41
RIVISTA DEGLI STUDI ORIENTALI VOLUME LXXVI FASC. 1-4 (2002) PISA ROMA ISTITUTI EDITORIALI E POLIGRAFICI INTERNAZIONALI 2003

Umamahesvara in Central Asian Art

Embed Size (px)

DESCRIPTION

The spread of the Umamahesvaramurti in the art of Tokharestan and Sogdiana in the VI-VIII centuries.

Citation preview

RIVISTADEGLI

STUDI ORIENTALI

VOLUME LXXVI

FASC. 1-4

(2002)

PISA ⋅ ROMA

ISTITUTI EDITORIALI E POLIGRAFICI INTERNAZIONALI

2003

RIVISTA DEGLI STUDI ORIENTALI

Trimestrale

Prezzo d’abbonamento per l’anno 2002:Interno 150,00 (privati); 200,00 (enti, con edizione Online)

Estero 350,00 (Individuals); 500,00 (Institutions, with Online Edition)(comprensivo dell’intera annata della rivista e dei due supplementi)

I versamenti possono essere eseguiti sul conto corrente postale n. 13137567o tramite carta di credito (Visa, Eurocard, Mastercard, American Express, Carta Si)

ISTITUTI EDITORIALI E POLIGRAFICI INTERNAZIONALI

PISA ⋅ ROMA

Casella postale n. 1 ⋅ Succursale 8 ⋅ I 56123 Pisa

Uffici di Pisa: Via Giosuè Carducci 60 ⋅ I 56010 Ghezzano La Fontina (Pisa)Tel. +39 050878066 (r.a.) ⋅ Fax +39 050878732

E-mail: [email protected]

Uffici di Roma: Via Ruggero Bonghi 11/b (Colle Oppio) ⋅ I 00184 RomaTel. +39 0670452494 (r.a.) ⋅ +39 0670476605

E-mail: [email protected]

La Casa editrice garantisce la massima riservatezza dei dati forniti dagli abbonati ela possibilità di richiederne la rettifica o la cancellazione previa comunicazione allamedesima. Le informazioni custodite dalla Casa editrice verranno utilizzate al solo

scopo di inviare agli abbonati nuove proposte (L. 675/96).

Sono rigorosamente vietati la riproduzione, la traduzione, l’adattamento ancheparziale o per estratti, per qualsiasi uso e con qualsiasi mezzo eseguiti,

compresi la copia fotostatica, il microfilm, la memorizzazione elettronica, ecc.,senza la preventiva autorizzazione scritta degli

Istituti Editoriali e Poligrafici Internazionali, Pisa ⋅ Roma

http://www.iepi.it

Copyright 2003 byUniversità degli Studi di Roma ‘‘La Sapienza’’

ISSN 0392-4866

INDICE DEL VOLUME SETTANTASEIESIMO

1-4

ARTICOLI......................................................................................................................................... Pagina

Y. LEV, Charity and Justice in Medieval Islam ............................................................................ 1Rapporto delle campagne di scavo 2002 in Uzbekistan:C. SILVI ANTONINI, F. FILIPPONI, A - Il sito di Uch Kulakh....................................................... 17F. NOCI, B - Kurgan Vardanzeh. Ricognizione archeologica preliminare.................................. 27M. COMPARETTI, C - Note sul toponimo Vardana-Vardanzı........................................................ 39C. LO MUZIO, The Umamahesvara in Central Asia Art .............................................................. 49B. LO TURCO, Il terzo prakaran

˙a dello Yogavasis

˙t˙ha (utpatti) e la dottrina sivaita della vibrazione

(spanda) ........................................................................................................................................ 87T. LORENZETTI, A Rare Meditational Sculpture from Tamil Nadu............................................. 121M.L. DI MATTIA, Il complesso templare di Nako nell’alto Kinnaur: un esempio dello stile indo-ti-

betano dei secoli X-XII. Parte II ............................................................................................ 137P. CORRADINI, On the Qidan and Jurcin Capitals ........................................................................ 169

NOTE E DISCUSSIONI

C.M. LUCARINI, D. CAMPANILE, A proposito di una nuova traduzione e commento a Filostrato,Vite dei sofisti ............................................................................................................................. 215

L. ROSTAGNO, Note su Domenico Gerosolimitano. Parte I......................................................... 231M. BERNARDINI, A Response to Dr. Orsatti ................................................................................... 263

RECENSIONI

The Phoenicians in Spain (G. Garbini) ......................................................................................... 265A. FINCKE, The Samuel Scroll from Qumran (A. Catastini) ....................................................... 267Santa Nino e la Georgia (E. Ercolino) ........................................................................................... 275D. THOMAS, Early Muslim Polemics against Christianity (B. Scarcia Amoretti) ........................ 278Saints et héros du Moyen-Orient contemporain (B. Scarcia Amoretti) ...................................... 279D. DAVIS, Panthea’s Children (A. M. Piemontese) ....................................................................... 282B. MARSHAK, Legends, Tales and Fables in the Art of Sogdiana (A. Santoro) ........................ 284L. ALSDORF, Kleine Schriften, 2. Auflage (P. Daffinà) ................................................................. 288E. WINDISCH, Kleine Schriften (P. Daffinà) ................................................................................... 290J. BROCKINGTON, The Sanskrit Epics (F. Squarcini) ...................................................................... 293A. FORTE, F. MASINI, A Life Journey to the East (M. Miranda) ................................................ 296L. BUTLER, Emperor and Aristocracy in Japan (V. Ferretti) ....................................................... 300

SCHEDE BIBLIOGRAFICHE

L. RAMIREZ BELLERÍN, Historia secreta de los Mongoles (P. Cannata) ....................................... 305M.S. GORDON, The Breaking of Thousand Swords (P. Cannata) ............................................... 307

LIBRI RICEVUTI ............................................................................................................................ 311

SUPPLEMENTI

C. BORI, Ibn Taymiyya: una vita esemplare. Analisi della fonti classiche della sua biografiaF. D’AGOSTINO, L. VERDERAME, Umma Messenger Texts

[1] 49

THE UMAMAHESVARA IN CENTRAL ASIAN ART

In the Central Asian Sivaite imagery, the Umamahesvaramurti, i.e. thesculptural or painted representation of Siva (Mahesvara) along with his wife Par-vatı (Uma), holds a place of special interest, in that it is one of the few icono-graphic subjects which lend themselves to an unambiguous identification.

Seven representations of the couple are so far known, six of them in paint-ing (one at Dilberjin, two at Khotan, three at Kyzyl) and one in clay sculpture(Penjikent). But for the Dilberjin mural (fig. 1), which has been attributed tothe late Kushan epoch (see further), all of them belong to a rather circum-scribed time span, that is the 6th-first part of the 8th century AD. They occur indiverse religious contexts. At Kyzyl (figs. 3-5) they appear in large scenes cen-tred on the Buddha; the Khotanese Umamahesvaras (figs. 6, 7) are depicted asisolated icons on wooden panels; as such they are likely to be linked to the Bud-dhistic milieu. At Penjikent (fig. 2) Siva and Parvatı seem to be set within a lo-cal religious background; so do they at Dilberjin (fig. 1). Although each casedisplays iconographic elements which are at variance with the original Indianmodels, a number of common traits make of our Central Asian Umamahesvarasa rather homogeneous group.

