō-ji Ki Fudō is a strict hibutsu, normally withheld even from temple visitors; the Shōren-in Ao Fudō is shown only at a handful of special openings each year; the Kōyasan Aka Fudō stays at Myōō-in. There is no open-access photograph of any of them. They are read here through scholarship and conservation reports, against a baseline that can be examined.

That baseline is the Cleveland Museum’s accession 1989.267, a cast-bronze Fudō Myōō, twelve and a half centimetres tall without its base, Japan, Kamakura period, early fourteenth century.¹ It does not hang on a wall and was never meant to. At its true size it is a desk object, a thing held at reading distance in the hand, not pedestalled at the far end of a hall the way the three painted icons are enshrined above an altar and behind a curtain. The metal is uncoloured, dark, granular where the casting did not fully take along the left flank; the figure sits with the sword held upright in the right hand and the noose looped in the left, the brow knotting into the standard scowl. It is, deliberately, an unassuming one: a standard-form Fudō, reproduced by the thousand across Japanese temple and devotional production. It is not one of the three named icons and is not offered as a stand-in for any of them. It is the control, the form held constant so that what the colour recensions add, lineage rather than attribute, can be isolated.² Read the bronze first; then read what the named paintings do not change.

The standard Fudō: the form held constant

The standard polychrome Fudō Myōō is one of the most stable iconographies in the Japanese Buddhist repertoire, and that stability is the whole point of using it as a control. The figure is wrathful and seated (or, in the temple-honzon tradition, standing) on a rock; the right hand grips the kongō-ken, an upright vajra-hilted sword; the left holds the kenjaku, a noose. The face carries the late-Heian codification: the tenchigan or “heaven-and-earth eyes,” one eye wide and the other narrowed, with the paired fang asymmetry (one fang up, one down) read in the medieval commentaries as the figure’s simultaneous duality and non-duality.² A single braid, the henpatsu, falls over the left shoulder. Two boy attendants stand below: Kongara, obedient and palms joined, and Seitaka, unruly, the tamed force.³

The Cleveland bronze carries this programme in miniature and in uncoloured metal, which is exactly what makes it legible as a baseline. With no pigment, the surface cannot make a colour argument; the eye goes to the posture, the sword, the noose, the scowl. The standard Met painted comparanda show the same programme in the more usual format: two fourteenth-century hanging scrolls, “Fudō Myōō and Two Attendants,” accessions 1975.268.27 and 1975.268.28, each with the seated wrathful figure flanked by the two dōji.⁴ Hold the bronze and the two Met scrolls together and the standard type resolves into something the colour recensions will not alter: the sword, the noose, the rock, the two boys. What the named paintings change is the body colour and the cult identity attached to it, and nothing else of the programme.

The Aka Fudō: Red, Myōō-in on Kōyasan

The Aka Fudō (“Red Fudō”) is the painting held at Myōō-in on Mount Kōya, the Shingon mountain headquarters in Wakayama. The body is rendered in red, and the figure carries the two standard attendants, Kongara and Seitaka.⁵ The painting is conventionally attributed to the Tendai monk Enchin (Chishō Daishi, 814–891) and is the subject of the most colourful etiological legend of the three: Enchin, seized by a vision of Acala, is said to have painted the image in his own blood, the legend that supplies the cult explanation of the red.³

Here is the first museum-versus-scholarship divergence worth stating plainly. The Enchin attribution is the traditional, devotional account. Modern art-historical analysis dates the surviving Myōō-in painting considerably later, to the Kamakura period, on stylistic grounds, which means the object venerated as Enchin’s blood-painting is, on the scholarly reading, a later cult image carrying an early attribution.³ This is not a marginal quibble: it is the precise pattern that recurs across all three recensions. The colour and the lineage-claim are old; the surviving silk is often younger. bodhi’s reading follows the art-historical dating where it is reported, the Aka Fudō as a Kamakura cult image under a Heian attribution, while recording that the temple’s own account is the Enchin-blood etiology, and that the residual uncertainty is the absence, in open scholarship, of a pinned technical date for the Myōō-in silk itself.

The Ao Fudō: Blue, Shōren-in

The Ao Fudō (“Blue Fudō”) is the National Treasure of Shōren-in, the Tendai monzeki temple in Kyoto. It is the best-documented of the three. The painting is a large silk hanging scroll, roughly 203 by 149 centimetres, dated to the mid-Heian period, around the eleventh century, and is regarded as one of the supreme works of Japanese Buddhist painting.⁶ Since 2014 it has been enshrined in the Seiryū-den, a purpose-built goma-dō hall at Shogun-zuka on the Higashiyama ridge, and is shown to the public only during a few special openings each year.⁶

The body is blue. The attendants are again Kongara and Seitaka, and the iconographic literature gives them their canonical reading here: Kiṃkara (Kongara) white and palms joined; Ceṭaka (Seitaka) red and holding a vajra, the obedient and the unruly, the tamed elemental forces.³ What distinguishes the Ao Fudō from the standard type is not the attendant pair, which is standard, nor any change to sword or noose, which are standard. It is the blue body, the singular eleventh-century silk, the National Treasure cult status, and the Tendai-monzeki transmission. The Ao Fudō is the cleanest case of the article’s thesis: drop the blue and the painting is iconographically a standard two-attendant Fudō; the blue, the date, and the lineage are what make it the Ao Fudō rather than an Fudō.

