cais1.gif (153930 bytes)

CAIS Persian Text.gif (34162 bytes)


The Circle of Ancient Iranian Studies

 Persian Section.PNG (9914 bytes)


About CAIS


Daily News

News Archive


CAIS Seminars

Image Library





Contact Us


Facebook-Button.jpg (107165 bytes)

On The Two Intertwined Dragons from Pendzikent


By: Arcangela Santoro


Among the lesser finds at Pendzikent stands out - owing to its iconographic peculiarity - a clay sculpture found in the ruins of the iwan Sector II, the temple area. ( Pl. I) The object itself, brought to light by excavations in 1952, has been published repeatedly (Beleniskij & Piotrovskij 1959: table. XXV; Belenitsky 1968: fig.105; Belenitzki 1982: fig. 92), although the only text that describes it and conjectures as to its function goes back fifty years (Beleniskij & Piotrovskij 1959: 66-67).

The singularity of this image and the possibility of reading it in a not exclusively ‘decorative' key have led me to propose a new in-depth analysis, a modest contribution which I dedicate to Boris Marshak, a great scholar and close friend.

Object: Clay sculpture
Subject: Two intertwined dragons
Dimensions: h. 27 cm; diameter 18 cm
Provenance: Pendzikent, Iwan - Sector II
Date of Discovery: 1952 excavations
Kept at: Pendzikent Museum (Tajikistan)
State of Preservation: Broken at bottom right and along the base

The back part comprises a semi-cylindrical block of clay, broken at the bottom. The front - modelled to form a more-or-less horseshoe-shaped niche - is entirely bordered by a high-relief moulding framed by two smooth fillets, the space between decorated by a set of heart-shaped leaves, which were moulded separately. Only a few of the latter have been preserved, but their place is clearly indicated by the hollows. The niche rests on a base with a smooth upper and lower fillet, the part between divided into boxes by double vertical lines.

Within the niche are portrayed two rampant fighting dragons, moulded in high relief, their bodies in profile heraldically confronting each other, with their central section knotted together. The figures represent a combination of several animals: the posterior is snakelike, while the front part is an association of various elements: a dog's or wolf's muzzle with jaws half-open, a short beard below the chin, ears and horns, a dorsal crest and short wings, and - lastly - the front legs of a feline. Owing to their position, the two dragons fit the shape of the niche perfectly. The upper and wider part houses their heads and the front portion of their bodies. The centre, where the space narrows, is entirely filled by the intertwining of the first portion of the snakelike part, while the base - where the niche widens and ends with two scrolls - is taken up by the symmetrical and specular portrayal of the terminal portion (a fish-shaped or foliate tail is shown on the left).

Beleniskij, after a brief description of the composition, underlines the fact that the decoration of the base moulding closely recalls another fragmentary clay sculpture coming from the same area, which can be positively identified as a headdress, and cautiously proposes that this should also be identified as part of a tiara (Beleniskij 1959: 67 and table XXXVIII, 3). This hypothesis has been put forward in the captions of all subsequent publications.

Clearly, his reading is still possible, though anything but certain. What is interesting, however, is not so much the object's function (which would even be rather difficult to define), as the theme of the two entwined dragons, a theme unique - as far as I know - among the items found at Pendzikent up to now.

A further two fragments of clay sculpture portraying dragons have also been found in the ruins of the iwan. Of the first (Beleniskij-Piotrovskij 1959: table XXXVI,1) only the animal's muzzle remains, portrayed in profile, with its jaws open. Of the second, the muzzle survives - also in profile with open jaws - and the front part of the body, showing the short wing and feline paw (Beleniskij-Piotrovskij 1959: table XXXVI, 2). The three images - the fragments just mentioned and the sculpture that is the subject of this study - undoubtedly have features in common (see in particular the shape of the muzzle, the wing and the paw in Beleniskij-Piotrovskij 1959: table XXXVI, 2), but nothing can be said about the rear parts of the two fragments. Above all, in the fragments, the two dragons are portrayed alone and - in all probability - in passant posture. Substantially similar, including the posture, is another portrayal of a dragon coming from the palace of Varahsa (Siksin 1963: 183 and fig.105). Of the stucco sculpture, only the head and neck - shown in profile - remain: the muzzle is portrayed in detail, with open jaws and pointed teeth; the sculpture of the dorsal crest is very precise, with open-ring protrusions; the neck is tubular, almost snakelike. Although the subject is similar, or partly similar, the "knotted dragon" composition remains unique.