Dilberjin (fig. 1). The mural was found in the “Temple of the Dioscuri”(room 2, west wall, to the right of the entrance to the cella)1. It was recoveredin the Kabul Museum, so that it is now virtually lost (the use of the presenttense in the following description is therefore purely conventional). Thepainted layer was only partially preserved. The upper part of the figures is miss-ing, in particular their heads are lost, so that we do not know whether Siva wasone- or three-faced. The couple, which represents the fulcrum of the composi-tion, is seated on the bull (vr

˙s˙a). Mahesvara (on the left) sits in sukhasana-like

pose, with his right foot resting on a foot-stool; he wears a dhotı, which covershis legs up to the ankles, bracelets and keyura. His erected penis is hidden bythe dhotı. He has four arms; of the upper two only details are still extant,

1 KRUGLIKOVA 1974: 44-47, fig. 30; 1976: 93-96, fig. 54; 1986: 65-66, 102-103, fig. 60; FITZSIMMONS

1996: 282-288, fig. 4; LO MUZIO 1999: 59-61, fig. 17.

50 Ciro Lo Muzio [2]

namely the bent fingers of the right hand, higher up on the left at some distancefrom the figure, and the lower contour of the left arm extended aside. Hislower right arm is bent, the hand with the palm facing forward, whereas withhis lower left arm (of which only the wrist with a bracelet is visible at the levelof Parvatı’s breast) Siva embraces the goddess.

Fig. 1. Dilberjin “Temple of the Dioscuri” (after Kruglikova 1986).

Parvatı (on the right) sits in three quarter view to the right, with her legscrossed at the ankles; she wears a diaphanous dress draped on the legs andbracelets; she hugs Mahesvara’s shoulders with her right arm; her left handrests on a round, convex object with a radiating motif or flower inscribed,placed upright on Mahesvara’s left thigh.

The huge bull lies in profile towards the left with its head turned to theright; it wears a circular yoke decorated with circles and a cloth laid on its back,on which Siva and Parvatı sit.

The couple is flanked by two worshippers. The one on the right, muchbetter preserved, is a male figure standing in front view, wearing a tunic (?)with round red-bordered neck-opening, a caftan with borders and sleeves dec-orated by red bands, a red belt and trousers. In his raised right hand he holds asmall round object, in his left, near the chest, a no longer identifiable attribute

[3] The Umamahesvara in Central Asian Art 51

(fruit?). He has black hair, beard and moustaches; two short ribbons are visiblebehind his shoulders.

Fig. 2. Penjikent, Temple II, chapel 15 (after Marshak 1996).

Beneath the main group four much smaller figures are to be seen, three ofthem to the right of Siva’s foot-stool, the fourth one farther on the left. All ofthem wear a belted tunic; the second and the fourth from the left have a head-gear. The first (male?) figure on the left has his right hand raised with forefingerpointing toward the couple. The same gesture is performed by the last figure onthe right, whereas the second one points his right forefinger at his own breast.

As we have seen, Siva’s only extant attribute is a circular convex-shapedobject resting on his left thigh, on which Parvatı puts her hand. It may recall acakra, one among Vis

˙n˙u’s most distinctive emblems, which is found, however,

in some early images of Siva as well2. In this case, however, the size as well as

2 Cf. OHIcO on a Huvis˙ka coin (Göbl 1984: 44, type 7) and a Gandharan Sivaite figure in the

MNAOR, identified as Harihara (TADDEI 1985; HÄRTEL 1989: fig. 2). In this case, however, the attribute

52 Ciro Lo Muzio [4]

the convex shape make the attribute closer to a shield, which is less easily ex-plainable in a Siva image. A similar object decorated with a large rosette withpointed petals, that we may tentatively interpret as a shield, appears betweenthe legs of one of the two “Sivaite” figures painted in the room 16a at Dilberjin(Kruglikova 1974: fig. 7). The size of the attribute and the absence of a handleshould however discourage us to think of a mirror, one of the traditional em-blems of Parvatı (Verardi 1982: 285, fn. 15), especially if we compare it to themirror held by the “martial” goddess portrayed in the room 12 (Kruglikova1974: fig. 41a; 1977: fig. 11; Grenet 1987: pl. XXIV, fig. 2). On the contrary, theshield on which the same deity rests her left hand seems very close to the em-blem we have been so far trying to identify.

On account of the position of Siva’s upper arms, raised aside, we can riskthe hypothesis that, as in other Central Asian Umamahesvaras (cf. Kyzyl: figs. 3-5; Khotan: fig. 6, 7), the god might have held the sun and moon sym-bols.

As to the two main worshippers, the caftan decorated with red bands andthe short beard are to be seen in the male personage portrayed on a well-knownpainted vase from Merv (5th-6th century) (Koselenko 1966; Chuvin [ed.]: 2002:fig. 223); for the caftan we have a rather close parallel in two donors in a Pen-jikent mural3, whereas the same beard is worn by several figures depicted inthe Afrasiab paintings (7th century) (Azarpay 1981: pls. 21, 22) and by someworshippers in a Bamiyan mural4. The Dilberjin worshipper holds an uniden-tifiable object in his left hand and a very small circular attribute in his right.What remains of the figure standing to the left of the divine couple (only tracesof the left part of the silhouette and of the dress) does not allow to ascertain itssex.

On iconographic grounds, the date proposed by Kruglikova (1986: 75,108), i.e. the 3rd century AD circa, appears too early. This issue will be dealtupon in detail further on.

Penjikent (fig. 2). A clay sculptural group depicting Siva and Parvatı hasbeen unearthed in a chapel (room 15) situated to the south of the porch givingaccess to the outer court of the temple II5. The upper part of both figures (halfbust and head) is lost, therefore the number of the god’s arms and faces is not

recalls a solar disk rather than a wheel (central plain disk and marginal rays). A closer comparison forthe Dilberjin round emblem is provided by the cakra held by a syncretistic deity depicted on a Swat re-lief (GNOLI 1963: fig. 4; AGRAWALA & TADDEI: 1966).

3 Cf. painting in a chapel of the temple II (end of the 5th century): AZARPAY 1981: figs. 23, 24;Oxus 1993: 67).

4 Cf. painting in the cave of the 38 m Buddha, 7th century: KLIMBURG-SALTER 1989: pl. IV, fig. 4D.2 E, D.3 W.

5 SKODA 1992; MARSAK, RASPOPOVA, SKODA 1993: 93-95; MARSHAK 1996: 435-438, fig. 7-8.

[5] The Umamahesvara in Central Asian Art 53

ascertainable. Siva sits on his vahana in sukhasana, he wears dhotı, which con-ceals his urdhvalinga, the sacred thread (upavıta) and, oddly enough, boots. Inhis right hand, resting on the thigh, he holds what seems to be a fruit (matu-lunga?); with his left arm he embraces Parvatı’s back, in the hand he holds a tri-dent. The goddess sits in three quarter view on Siva’s left thigh; she wears ashort upper garment fastened with a belt and a skirt. Her left arm is wrapped ina shawl, of which she holds a hem in her hand; her right arm is lost. Hair tuftspainted black are reported to be still visible on her left shoulder (Skoda 1992:319). The bull is portrayed as lying in profile to the left, with the head turnedupwards. Potsherds found in the clay mass date the sculpture in the early 8th

century.An iconographic detail is to be remarked, namely the boots worn by Siva,

which, except for OHIcO, are not to be found in Indian as well as in CentralAsian representations of the god. Along with Mode (1992: 329), we note theunusual position of the bull’s head, and in particular the long curve of its neckwhich reminds us rather the buffalo slain by Durga Mahis

˙asuramardinı than Si-

va’s vahana in any other Umamahesvara, be it Indian or Central Asian.