The Ki Fudō: Yellow, Onjō-ji (Mii-dera)

The Ki Fudō (“Yellow Fudō”) is enshrined at Onjō-ji, Mii-dera, the Tendai Jimon temple below Mount Hiei in Shiga. It is a National Treasure and the strictest hibutsu of the three: a hidden image, normally not displayed even to temple visitors, traditionally shown only to esoteric masters in the context of initiation.⁷ The body is yellow. The temple account ties it to Enchin’s vision during austerities in 838, the “golden man” who appeared before him, which the tradition reads as the etiology of the yellow.⁵

This is where the museum-versus-scholarship divergence is sharpest and best-evidenced, and it is worth giving the named source its due weight. Yanagisawa Taka’s 2005 conservation report in Bijutsu Kenkyū, published after the 1996–1998 conservation campaign on the painting, argues that the Onjō-ji Ki Fudō was not painted in Japan at all. Yanagisawa identifies it with a painted Acalanātha image Enchin recorded receiving from the ācārya Zhihuilun at Da Xing-shan-si in 855, and reads the technical evidence (pigments, ground, brushwork “absolutely not found on any other examples of Buddhist painting created in Japan”) as consistent with central-Chinese workshop practice, placing production after Enchin’s return in 858 rather than at the 838 vision.⁸ On this reading the Ki Fudō is not the earliest Japanese Fudō painting but an imported Tang one — the seed-image from which the Japanese tradition then grew.

bodhi’s reading follows Yanagisawa: the conservation evidence is technical, recent, and specific, and it outweighs the traditional 838-vision dating for the question of where and when the surviving silk was made. The residual uncertainty is honest and twofold: the Ki Fudō’s hibutsu status means the painting is examined far less often than its fame implies, and the Yanagisawa pages are cited here at article level from the conservation summary, not pinned to the physical volume of Bijutsu Kenkyū 385, so an operator pass against the print journal is flagged in the sidecar.

Telling them apart: the comparison table

Recension	Body colour	Temple / collection	Period of the surviving image	Lineage of the cult image	Attendant configuration	Open-access image?
Aka Fudō (Red)	Red	Myōō-in, Kōyasan (Shingon)	Kamakura (scholarly); attributed to Enchin, Heian (traditional)	Enchin “blood-painting” legend; venerated locally since the Heian period	Kongara + Seitaka (two dōji)	No (temple-bound)
Ao Fudō (Blue)	Blue	Shōren-in / Seiryū-den, Kyoto (Tendai monzeki)	Mid-Heian, 11th century	Tendai-monzeki cult image; National Treasure	Kongara (white, palms joined) + Seitaka (red, vajra)	No (special openings only)
Ki Fudō (Yellow)	Yellow	Onjō-ji / Mii-dera, Shiga (Tendai Jimon)	Tang Chinese import, post-858 (Yanagisawa 2005); Enchin 838 vision (traditional)	Enchin / Zhihuilun transmission; strict hibutsu	Standard two-attendant (account-dependent)	No (strict hibutsu)
Standard Fudō	Polychrome or uncoloured (e.g. bronze)	Ubiquitous; here Cleveland 1989.267	Any period; anchor object Kamakura, early 1300s	None: a type, not a singular icon	Kongara + Seitaka (canonical)	Yes: Cleveland CC0 anchor; Met 1975.268.27 / .28 painted comparanda

The fastest disambiguation is therefore not “look at the attributes,” since the attributes are constant. It is: read the body colour first, then ask which named temple image that colour belongs to, then verify the lineage and the secret-image status. The standard Fudō has no such lineage; that absence is exactly what makes the bronze a clean control.¹

The common misreading, and what the anchor cannot do

The persistent error is to treat body colour in these three as a decorative or stylistic choice, to file “Red Fudō” alongside, say, a polychrome versus a monochrome rendering, as if the painter chose red the way a painter chooses a palette. In the colour-type recensions the colour is lineage-diagnostic: it indexes a specific transmitted cult image with its own etiology (Enchin’s blood for the red, the golden vision for the yellow, the eleventh-century Tendai silk for the blue). A second, related conflation is to assume the named icons must differ iconographically from the standard type, to hunt for a different mudrā or a different implement. They do not. The discriminating facts are colour, the singular object, the transmission lineage, and the secret-image status, none of which is an attribute on the body.