We must, however, examine more closely the context in which the sculpture was found. As mentioned above, it was found in the ruins of the iwan, Sector II, featuring the well-known and imposing clay frieze portraying aquatic figures (Beleniskij-Piotrovskij 1959: tables XXVII-XXXIII ). Whatever precise cult value may be attributed to the frieze 1, the depiction of the dragons ties in well enough - at least iconographically - with the monstrous and composite figures that crowd the aquatic landscape. It is worth recalling that among the subjects of the frieze are images of purely classical origin, which assuredly reflect the Gandharan experience: one representative figure is a triton, wholly comparable to a set of Gandharan bas-reliefs (Dani 1968-69: figs.19-20). Our dragons, too, may easily be compared with the composite figures that decorate the lateral portions of the top of false gable and the string panels of a stair of Gandharan sculptures. The latter are, however, always isolated dragons, with a body associating the parts of various animals and a posture wholly conforming to the architectural space available: thus, the front portion of the body with its canine muzzle, dorsal crest, short wings and feline paws fits the wider area perfectly, while the rear of the snakelike body fits happily with the narrowing of the terminal part (Dani 1868-69: figs 26-30; Kurita 1990: figs 718-721). In almost all cases, comparison with Gandhara ends with a typological similarity, seeing that - as we have already said - the images are either single or, in the case of friezes, passant images in Indian file (Zwalf 1996: 254 and fig. 345).

There are, however, at least two exceptions.

A frieze now in the British Museum portrays a series of earth dragons (canine muzzle, wings, dorsal crest, feline body) facing each other two by two and standing on their back legs, while their bodies, rendered with a tubular shape, produce an undulating line that generates a superimposed double S (Zwalf 1996: 253-254 and fig. 344).

The second specimen ( Pl. II) is of great significance and deserves careful attention. It is a rectangular panel, of unknown provenance, and belongs to a Japanese private collection (Kurita 1990: fig.716), delimited above by a moulding with vegetal elements and on the sides by two half-columns with Indo-Corinthian capitals. Within this, two high-relief dragons stand out in profile, rampant on their back legs, seizing each other with their front legs in a clearly combative attitude. Typologically they are earth dragons, already known from other Gandharan reliefs: a canine head with a pronounced twisted almost trumpet-like muzzle, feline ears, body and paws, foliate wings. Up to where they join the back legs, a tubular, almost snakelike shape renders the neck and body. At the waist, the two dragons are linked by what would technically be defined as a reef- or flat knot, generated by the hindquarters conceived as a single component and the forequarters and heads, also conceived as a single component. It may be more easily explained by saying that the knot is generated, not by the clasping and entwining of the two whole figures, but by the joining of one part that is all tail with the other part that is all head.

Despite the different typologies of the dragons in the Gandharan relief (earth dragons) and the Pendzikent ones (water dragons), their compositional resemblance stands out: in both cases, the dragons are portrayed in profile, rampant and combatant and - above all - are knotted together at the centre of their bodies.

Although such a singular composition can be read in a purely decorative key, semantic values cannot be excluded.

A decisive factor in view of this second hypothesis is that in Islamic iconography, the "knot or node" has a precise meaning. Hartner, in a study going back 60 years, but still fundamental and unsurpassed, deals with the matter in great detail (Hartner 1938) and should be consulted for a complete analysis of the problem. Here, I shall merely recall that the knot and/or knotted animals (whether dragons or snakes) are a "sign that stands for" the phenomena of lunar and solar eclipses. In ancient mythological tradition, widespread among different and distant peoples and cultures (Sorrentino 1990), the terrifying phenomenon of the disappearance of the sun and moon is attributed to a monstrous being that swallows the two supreme luminaries. 'As to the nature of this great antagonist of light and life, there rules an astounding agreement among the peoples, as most of them suppose it to be a giant snake or dragon, menacing the great luminaries, and devouring them at certain irregular intervals' (Hartner 1938: 131). In India and Iran, the demon that devours the sun and moon is called respectively Rahu and Djawzahr.