Kyzyl. Devil Cave (passage B, annexe C) (fig. 3). The painting was detachedfrom the left wall6. In this case the Brahmanical couple appears in the fore-ground of a parivara scene, to the right of the Buddha. Both deities are haloedand sit in the European fashion in three quarter view towards the left, withcrossed ankles, in añjalimudra. Siva wears a “tiger” skin (black with white circlesand a white border), a long scarf covering his shoulders and winding round hisarms, leggings, two bands crossing on the breast clasped by a circular plaque,necklace, bracelets. He has three faces (the lateral ones are smaller); the frontface has third eye, jat

˙a with central round ornament, earrings; left face with

moustaches, and a small animal (calf?) head above the forehead; right facescarcely discernible. The god has four arms (only three of them are visible); heraises his back right hand holding a black circular attribute with white dots thatmay lead us to think of a fruit, although it is most probably to be interpreted as asolar disk. He sits on a long-horned bull lying in profile to the left, with rightfront leg stretched and left one bent.

Parvatı wears a tight garment decorated by groups of three dots and, onthe sleeves, by circle rows, scarf, crown and bracelets; she has an elaboratedhair-do. The goddess does not appear to sit directly on Siva’s vahana, but no

6 GRÜNWEDEL 1912: 136, fig. 297; 1920: II, 50, fig. 47; GAULIER, JERA-BEZAR, MAILLARD 1976: II,39, 50, fig. 106.

54 Ciro Lo Muzio [6]

other kind of seat is recognizable7.This is the only case in our repertory in which Siva does not embrace Par-

vatı, the couple being portrayed in one of the most standardized worshippingattitudes to be found in the painting of Kyzyl.

This mural is dated around the 600 AD (Gaulier, Jera-Bezard, Maillard,1976: II, p. 39).

Kyzyl. Maya Cave (3rd Foundation, cave 224; Grünwedel’s cave 5) (fig. 4)8.The Umamahesvara appears within a parivara scene, which includes the theme

Fig. 3. Kyzyl, Devil Cave (after Grünwedel 1912).

7 We cannot be sure that the cuirassed black-skinned figure to be seen behind the couple may re-present Karttikeya (GRÜNWEDEL 1912: 136). Just above the group another divine pair is to be seen(GRÜNWEDEL 1912: fig. 296): on the right, a male figure wearing caftan and large trousers, holding a sun-flower in his right hand, on the left, a female figure milking a child (Pancika and Haritı? cf. Gaulier,JERA-BEZARD, MAILLARD 1976: II, 39, 51, fig. 112).

8 GRÜNWEDEL 1912: 176, fig. 410.

[7] The Umamahesvara in Central Asian Art 55

of the Conversion of At˙avika; the painting was part of the iconographic pro-

gramme of the right wall (with respect to the stupa-pillar) of the cella. Siva andParvatı are represented in the foreground, to the right of the Buddha. Siva sitson the left in the European fashion with crossed ankles, wearing a tunic whichleaves his right shoulder uncovered. The god is blue-skinned, three-faced (cen-tral face with jat

˙a, three eyes; right face female [?]; left face hidden) and six-

armed; he embraces Parvatı’s shoulders with his left arm, holding her chin inhis right hand; the other extant (left) hands hold an astral symbol and a circularattribute (shield?). The god sits on the bull, of which only the fore part is visible(head in front view, bent legs, circular yoke). Parvatı wears a long-sleeved gar-ment; she is haloed. No other details are discernible.

Fig. 4. Kyzyl, Maya Cave, 3rd Foundation (after Grünwedel 1912).

56 Ciro Lo Muzio [8]

To the left of the Buddha, in the foreground, the yaks˙a At

˙avika takes a

dark-skinned child by the hand; both of them are standing with their legs half-plunged in a pond9.

For this painting a date no later than 650 AD had been proposed (Gaulier,Jera-Bezard, Maillard 1976: II, p. 39). However, 14C analysis carried out onsome murals detached from the same cella, now in the Museum für IndischeKunst in Berlin, have yielded a time-span comprised between 261-403 AD(painting with the Bodhisattva Maitreya in the lunette above the door, cf.Yaldiz et alii, s.d.: 210, no. 306) and 416-525 AD (painting with the parinirvan

˙a,

from the pradaks˙ina wall, cf. ibid.: 213, no. 309), i.e. between the mid-3rd and

the mid-6th century. On iconographic grounds (see further our discussion ofthe Indian evidence), we are inclined to believe that our painting belongs to thelatest part of this period10.

Fig. 5. Kyzyl, Gorge Cave (after Sérinde 1995).

9 This is an unexplained detail which is not witnessed in the Gandharan reliefs depicting this le-gend as well as in literary sources, cf. SANTORO 1988: 229.

10 The application of scientific methods (14C, thermoluminescence, etc.) is bringing about a dra-matic change in the traditional chronological layout of Xinjiang paintings, especially as far as its initialphase is concerned, cf. Howard 1991; Yaldiz et alii, s.d.: 191-192. I take the occasion to thank dr. F. Fi-lipponi who was so kind as to provide me with useful information about this issue.

[9] The Umamahesvara in Central Asian Art 57

Fig. 6. Dandan Oiliq (Khotan), painted wooden panel (after Williams 1973).

58 Ciro Lo Muzio [10]

Kyzyl, Gorge Cave, right side wall (fig. 5)11. Here once again is the Umama-hesvara depicted in a preaching scene, in the foreground, on the right; interest-ingly enough also in this case we find the yaks

˙a At

˙avika. Siva and Parvatı sit on

the bull in the European fashion with crossed ankles; both of them are haloed.Siva is black-skinned and wears a tiger-skin tied on the left shoulder, a scarf,leggings, a long garland, large circular earrings, a necklace and bracelets; he isithyphallic (with penis concealed by the dress) and three-faced: central facedark, slightly bent towards the goddess; side faces fair skinned; above them ani-mal heads. He has six arms; here also is he portrayed in the “chin-touching” at-titude; the upper hands probably hold sun and moon (only his right is visible),the middle right hand rests on the thigh, the middle left is no longer visible.Parvatı holds no attribute, her head is slightly bent upwards in three quarterview. The bull lies in profile toward the left, with fore legs bent, head in threequarter view to the right.

In the foreground, on the left, the yaks˙a At

˙avika stands in a pond taking a

child by the hand12. Above the yaks˙a, a blue-skinned, six-armed divine figure

holding sun, moon and sankha, i.e. a deity sharing several traits of Vis˙n˙u.

This painting has been dated in the 7th century (Sérinde 1995: 309; but seethe remarks made above on the Maya Cave painting).

It is to be observed that in all three instances Siva holds the solar and lunarsymbols. In the Gorge Cave only (fig. 5) are both deities clearly seated on thebull and Siva is ithyphallic; in this cave and in the Maya Cave (fig. 4) the god issix-armed, he hugs Parvatı’s shoulders putting a hand under her chin; bothdeities look at each other. Another detail common to both representations areSiva’s large circular earrings. In two cases (Gorge and Maya Caves) the sceneincludes the theme of the conversion of the yaks

˙a At

˙avika, who is represented

on the left of the Buddha. In the Devil Cave, Siva and Parvatı sit side by side inañjalimudra in a parivara scene.