The honest limit of this article is its own anchor. The Cleveland bronze does the one thing it can do, holding the standard programme constant in a form that can actually be examined and reproduced, and it cannot do the things the named paintings would: it carries no colour argument, no transmission lineage, no hibutsu aura, and it is sculpture set against three paintings. It is a control, declared as a control, not a substitute. The three colour Fudō remain, by their nature, read rather than seen; that condition is part of what the recension is. What can be set down here with confidence is the structure of the distinction and the named scholarship that fixes the Ki Fudō’s origin and unsettles the Aka Fudō’s date. What cannot, without the temples opening their images and the print volume of Bijutsu Kenkyū 385 in hand, is the last layer of technical pinning; that gap is named rather than papered over.

Related

• Fudō Myō-ō: the sword, the lasso, the asymmetric face, and the standard programme in full
• Met 44842: a close reading of the standing Heian Fudō Myōō
• The Five Wisdom Kings: disambiguating Fudō within the Godai Myōō set
• Fudō Myōō (entity)
• Tendai (entity)
• Shingon (entity)

Footnotes

1. The Cleveland Museum of Art, “Fudō Myōō (不動明王像),” accession 1989.267, object 154671: Japan, Kamakura period (1185–1333), early 1300s; bronze with wire attachments; with base 16 cm, without base 12.7 cm; Bequest of Mrs. Severance A. Millikin. share_license_status CC0 re-confirmed via the Cleveland Open Access API, 2026-05-16. Used in this article strictly as the standard-form comparative anchor; it is not one of the three named colour-type icons and is not presented as a proxy for any of them. ↩ ↩²

2. Bernard Faure’s treatment of Fudō Myōō in the medieval Japanese esoteric tradition supplies the reading of the wrathful programme used as the constant here: the tenchigan eye-and-fang asymmetry as duality/non-duality, and the variation in attendant count (two, eight, thirty-six, occasionally forty-eight) as configuration rather than identity change. Cited at chapter level; specific page references are not pinned and are flagged for an operator pass in the sidecar. ↩ ↩²

3. Acala/Fudō Myōō iconographic-subject reference (Wikidata Q337624), cross-checked against Faure and the temple records: the standard two-attendant pair (Kongara, palms joined; Seitaka, the unruly tamed force), the tenchigan/opposed-fang asymmetry codified from the late ninth century, and the Aka Fudō blood-painting legend with the modern Kamakura redating of the surviving Myōō-in silk. ↩ ↩² ↩³ ↩⁴

4. The Metropolitan Museum of Art, “Fudō Myōō and Two Attendants,” accessions 1975.268.27 (object 45203) and 1975.268.28 (object 45204): Japan, 14th century, hanging scrolls; public domain (CC0), re-confirmed via the Met Collection API, 2026-05-16. Cited as standard two-attendant painted comparanda; neither is a hero on bodhi. ↩

5. Himalayan Art Resources, “Japan: Achala (Three Famous Fudo-myoo Paintings),” set 6820 (accessed 2026-05-16): collates the Aka Fudō (Myōō-in, Kōyasan; red; Kongara and Seitaka; Enchin attribution; venerated locally since the Heian period) and the Ki Fudō (Mii-dera; yellow; Enchin retreat-vision account; hibutsu) with temple, period, and attendant data used in the comparison table. ↩ ↩²

6. Shōren-in temple record, “National Treasure — Ao Fudō Myōō with Two Attendants” (Seiryū-den, Shogun-zuka; accessed 2026-05-16): silk hanging scroll, ca. 203 × 149 cm, mid-Heian (11th century), National Treasure; enshrined since 2014 in the purpose-built Seiryū-den goma-dō and shown only during a small number of special openings each year. ↩ ↩²

7. Onjō-ji (Mii-dera) temple record (accessed 2026-05-16): the Ki Fudō is a National Treasure enshrined as a strict hibutsu, normally withheld from public view and traditionally revealed only to esoteric masters in the initiation context; associated with Enchin (Chishō Daishi). ↩

8. Yanagisawa Taka (柳澤孝), “Onjōji shozō kokuhō Kifudō-myōō-zō ni kansuru shin-chiken — shūri hōkoku” (“New Information Regarding the Yellow Acalanātha [Ki-Fudō] Painting of Onjōji: Conservation Report”), Bijutsu Kenkyū (美術研究) no. 385 (2005), pp. 1–16. Reports the 1996–1998 conservation campaign and argues, on technical-material grounds, that the Ki Fudō was produced in China and brought to Japan by Enchin after his 858 return — identifying it with the Acalanātha image Enchin recorded receiving from Zhihuilun in 855, against the traditional 838-vision attribution. Cited from the conservation-report summary at article level; the specific pages within no. 385 are not pinned to the physical volume; operator pass needed (flagged in the sidecar watch list). ↩