There is common consent among scholars that the Iranian tradition in the astronomic and astrological field has been strongly influenced by the Indian ( Renou et Filliozat 1985: 193-194; Pingree 1987: 859 ). We may briefly recall what it says about the phenomenon of eclipses.

Leaving aside the story of the Vedic asura Svarbhanu, identified by many scholars with Rahu - with the exception of Jamison (1991: 142) , who identifies it with Agni -, the first mentions of Rahu as planet and monster that devours the moon are found in the Atharvaveda (Pant 1981: 279). Ketu appears in the same context, but only as a planet and a comet, meteor or falling star. Throughout the pre-Siddhantic period, only Rahu was held "responsible" for eclipses.

The demon is well known not only to the Brahmanic tradition, but also to the Buddhist. On several occasions, the Jatakas refer to the moon gripped between Rahu's jaws, or being freed from Rahu's jaws. This theme is a standard topos, repeatedly utilised as a comparison in different contexts. Thus, in the Titthajataka, of the monk who conquers Arhatship, it says, ' He, like the moon that wins her way at last from Rahu's jaws, has won supreme release' (Cowell 1981: I, 65). Yet again, in the Kunalajataka, we read, 'A man …..yet fall'n 'neath woman sway, no more will shine than moon eclipsed by Rahu's power malign' (Cowell 1981: V, 244). Neither should we forget that in the Buddhacarita Siddhartha's son is called Rahula 'with the face of Rahu's adversary' ( Johnston 1972: II 48, 29). Lastly - and not because it is less important -, I quote the story recounted in the Candrasutra in which the Buddha converts Rahu, thus liberating the moon (E.Waldschmidt 1970).

It is in the Puranas that we first find the whole story of Rahu and the reasons for his enmity with Sun and Moon. The story forms part of the cosmogonal myth of the churning of the Ocean. 'It was he (i.e. Rahu) who, previous to the churning of the milk ocean, commanded the demons, then allied with the celestial gods in the struggle against the world serpent, Ananta. After the victorious event, he succeeded in an unguarded moment in sipping the amrita drink; but the sun and the moon, who had watched his crime, denounced him to the gods, and instantly Vishnu, approaching in haste, severed his head from his body. Nevertheless, the amrita had already produced its effects and rendered him immortal like the celestials. Consequently, Rahu's head as well as his body, Ketu, intransigent enemies of the two great luminaries, ever since try to devour the sun and the moon whenever occasion serves, and thus cause solar and lunar eclipses (Hartner 1938: 131).

As far as Iran is concerned, the most ancient written evidence is found in the Bundahisn, a text compiled in the ninth century CE, but in which 'traditional, orthodox beliefs derived from the Zoroastrian scriptures appear side by side with later and even contemporary scientific opinions' (MacKenzie 1964: 511). Here, the two entities that oppose the sun and moon appear 'as Dark (i.e. presumably 'eclipsed') Sun and Dark Moon. These hostile entities which intercept the light of the luminaries are the head and tail of the Dragon Gocihr… At the beginning of celestial motion the head (the ascending node of the moon's orbit upon the ecliptic, Sanskrit Rahu) was in Gemini, the tail (Sanskrit Ketu) in Sagittarius.' ( Brunner 1987: 867).

The mythological reading of the phenomenon survives even when the scientific causes are clear. To quote Hartner again, 'We might suppose that clear insight into the physical causes of eclipses would have thrown the mythological tradition into the background. But this has not been the case. What we observe is that the mythological and astronomical elements contract an intimate fusion. The nodes of the moon's orbit are simply identified with the eclipse monster itself: with the Hindus, Rahu becomes the ascending, Ketu the descending node; with Persians and Arabs, the head and the tail of the Djawzahr play the same role.' (Hartner 1938:131).