Dandan Oiliq (Khotan) (fig. 6). Wooden painted panel; 8th century13. Sivasits in front view, with crossed legs; he wears a tight, long-sleeved, white gar-ment, tiger-skin concealing his urdhvalinga, upavıta, necklace, keyura, bracelets.He has three faces: central head in three quarter view to the left, with diadem,black jat

˙a; right face with diadem; left face terrific, with third eye, frowned eye-

brows, wide open mouth. Four arms: in his back right and left hands he holds

11 Berlin, Museum für Indische Kunst, inv. MIK III 8725. GRÜNWEDEL 1920: taf. xxvi, fig. 4, II,65-67; YALDIZ 1987: Taf. XI. The painting is illustrated also in Sérinde 1995: 309, but it is not included inthe descriptive item (no.238), which gives account of other sections of the mural.

12 On this detail, see fn. 11.13 Lahore, Central Museum, inv. D.X.8. STEIN 1907: II, pl. LXII, obv.; WILLIAMS 1973: 142, fig.

52.

[11] The Umamahesvara in Central Asian Art 59

sun and moon, front right arm embracing Parvatı’s shoulders, front left handholding vajra and resting on the thigh; halo and mandorla. The goddess, de-picted in smaller scale, is black-haired and wears a diadem; she kneels on Siva’sright thigh, staring at the god’s face and leaning toward him a bowl she holds inher right hand.

A number of iconographic details make the Dandan Oiliq Siva very closeto another isolated image of the god from the same site (Williams 1973: fig. 51;Whitfield 1984: pl. 70). In fact in both cases we find tiger skin, astral symbols,vajra, ithyphallism, three faces (one of which is terrific) and a diadem with a cir-cular ornament.

Fig. 7. Khadalik (Khotan), painted wooden panel (after Williams 1973).

60 Ciro Lo Muzio [12]

Parvatı’s pose is unusual, as she kneels on Siva’s right thigh, looking at himand leaning a cup in his direction.

Khadalik (Khotan) (fig. 7). Wooden painted panel, 8th century14. Only theleft part of the panel is preserved and the painting on it is in bad condition. Wecan discern Parvati’s head turned towards the right, with a head ornament inthe form of a rosette, her right arm bent before her breast, Siva’s lower rightarm embracing the goddess and holding an unclear attribute, his upper rightarm holding the sun disk. Siva is blue-skinned. It is not clear whether Parvatı issitting or kneeling on Siva’s right thigh, however the iconographic scheme musthave been very similar to the one on the Dandan Oiliq panel describedabove.

To sum up, with the only exception of the Devil Cave mural at Kyzyl (fig.3), in all Central Asian representations Siva and Parvatı are portrayed as sittingon the bull embracing each other. This is a relevant clue to the understandingof their iconographic source and chronology. The epoch of the spread of theUmamahesvaramurti can be fixed between the 6th and the 8th century, i.e.within a relatively short period, which is in agreement with the typological uni-formity of almost all the representations we have listed, the main difference be-ing the absence of the bull in the Khotanese paintings.

It goes without saying that the chronology proposed for the Dilberjin mu-ral (3rd century A.D.) sounds like a false note15. As a matter of fact, doubts havebeen cast about the whole chronological frame of the site. In a previous article(Lo Muzio 1999), we have tried to show that one of the pivotal arguments ofKruglikova’s chronology – i.e. a Graeco-Bactrian date (approximately the firstquarter of the 2nd century BC)16 for the painting with the Dioscuri found in thesame temple (period I) – is to be rejected; on iconographic grounds an attribu-tion of the mural to the 2nd or even the 3rd century AD appears more reason-able. As to the Umamahesvara, Kruglikova (1986: 75) relates it with the reign ofthe Kushan king Vasudeva I, that is the 3rd century AD (but earlier she thoughtof Vima Kadphises, 1st century AD: Eadem 1974: 46)17. However she fails tosubstantiate her attribution of the painting to the Kushan epoch, as no persuad-

14 New Delhi, National Museum, inv. Kha. i. 51. WILLIAMS 1973: fig. 51.15 Scepticism about such an early date was expressed by Fussman (1978: 428, on numismatic

grounds) and Bernard (BERNARD - FRANCFORT 1979: 126 fn. 7, with regard to the painting with theDioscuri).

16 KRUGLIKOVA 1974: 26-27; 1986: 7 ff.17 In both cases the hypothesis is based on the supposed Sivaite faith of the two Kushan sover-

eigns. Such a considerable (and unexplained) shift in the opinion of the Russian scholar patently betraysthe weakness of stratigraphic data.

[13] The Umamahesvara in Central Asian Art 61

ing Indian or Gandharan parallel can be brought forward. Verardi (1982: 285fn. 15) rightly stressed the point that the Dilberjin couple derives from a quitespecific iconographic type, widespread in post-Kushan India. On the otherhand a long list of inconsistencies affecting Kruglikova’s accounts of the Dilber-jin excavations has been more recently highlighted by Fitzsimmons (1996) onthe grounds of a thorough revision of stratigraphic and structural data; on ac-count of the affinity with the Tapa Skandar marble group depicting Siva andParvatı (fig. 17), he assigns the Dilberjin Umamahesvara to the 7th-8th century(ibid.: 287). But we shall turn back on this issue further on.

Fig. 8. Surkh Kotal, fragment of frieze (F1) (after Schlumberger-Le Berre-Fussman1983).

There are two further pieces of evidence that have been brought forwardin witness of an early spread of the cult of Siva and Parvatı in Kushan Bactria,namely a fragment of the sculptural decoration of the Surkh Kotal shrine and astone sculpture from Airtam (Old Termez). We did not think it right to includethem in our list of ascertained iconographic evidence (for the reasons that willbe given below), nonetheless we cannot fail to discuss them.

According to Fussman, Siva with the bull Nandi is shown on three frag-ments of frieze (F1, F2, F10) found in the cella of the main temple (A) of Surkh

62 Ciro Lo Muzio [14]

Fig. 9. Surkh Kotal, fragment of frieze (F2) (after Schlumberger-Le Berre-Fussman1983).

Kotal (figs. 8-10)18. In one case (F2) the god is accompanied by Parvatı(fig. 9). The presence of Sivaite images in a religious building of declaredKushan dynastic patronage would represent a crucial piece of evidenceon the place assigned to the worship of Siva within the royal religiousideology (an issue since long debated with regard to the numismatic OHIcO19),as well as within the context of this particular temple. Nevertheless, thevery small scale and the approximate treatment of the figures, as wellas their general attire (naked with mantle or anklets) and bad state ofpreservation (all of them are headless), do not make them very differentfrom the Erotes depicted on other fragments of frieze unearthed in thesame building (Schlumberger, Le Berre, Fussman 1983: pl. 55.155-159).If we add that on each of the fragments at issue two bulls are to beseen, it is difficult to follow Fussman when he maintains that: «les personnages(...) des blocs F1, F2 et F10 sont clairement identifiables pour qui adéjà manié des monnaies kouchanes: il s’agit de Siva appuyé sur le taureauà bosse Nandin (F1, F10) et accompagné de Parvatı (F2)». As a matterof fact, when Siva is accompanied by his vahana, only one bull is visible.This holds true since the earliest representations of the god up to ourdays, including his countless depictions in the Kushan coinage20. The

18 SCHLUMBERGER, LE BERRE, FUSSMAN 1983: 110, 112, 149, pl. 55. 152-154, 160; FUSSMAN 1989:198.

19 TANABE 1991-92; LO MUZIO 1995-96; CRIBB 1997; ZEYMAL 1997.20 The exceptional representation of two bulls lying symmetrically at the feet of Mahesvara on a

well-known wooden painted panel from Dandan Oiliq, Khotan (7th century AD circa) (WILLIAMS 1973:fig. 51; WHITFIELD 1984: pl. 70) is likely to be related to the later iconographic type of the kingly or di-vine figure seated on zoomorphic throne, be it supported by lions, rams, horses or camels; a widespread

[15] The Umamahesvara in Central Asian Art 63

identification proposed by Fussman should therefore be taken with cau-tion.

Fig. 10. Surkh Kotal, fragment of frieze (F10) (after Schlumberger-Le Berre-Fussman 1983).

With no less caution should we consider another supposed clue to aSivaite presence in the Surkh Kotal shrine, namely the trident-shaped marksengraved on some steps of the monumental stairway leading to the temple, inwhich Fussman does not hesitate to see Sivaite trisulas (fig. 11)21. A clue whichis not easy to manage from a chronological viewpoint, as the author admits. Onstratigraphic grounds the marks are chronologically unrelated to the temple A(and therefore to the “Sivaite” fragments discussed above), on the other handthey are hardly to be linked with the later fire-temples B and D. Therefore weshould place the tridents either somewhere between the two main phases or af-ter the temple B had ceased to exist. In both cases, to explain the frequentationof the shrine by Sivaite pilgrims, who were free to engrave trisulas on the steps

pattern in Sasanian and Central Asian art. On the contrary, the motif of two addorsed bulls (zebus) isfound in a purely ornamental context in capitals from Bactrian sites of the Kushan period, namely ShamQala (not far from Surkh Kotal) and Kara Tepe, see STAVISKIJ 1973: figs. 9, 10.

21 Trident marks appear also on some stone blocs reemployed in the pit at the bottom of the stair-way (SCHLUMBERGER - LE BERRE - FUSSMAN 1985: 73-74, pl. 122).

64 Ciro Lo Muzio [16]

Fig. 11. Surkh Kotal, stairway leading to the temple (after Schlumberger-Le Berre-Fussman

1983).

of the monumental stairway, is no easy task, as there are no proofs ofa Sivaite interlude during the existence of the Bactrian temple or of alater consecration of the shrine to the cult of Mahadeva. According toFussman «les nombreux pèlerins qui gravèrent les trisula sur les marchesdu grand escalier de Surkh Kotal avaient trouvé au sommet de la collinede quoi penser qu’il s’agissait d’un sanctuaire sivaïte: soit un bloc sculptéanalogue aux blocs de frise; soit une statue de Siva ou interprétée commeétant de Siva; soit une pierre considérée comme un linga» (ibid.: 152).However, to begin with the very assumption that Siva had «nombreuxpèlerins» in Kushan Bactria, all these hypotheses pertain more to the realmof conjectures. We think it safer not to rule out that the Surkh Kotaltridents might have been stone-cutter tamghas22. In this case we wouldbetter compare them to the trident–shaped marks born by some stoneblocs of the defensive wall of the late Sasanian Darband, in Daghestan

22 As can be inferred from the photograph (fig. 11), no bloc bears more than one trident, a regula-rity which seems more suitable to construction workers than to worshippers.

[17] The Umamahesvara in Central Asian Art 65

(most probably built during the reign of Khusraw I Anusirwan, 531-579AD), where no Sivaite implication can be surmised (fig. 12)23.

Fig. 12. Darband (Daghestan), Trident-shaped marks on stone-blocs (after Gadziev-Ku-drjavcev 2001).

The Airtam stone sculpture is composed of a bloc serving as a pedestalto a relief representing a standing couple, of which only the lower partis preserved (fig. 13). A six-lined inscription in Bactrian script engravedon the base dates the sculpture to the 4th year of Huvis

˙ka’s reign (2nd

century AD)24. The stratigraphic context is not very clear and the blocwas not found in its original location; at any rate a relationship withthe Buddhist ruins found in that sector cannot be excluded. The twopersonages stand side by side; the female figure, on the right, wears along himation-like dress and heavy anklets bearing a rosette each, theright leg is crossed in front of the left; the male one keeps no tracesof clothes. Bernard (1981: 327-328) deems it plausible to identify themas Siva and Parvatı. As a matter of fact, the general impression one getsfrom this sculpture is that its iconographic source might have been Indianor possibly Gandharan, as the dress of the female figure seems to convey.What remains of the relief, however, does not preserve so specific traitsas to allow us to think of Siva and Parvatı rather than any other tutelarycouple or even a generic mithuna, the auspicious couple that often flanksthe entrance to Buddhist caityas25. In conclusion, neither Surkh Kotal norAirtam seem to provide unquestionable proofs on the spread of the cultand iconography of Siva and Parvatı in Kushan Bactria.

23 For the rich repertory of tamghas in the Darband fortification, see GADZIEV-KUDRJAVCEV

2001.24 BERNARD 1981; see also ABDULLAEV et alii (eds.) 1991: I, no. 193 (with the text of the inscription)

and RTVELADZE et alii (eds.) 1991: no.23. The inscription refers to works carried out in a temple, to thebuilding of large gates (?) for the gods and of dams. No god is mentioned, however.

25 With regard to Bernard’s interpretation, scepticism has been expressed by Fussman (1987:359) and Rtveladze (in CHUVIN [ed.] 1999: 100). The latter does not discard the hypothesis that the relief

66 Ciro Lo Muzio [18]

An overview of the development of the Umamahesvara in Indian art willhelp us trace the original models of its Central Asian representations as well asplace the Dilberjin painting in a credible chronological frame. Our analysis ofthe Indian repertory will lay particular stress on the iconographic schemes, inother words on the position and the attitude of the two deities, as they are, inour opinion, pivotal elements for singling out the source of inspiration of ourCentral Asian images.

Fig. 13. Airtam, stone relief (after Rtveladze et alii 1995).

No Gandharan specimen of this subject was known until the recent publi-

may represent a seated king (Huvis˙ka) flanked by a standing woman. However, the representation of a

barefoot and trousersless Kushan sovereign does not seem very sound.

[19] The Umamahesvara in Central Asian Art 67

Fig. 14. Kashmir or Northern Pakistan, linga-shaped dyptich (after Rowan 1985).

cation, by J. Sherrier (1991), of three reliefs depicting Siva and Parvatı, all ofthem belonging to the author’s collection. In one case (ibid.: pl. 48.1a) Siva andParvatı figure in a “quartet” along with Karttikeya and the six-headed goddessS˙as˙t˙hı (cf. Agrawala 1995: fig. 1); in another relief (ibid.: pl. 48.2a) they are ac-

companied by Karttikeya only, represented between them in smaller scale andriding a peacock; in the third relief (ibid.: pl. 48.4a-b) we find again their sonportrayed between them in much smaller scale, and two worshippers, male andfemale, depicted on Siva’s and Parvatı’s side respectively. A point worth being

68 Ciro Lo Muzio [20]

stressed is that all three examples show the couple as standing, and none ofthem includes Siva’s vahana, the bull. The same applies to the Huvis

˙ka coin

showing the couple OMMO/OHIcO (Uma/Siva)26.At Mathura the Siva-Parvatı couple is a relatively common subject, yet

their representations show a rather strict adherence to an iconographic patternwhich has very little in common with our Central Asian specimens, and whichgoes on being witnessed in India until the Gupta epoch (Kreisel 1986: 154-160,figs. 102-108). The Mathuran depictions of the pair show the two deities asstanding in frontal view, side by side. Siva has one head, two arms and ur-dhvalinga; he hugs Parvatı’s shoulders with his left arm. The bull appears be-hind them, standing or lying. Moreover, in most cases we deal with rather smallterracotta reliefs, whereas stone specimens of larger size, but following theiconographic type described above, appear only in the Gupta epoch (ibid.: 156ff., figs. 105, 106a-d). In the same way (but without contact) is the couple repre-sented on a terracotta relief from Kaushambi (second half of the 4th centuryA.D.) (Banerji 1913-14: 292, pl. LXXb; Banerjea 1956: 467, pl. XXXVIII, fig.2; Kreisel 1986: fig. A24). A few later examples from the North-West of theSubcontinent can be mentioned. Siva and Parvatı are represented as standingside by side on a 6th century relief from Baramula, in Kashmir, where we findsun and moon among Siva’s attributes (Kumar 1975: pl. XI), as well as on alinga-shaped dyptich (fig. 15) from Northern Pakistan (Taddei 1964-65) orKashmir (Goetz 1965; Rowan 1985: 279-80, pls. 39A, 39B), dated to the 7th-8th

(Taddei) or to the 10th century (Goetz). In this case the couple is portrayed asstanding before the bull; Siva puts his arm on the goddess’ right shoul-der.