If we pass from the written tradition to iconography we find it is the Indian world that offers the most ancient portrayals of Rahu and Ketu (Sivaramurti 1958). The two demons appear starting from the sixth century CE in the navagraha group, together with the five planets, the sun and moon, and - like the other seven - are shown in anthropomorphic form. More precisely, Rahu is shown in anthropomorphic form whereas, in some images, Ketu is shown with a human torso and a snakelike lower body, often knotted.2 The anthropomorphic iconography also persisted in south-eastern Asia, where, however, the lower part of Rahu's body is enveloped and hidden by clouds (Bhattacharya 1956; Idem 1964), and in the Buddhist Silk Road art at the Khara Khoto site (Kocetova 1947: 492-495).3

In the Iranian area, the first portrayals 4 date back to the twelfth century CE ( Taddei 1958). Islamic iconography, instead of a human form for the demon of the eclipse, portrays serpents or dragons with a node, or pairs of dragons with entwined bodies. They may be autonomous images and, as it were, complete in themselves, or form the terminal part of the signs of the zodiac (e.g. Sagittarius), or even appear in the abstract form of a heart-shaped knot (Hartner 1938: 138).

In conclusion, unlike the substantially concordant and ancient textual tradition, the corresponding iconographic tradition belongs to different periods (from the sixth century CE in the Indian area to the twelfth century CE for the Iranian area) and is figuratively discordant, the Indian being anthropomorphic, and the Islamic zoomorphic or abstract. And to the latter, the node theme is fundamental.

We shall now return to the entwined dragons of Pendzikent, from which we started.

Azarpay has already noted that 'The theme of the entwined dragons in Islamic art finds Central Asian forerunners that doubtless contributed to the widespread use of the motif in Islamic art patronized by Turkish dynasties', citing - among the most ancient examples - the very sculpture we have been discussing (Azarpay 1978: 306, note 20). Thanks to the Gandharan relief analysed above, the first appearance of the theme can be backdated to the third-fourth centuries CE. A third image, coming from the Iranian area and datable to the late Parthian period, contributes to our documentation - in actual fact somewhat scarce - for the more ancient period. I refer to a stucco capital found at the Qal'eh-i Yazdigird ( Pl. III) - where two rampant, divergent and knotted earth dragons appear. 'The intertwined dragon motif on an engaged column capital clearly owes much to the ancient traditions of Near Eastern Art, which abounds with scenes of crossed and rampant beasts. Curiously, the closest examples to the Qal'eh-i Yazdigird intertwined dragons must be the pair of 'fighting dragons' from 7th/8th century Penzikent. In view of the suggested significance of the twisted knot or 'node' in Islamic iconography an astrological association in the case of the Qal'eh-i Yazdigird piece should not be discounted' (Keall 1977: 7 and pl. IIIb).

I consider that this latter statement of Keall's can also be extended to the Gandharan panel and the clay sculpture we are discussing.

In other words, while it is certain that, from an iconographic point of view, the Qal'eh-i Yazdigird image, the Gandharan one and the one from Pendzikent must be considered as forerunners of the astronomic-astrological iconography of the Islamic period, we cannot - and perhaps should not - exclude that this is also valid from a semantic point of view.

One objection to this hypothesis may be found however in the fact that in Islamic iconography, knotted figures - even the various types summarized above - are indubitably an eclipse sign when they form part of an astrological-astronomic context, i.e. when they accompany signs of the zodiac or of stars such as the moon. Such a context is missing in the case of the three examples we are discussing. At the same time, Islamic iconography is not lacking in images of knotted dragons or snakes, whether isolated or found in different contexts, for which scholars have put forward various interpretations.5 One recurrent and more convincing reading attributes a protective and apotropaic value, perhaps in some way connected with its primary meaning. In this connexion, it is interesting to recall what Ittig says about two entwined dragons ( Pl. IV) that appear on a talismanic bowl (twelfth-fourteenth century CE?) in the Toronto museum. 'According to the oral traditions of the Arabs, the eclipse was a harbinger of great disaster. In keeping with this is the use of entwined dragons in Islamic gate emblems as prophylaxis against the entry of evil spirits. The symbol of the eclipse may therefore have been placed in the Toronto bowl to frighten away the evil ginn which had entered the body of the afflicted party' (Ittig 1982: 91 and pl. VII,a).