The two deities are shown as sitting, but separated, on a few terracottas ofthe Kushan epoch from Bhita (U.P.) and Amreli (Gujarat) (Joshi 1989: 49). Ona 4th century terracotta relief from Rang Mahal (fig. 14)27, Siva and Parvatı areseated, but neither in this case is there any contact between them. Siva’s vahanaresembles more a lion than a bull. The solar and lunar symbols, here held by asmall long-haired figure emerging above the god’s head28, provide an interest-

26 ROSENFIELD 1967: 94, c.166; CRIBB 1997: 35, pl. G8. Against the equation OMMO=Uma, seeGÖBL 1984: 43, who reads the goddess name as NANA[NONO]; such interpretation seems to be sup-ported by another coin of the same monarch in which OMMO is replaced by NANA, depicted in a verysimilar attire and position, but holding different attributes (CRIBB 1997: 35, pl.G9). OMMO and NANA,however, are separately mentioned among the gods listed in the lines 9 and 10 of the Rabatak inscription(SIMS-WILLIAMS – CRIBB 1995-96: 79, 108). Strangely enough, the Rabatak “pantheon” does not includeOHIcO, whereas in an interlinear addition Mahasena and Visakha are mentioned.

27 R.C. AGRAWALA 1956: fig. 1; V.S. AGRAWALA 1960: pl. XXIV.14; KREISEL 1986: pl. A16.28 Maxwell (1983: 41-44) identifies the personage holding the astral emblems as a raudra figure

symbolizing the control over Time, i.e. over birth, life, death and rebirth, the first appearence of which is

[21] The Umamahesvara in Central Asian Art 69

Fig. 15. Terracotta relief from Rang Mahal (after Kumar 1975).

recorded in an earlier relief (ca. 50 AD) from Musanagar (U.P.) depicting Siva in a proto-visvarupa form(ibid.: fig. 1; KREISEL 1986: Abb. A15).

70 Ciro Lo Muzio [22]

ing link with the Central Asian Sivaite pairs29. Moreover, the presence of a maleand a female figure kneeling on both sides of the couple is worth beingstressed.

Fig. 16. Aihole, stone relief (after C. Sivaramamurti, India, Ceylon, Nepal, Tibet, Torino

1988, p. 62).

It seems that the iconographic type showing Siva and Parvatı sitting andembracing each other – as they are depicted in Central Asia – is the one whichis properly referred to as Umamahesvaramurti. According to the Vis

˙n˙

udhar-mottarapuran

˙a: «Siva and Uma should be seated on a seat, embracing each

other. Siva should have the jat˙amakuta on his head with the crescent moon

stuck on it; he should have two arms, in the right one of which there should bea nılotpala flower and the left one should be placed in embrace on the leftshoulder of Uma. Umadevı should have a handsome bust and hip; she shouldhave her right hand thrown in embrace on the right shoulder of Siva andshould keep in her left hand a mirror» (Rao 1968: II, pt. 1, 132-133). A similar

29 Apart from the evidence provided by the reliefs from Musanagar and Rang Mahal as well assome later sculptures from Kashmir, sun and moon cannot be listed among the most typical emblems ofSiva. Moreover in Central Asian art these attributes are not specific to Siva, being found in the hands ofmany divine figures and hinting to a broad popularity of astral cults. On this subject, which surely needsfurther investigation, see BUSSAGLI 1949 and 1962; KLIMKEIT 1983.

[23] The Umamahesvara in Central Asian Art 71

Fig. 17. Tapa Skandar, stone sculpture (after Kuwayama 1976).

description of the group, with the addition of some relevant details, is found inthe Rupamand

˙ana (loc. cit.), according to which Siva should have four arms, in

72 Ciro Lo Muzio [24]

one of his right hands he should hold a trisula, in the other a matulunga fruit;one of the left arms should embrace Uma, whereas the other left hand shouldhold a snake; the group should include Siva’s bull, Gan

˙esa, Kumara and the

emaciated figure of the r˙s˙i Bhr

˙ingi. But for a few details, the latter description is

perfectly matched by a relief from Aihole (fig. 16), to-day in the Prince ofWales Museum (Mumbai); therefore in the 6th century this iconographicscheme was an accomplished fact. It is not just as simple to fix its beginning.The Vis

˙n˙

udharmottara, which is the result of a stratification that may havelasted for some centuries (probably up to the 9th c.) on a core dating from the5th century30, provides no precise chronological clue, but for an approximateterminus post quem, i.e. the 5th century. At any rate, there is no artistic evidenceof the Umamahesvara type showing the two deities seated and embracing eachother earlier than the 5th century, a remarkable spread of that specific icono-graphic being already witnessed in the 6th century. Beside a relief from Chanda,in Maharashtra (Kramrisch 1981: fig. 49) and a relief from Uttar Pradesh (ibid.:fig. 50), that Joshi (1989: 50) considers as the earliest evidence, a Kashmiri reliefof the Fatah collection (fig. 18), that we will examine in more detail later on,shows that in the 6th century the Umamahesvaramurti was known also in theNorth-West of the Subcontinent.

Fig. 18. Kashmir, stone relief (after Barrett 1957).30 On this subject, see COLLINS 1988: 38-39, and fn. 26.

[25] The Umamahesvara in Central Asian Art 73

As to the variant depicting the deities as looking at each other, while Sivatouches her wife’s chin, the proposal to date its appearance in Indian iconogra-phy as late as the 8th century, specifically in the Pratihara artistic milieu (Joshi1989: 51), has to be rejected, the “chin-touching” attitude being already wit-nessed at the latest in the 6th century at Kyzyl (figs. 4, 5). We should expect it tohave been known in India some time before, probably in the 5th century.

If we take into account the position in which Siva’s vahana is represented(and we find it in most Indian Umamahesvaramurti, often along with Parvatı’slion), further distinctions can be pointed out. The bull can be shown as stand-ing or lying behind Siva (as in the Aihole relief, fig. 16) or, in much smallerscale, lying at the feet of the god, a manner which was to become almost canon-ical in the following centuries. The vr

˙s˙arudhamurti, i.e. Siva and Parvatı sitting

directly on the bull’s back, is the source to which our specimens from Dilberjin,Penjikent and Kyzyl (in the Gorge and Maya Cave [fig. 5], whereas in the DevilCave [fig. 3] Siva only sits on his vahana) are to be related. As to the date of itsintroduction, nothing more precise can be said than what we have remarkedabove with regard to the Umamahesvaramurti in general31. As to its homeland,the vr

˙s˙arudhamurti is thought to have been particularly popular in Western

Madhya Pradesh and Rajasthan (Joshi 1989: 54), however most relevant is forour subject the evidence provided by the region between the Hindu Kush andKashmir.