In conclusion, my opinion is that the three portrayals of entwined dragons we have examined do not have a merely decorative value. The future history of the theme would lead me to propose a reading connected with an astronomic-astrological and/or prophylactic context. Such a hypothesis, though higly probably and reasonable, is however far from certain. If I may, I should like to note that several years ago, when I showed the Gandharan relief with the dragons to Maurizio Taddei, his first comment was "Rahu and Ketu". If accepted, such an identification would explain the singularity of the image which, as described above, constructs the node, not through the entwining of the two dragons' bodies, but through the combination of one being that is "all head" (i.e. the front part of the body) with another being that is "all tail" (i.e. the back part of the body). Far from being the result of the sculptor's lack of skill, such a solution could be an effective visual expression of the myth of Rahu and Ketu, a myth that survives in the formula 'caput et cauda draconis'.

Bibliographical References

Azarpay, G. ( 1978) The Eclipse Dragon on an Arabic Frontispiece-Miniature. Journal of American Oriental Society, 98, 363-374

Belenitski, A.M. (1958), Nouvelles découvertes de sculptures et de peintures murales a Piandjikent. Arts Asiatiques V, 163-182

Beleniskij, A.M. ( 1973), Monumental'noe Iskusstvo Penzikenta, Moskva

Belenizki, A. M. ( 1980) Mittelasien Kunst der Sogden. Leipzig

Belenitski A.M. et Marshak, B. ( 1971) L'art de Piandjikent à la lumière des dernières fouilles ( 1958-1968) Arts Asiatiques 23, 3-27

Beleniskij, A.M. & Piotrovskij ( 1959), Skul'ptura i zivopis' drevnego Pjandzikenta, Moskva

Bhattacharya, K. ( 1956), Notes d'iconographie Khmère, Arts Asiatiques, III, 183-194

Bhattacharya, K. ( 1964), Notes d'iconographie Khmère, Arts Asiatiques, X, 91-94

Brunner, C.J. (1987), Astronomy and Astrology in the Sasanian Period, s.v. Astrology and Astronomy in Iran, Encyclopaedia Iranica, II, pp.862-868

Cowell, E.B. (ed.) The Jataka or Stories of the Buddha's Former Births, 6 vols., London ( Irst ed. 1895)

Dani, A.H. ( 1968-69) Chakdara Fort and Gandhara Art. Ancient Pakistan, IV ( 1968.69), Special Number

Hartner, W. (1938) The Pseudoplanetary Nodes of the Moon's Orbit in Hindu and Islamic Iconography. Ars Islamica, V, 112-154

Ittig, A. (1982), A Talismanic Bowl, Annales Islamologiques, XVIII, 79-94

Iannaccone, I. e Tamburello, A. ( a cura di) (1990) Dall'Europa alla Cina : contributi per una storia dell'Astronomia. Napoli

Jamison, S. W. (1991) The Ravenous Hyenas and the Wounded Sun. Myth and Ritual in Ancient India. Ithaca and London

Johnston, E.H. ( 1972),Asvaghosa's Buddhacarita or Acts of the Buddha, Delhi-Varanasi-Patna ( Irst ed. 1936)

Keall, E.J (1977) Qal'eh-i Yazdigird: the Question of its Date Iran XV, 1-9.

Kocetova, S. M. ( 1947) Bozestva svetil v zivopisi Hara-Hoto , Trudy Otdela Vostoka Travaux du Département Oriental du Musèe de l'Ermitage, IV, 471-502

Kurita, I. ( 1990), Gandharan Art II The World of the Buddha, Tokyo

MacKenzie D.N. (1964), Zoroastrian Astrology in the Bundahisn, Bulletin of the School of Oriental and African Studies, XXVII, 511-529

de Mallmann, M.Th. (1963), Les enseignements iconographiques de l' Agnipurana, Paris

Oney G. (1969), Dragon Figures in Anatolian Seljuk Art, Belleten, XXXIII, 193-216

Pant, R. ( 1981), History of Rahu and Ketu, Vishveshvaranand Indological Journal Prof. Jagannath Agrawal Felicitation Volume, XIX, 277-283

Pingree, D. (1987), History of Astronomy in Iran, and Astrology in Islamic Times, s.v. Astrology and Astronomy in Iran, Encyclopaedia Iranica, II, 858-862 and 868-71