Two sculptures deserve particular attention: the marble group from TapaSkandar (fig. 17) in Afghanistan32, and a small Kashmiri relief of the Fatah col-lection (fig. 18)33. In both representations the couple sits directly on a lyingbull, Siva is in sukhasana and leans one of his arms on Parvatı’s right shoulder;the goddess wears a long folded robe and holds a lotus in her right hand34.Moreover, both sculptures show a similar rendering of the lower edge of the

31 The so far unique piece of evidence provided by the Pearse seal, labelled as «Gandharo-Sasa-nian» (3rd century?), on which the deity is shown seated on his vahana and holding some of the attribu-tes usually held by OHIcO (BANERJEA 1956: 468-469, pl. XXXIV, fig. 1; KREISEL 1986: 94, fig. 83), has ledKreisel (loc. cit.) to surmise a Gandharan origin for this iconography. However, we have to rememberthat, either alone (cf. the Siva on a Gandharan relief found in Mesopotamia, INVERNIZZI 1968-69: fig.102; TADDEI 1971; a relief from Akhun Dheri: HÄRTEL 1989: fig. 3) or along with Parvatı (reliefs of theSherrier collection, cf. above), Siva is never depicted as seated on vr

˙s˙

a in Kushan coinage or in Gandha-ran sculpture. In this regard, a rather odd relief belonging to a Japanese collection is worth mentioning:on the left, Pharro and Ardokhso are seated on a lying bull, on the right two female musicians are stan-ding or moving towards the couple (TANABE 1986: 37, II-6). Unless the photograph has been misprinted,the traditional position of the two deities is inverted (cf., e.g., FOUCHER 1905-51: figs. 382-384). Are we tosuspect that the motif of Siva-Parvatı sitting on vr

˙s˙

a was already known in the North-West and that acontaminatio between the two divine couples had taken place, or should we rather consider with suspi-cion this relief that perhaps displays too many unusual features?

32 For a detailed description of this sculpture, see KUWAYAMA 1976: 381-383, figs. 14-17.33 BARRETT 1957: 59, fig. 12; PAUL 1986: 129, fig. 60.34 BARRETT (loc. cit.) surmises she could have held a mirror in her left hand.

74 Ciro Lo Muzio [26]

dress hanging from the border of the basement. In the Tapa Skandar group,that we should perhaps consider a Somaskandamurti, as it shows the childSkanda at his mother’s feet clinging to her skirt, Siva is ithyphallic and four-armed, his only extant attribute is the trident (the shaft is missing) that the godholds from the prongs35. In the Fatah relief the god is three-headed (as in mostof his representations from the same area), six-armed and preserves all of hisemblems: sun and moon36, trident (only part of the shaft is preserved), lotusflower, rosary and club; the two deities look at each other and both of themwear large ear-rings. As to chronology, for the Fatah relief the 6th century (Paul1986: 129) seems a much more plausible date than the 7th-8th century, proposedby Barrett (1957: 59). The Tapa Skandar group is a remarkable specimen of themarble sculpture that flourished during the reign of the Turki Shahi, i.e. the 7th-8th century (Kuwayama 1976); however, given the rather close iconographicaffinity with the Fatah relief, we would suggest for it a date nearer to the earlierterm of the time span proposed by Kuwayama, i.e. the 7th century.

What we have been discussing so far can be summarized as follows. InCentral Asia Siva and Parvatı are depicted as sitting on the bull (except forKhotan, figs. 6, 7) and embracing each other (except for Kyzyl, Devil Cave, fig.3). In two instances at Kyzyl (figs. 4, 5) they look at each other and Siva touchesParvatı’s chin. In Indian art none of the above said attitudes is attested until theend of the Gupta epoch. Before then, when the two deities are shown as seated(Bhita, Amreli, Rang Mahal, see above), there is no contact between them. The“chin-touching” attitude must have been introduced not later than the 5th cen-tury; there are no representations of the couple sitting directly on the bull (vr-

˙s˙arudhamurti) before the same epoch. This iconographic type is witnessed in

the North-West of the Indian subcontinent from the 6th c. onwards (Kashmir,Tapa Skandar).All this having being said, Kruglikova’s proposal to date the Dilberjin paintingto the 3rd century can be safely rejected. However, rather than in the 7th-8th cen-tury (Fitzsimmons 1996: 288), we deem it cautious to date the mural some-where between the 6th and the 7th c., at any rate not earlier than the 6th.

Having so far focused on the iconographic peculiarity of Central AsianUmamahesvaras and on their plausible sources in Indian art, we should findout the very reasons and the religious implications of the spread of this Indianicon in Western and Eastern Turkestan.

35 The pedestal of the sculpture bears a brahmı inscription in which the gods Brahma, Vis˙n˙u and

Mahesvara are mentioned, see GUPTA - SIRCAR 1973.36 The attribute held in Siva’s proper uppermost right is more likely to be interpreted as a moon

emblem, rather than a kapala, as Barrett proposed (1957: 59).

[27] The Umamahesvara in Central Asian Art 75

In their motherland, Siva and Parvatı boast an iconographic repertorythat, for variety, richness and diffusion, is unparalleled by other Indian divinecouples. It ranges from the themes more specifically linked to marriage – Kalya-n˙asundaramurti and Vaivahikamurti – to those which can be considered as

family pictures, i.e. Umasahitamurti, Umamahesvaramurti and Somaskan-damurti.

Just as all major Indian icons, the image of the Sivaite couple lends itself tobe interpreted at different levels. From a highly speculative standpoint, it is arepresentation of the union between purus

˙a and prakr

˙ti, essence and sub-

essence (Kramrisch 1981: 58). The two gods «are the first self-revelation of theabsolute, the male being the personification of the passive aspect which weknow as Eternity, the female of the activating energy (sakti), the dynamism oftime. Though apparently opposite they are in essence one» (Zimmer 197422:139).

Siva and Parvatı are no less suitable to embody common people’s aspira-tion to wealth and fecundity within family life. In the Brahmanical pantheonthey are the main depositories of the values connected to marriage. Their unionis the celestial and auspicious archetype of conjugal love.

Commenting on a 10th century relief from Bengal, in which an Umamahe-svara is flanked by a whole family – the father with a son on Siva’s side, themother with two daughters on Parvatı’s – Zimmer wrote: «Thus the humancouple are shown to participate in the mystical union of the divine; they, too,are of one flesh» (19742: 138, fig. 3). A similar picture is already found in themuch earlier relief from Rang Mahal (fig. 14). As pointed out by R.C. Agrawala(1956: 63), the attire and the ornaments of the couple are similar to thoseshown by the two deities, as to claim a sort of assimilation between the humanand the divine couple, the archetype of the ideal marriage, in which «She [Par-vatı] has her living counterpart in every woman as the God in every man» (Zim-mer 19742: ibid.: 140).

How can we make sense out of the diffusion of their icon in Central Asia?In Xinjiang the Umamahesvara always occurs in a Buddhistic context, in whichBrahmanical deities are not certainly an anomalous presence. At Penjikent andat Dilberjin, however, the appearance of the Sivaite couple, as well as of singlerepresentations of Siva (or divine figures with Sivaite traits) might stimulatecontroversial interpretations. Yet we feel that, beyond the diversity of the reli-gious milieux in which the couple is represented, a univocal explanation can beventured. Our first step will be once again the Dilberjin temple.