Renou, L. et Filliozat, J. ( 1985), L'Inde Classique. Manuel des Etudes Indiennes, 2 vols., Paris ( Irst ed. 1947/49)

Siksin, V.A.( 1963), Varahsa, Moskva

Sivaramamurti, C. (1958) ) Astronomia e Astrologia India, s.v., Enciclopedia Universale dell'Arte, II, Roma-Venezia , 134-138

Sorrentino, A. (1990), "Mangiare il sole": il mito astrale dell'eclisse in India e in Cina in Iannaccone e Tamburello :1990, 67-74

Taddei, M. (1958) Astronomia e Astrologia Islam, s.v., Enciclopedia Universale dell'Arte, II, Roma-Venezia ,130-134

Verardi, G. (1982), Osservazioni sulle sculture in argilla e su alcuni ambienti dei complessi templari I e II di Pendzikent, Annali dell'Istituto Orientale di Napoli, 42, 247-304

Waldschmidt, E. (1970) Buddha frees the disc of the moon ( Candrasutra), Bulletin of the School of Oriental and African Studies, XXXIII, 179-184

Zwalf, W. ( 1996) A Catalogue of the Gandhara Sculpture in the British Museum, 2 vols., London 1996



List of Illustrations


Pl. I Two intertwined dragons from Pendzikent, Pendzikent Museum

Pl. II Two intertwined dragons from Gandhara ( after Kurita 1990: fig.716)

Pl. III Two intertwined dragons from Qal'eh- i Yazdigird ( after Keall 1977: pl. III, b)

Pl. IV A talismanic bowl, Toronto Museum (after Ittig 1982: pl.VII, a)


1 Scholars do not agree as to the real significance and source of the subject portrayed. Although Russian archaeologists incline towards a cult reading- seeing it as a representation of the River Zeravshan- its composition based on Indian models and more specifically from Grotto 5 at Udayagiri ( Beleniskij-Piotrovskij 1959: 76-77; Belenitski 1958: 164-166), Verardi links the scene to Buddhist production, and more precisely to the aquatic scene in a chapel at Tapa Sotor (Verardi 1982: 252-257)

2 References to the snake tail of Ketu in Agnipurana ( de Mallmann 1963: 86)

3 I should like to thank Dr. Semenov, Head of Central Asiatic Department, Hermitage Museum, who provided me with a copy of this article.

4 I entirely leave apart the more ancient history of the entwined animals theme, which does not directly concern the present case, and can be found in Azarpay 1978: 369-371

5 As an extreme example of different readings, I cite the list of Oney 1969: 213-216, according to whom the dragons in Seljuk art may have the following symbolic meanings: 1. Symbolism of motion, harmony or universe; 2. Symbolism of fight against darkness and evil; 3.Symbolism of underground and darkness; 4. Symbols of Cauezahr, i.e. planets; 5. The dragon year in the Turkish-Chinese calendar with twelve animals; 6. Water symbols.



Top of Page



Please note this article is originated from: Webfestschrift Marshak Ērān ud Anērān - Electronic Version of studies presented to Boris Ilich Marshak on the Occasion of His 70th Birthday.

Editors: Matteo Compareti, Paola Raffetta, Gianroberto Scarcia



Please note: CAIS has the privilege to publish the above article originating from the above-mentioned source, for educational purposes only (Read Only). This article has been published in accordance with the author(s) / source' copyright-policy -- therefore, the ownership and copyright of this page-file remain with the author(s) / sourceFor any other purposes, you must obtain a  written permission from the copyright owner concerned. (Please refer to CAIS Copyright Policy).




my_Iran.jpg (13682 bytes)

"History is the Light on the Path to Future"


Persian_NOT_Farsi_by_Shapour_Suren-Pahlav_3D2.gif (177309 bytes)


Encyclopaedia Iranica

BIPS.jpg (15695 bytes)

The British Institute of Persian Studies

"Persepolis Reconstructed"

Persepolis_reconstructed2.jpg (36944 bytes)


The British Museum

The Royal

Asiatic Society

Persian_Gulf_Facebook.jpg (1935028 bytes)

The Persian Gulf

Facebook Page

Please use your "Back" button (top left) to return to the previous page

Copyright © 1998-2015 The Circle of Ancient Iranian Studies (CAIS)