As a matter of fact, the destination of this building, no less than itschronology, has soon appeared a puzzling question, one of the main reason be-ing that the subjects of the paintings unearthed in it (the Dioscuri and theUmamahesvara) seem to testify to a drastic change in the cult. This is an issue

76 Ciro Lo Muzio [28]

we have dealt upon at length elsewhere (Lo Muzio 1999), nonetheless we thinkit useful to summarize our opinion. The painting with the Dioscuri (Periods Iand II) is in itself no evidence that the temple was consecrated to the GreekTwins. The mural was executed on the Western wall of the portico, on bothsides of the door giving access to the sanctum; hence the two gods were rather asort of assistants of a major deity, the image of which is no longer extant butwas probably housed in the cella. An overview of the Graeco-Roman icono-graphic repertory of the Dioscuri led us to risk the hypothesis that the Dilberjintemple might be considered as the easternmost evidence of a theological andiconographic pattern which enjoyed large popularity both in the Western andin the Eastern part of the Roman Empire: the Dioscuri assisting a goddess. Apattern that probably found in the Balkh oasis fertile soil in similar beliefs (i.e.,the cult of the Divine Twins associated to a goddess) that were deeply rooted inthe Indo-Iranian religious domain. In other words, despite their disguisingClassical attire, the Dilberjin Dioscuri are more likely to be interpreted as Bac-trian Twins assisting a local (Iranian) goddess, and that their main role was theprotection of marriage and, probably, children.

In the execution of a mural depicting Siva and Parvatı (fig. 1), in PeriodIV, many scholars, beside Kruglikova, have seen the unambiguous evidencethat the temple had become a Sivaite shrine. Bernard (1987: 61), for instance,sees no reasons why we should not consider Siva as the main god worshippedin the temple, perhaps since the third or even the second period37. In this casewe should suppose that an image of the god Siva (or a linga) was placed in thecella. In fact, just like the mural with the Dioscuri, the Umamahesvara waspainted on the western wall of a vestibule (room 2), not in the cella; so it wasnot the main object of worship. In this regard we have suggested an alternativehypothesis, namely that in a later epoch and under the action of a different cul-tural influence, possibly originating in North-Western India, the replacementof the Dioscuri with Siva and Parvatı meant no drastic change, as there is an al-most perfect overlapping between the Greek Twins and the Indian couple asfar as their functional domain is concerned. If the presence of the Dioscuri wasjustified by their being «au service d’une déesse», the same might have been truefor the Umamahesvara (Lo Muzio 1999: 61): the Dilberjin temple may not havechanged its destination throughout its existence.

However conjectural, our suggestion seems at least not to be contradictedby the Penjikent Umamahesvara (fig. 2). The clay sculpture was found in theroom 15, which is one of the two chapels (rooms 14 and 15) that flanked a cen-tral porch (13) giving access to the main courtyard of the temple II. The point

37 Cf. also in BERNARD - FRANCFORT 1979: 128.

[29] The Umamahesvara in Central Asian Art 77

has been stressed that the earlier door linking the chapel with the courtyardwas walled in the beginning of the 8th century, probably when the Sivaite sculp-ture was placed in its niche. Since then the small shrine could be entered onlyfrom the street (Marsak, Raspopova, Skoda 1993: 93-95). Be that as it may, wehave to remark that the chapel was set within the precinct of the sacred areaand, what makes the thing all the more interesting, in a symmetrical positionwith the six-columned chapel 14 in which a clay sculpture of the goddess Nanasitting on a lion has been found. Another clay sculpture depicting the same de-ity mounting her animal vehicle was found in the porch (13), to the left of thedoor leading to the courtyard. We are led to suspect that, although mutuallyisolated, the two chapels (14 and 15) might have been ideally interrelated andthat a sort of affinity was understood among them, a relationship between thecult addressed to the main goddess of Sogdia and the worship of the Brahmani-cal couple, justifying their topographical proximity38. If this were a valid sug-gestion, it could add weight to the hypothesis we have advanced for the Dilber-jin Umamahesvara.

An aspect which is worth being stressed is that the Penjikent Umama-hesvara, and in particular the image of Siva, has almost nothing in commonwith the other Sivaite images yielded by the same site, namely the well-knownand much debated Wesparkar39 and the “blue god”40 (which in their turnare clearly inspired to different iconographic types). It seems, in otherwords, that these divine figures (or, in some cases, their image) reachedthe Sogdian town each following a path of its own and that they wereworshipped by different social groups, each of them in its own way linkedwith India. Of all the Sivaite images discovered so far at Penjikent theUmamahesvara seems the purest, which is the reason why, according toMarsak (1996: 435-436), it could not be fully integrated in the Sogdianpantheon, as the isolation of its chapel shows. Its iconography appearsas the least contaminated, the boots worn by Siva being the only concessionmade to the Central Asian taste. Even the linga concealed by the dhotıneeds not be explained on account of «the reluctance of Sogdian artiststo represent nudity» (Grenet 1994: 46), being on the contrary a commonmanner of representing Siva’s penis in the Umamahesvaramurti, both inIndia (from Rang Mahal onwards) and in Central Asia. We must addthat this is the first case in Sogdia that a Hindu image has been found

38 A similar hypothesis has been advanced by Mode (1992: 330) on account of the old relationshipbetween OHIcO (Siva) and NANA sometimes represented as a couple on the Kushan coinage (cf. abovesee above, fn. 18).

39 HUMBACH 1975; TANABE 1991-92; MARSHAK, in AZARPAY 1981: 30; MARSHAK 1995-96: 305.40 In the rooms VII/14 (BANERJEE 1969: fig. 2; CHUVIN [ed.] 2002: fig. 170) and VI/8 (BELENICKIJ

1973: 23).

78 Ciro Lo Muzio [30]

in a shrine; a shrine that was expressly dedicated to its worship andthat was situated within one of the two main sanctuaries of the town.

Who were the worshippers of Siva and Parvatı at Penjikent? There are dif-ferent possible interpretations, all of them, however, more or less conjectural.We feel rather sceptical with regard to the perspective suggested by Mode(1992: 330), namely that Siva and Parvatı are nothing but a disguised form ofthe local deities Wesparkar and Nana. From what we have remarked above, thelatter does not seem a sound hypothesis, so much so that, as we have alreadyunderlined, a shrine containing a “canonic” sculpture of Nana was situated at adistance of a few metres from the chapel containing the Umamahesvara. Inspite of the relationship we have supposed might have existed between the twochapels, the gods that they housed were likely perceived as quite distinct andprobably venerated by different people. That is why we consider more attrac-tive the hypothesis that the sculpture at issue bears witness to the presence ofSiva worshippers in the Sogdian town (Skoda 1992: 327; Marshak 1995-96:305), although it leaves us in the uncertainty of what exactly should we meanwith “Siva worshippers”. In other words, do we have to think of Indian immi-grants or of Indianized Sogdians?

The contacts between Sogdians and Indians seem to have been intense onthe tracks of the international trade routes leading to China. There is clear evi-dence of their commercial and cultural intercourse in the Tarim basin, in theSouthern oasis of which the establishement of Indian immigrants (most proba-bly from the North-Western region of the Subcontinent) dates at least from the2nd-3rd century AD41. To my knowledge, however, no unquestionable evidencehas so far been provided for the presence of an Indian community in the Sog-dian motherland and, in our specific case, at Penjikent, although it cannot beexcluded. On the contrary, we have extensive epigraphical material in the Up-per Indus valley attesting to an active frequentation of this region by Sogdianmerchants. Out of the 650 Sogdian graffitis yielded by the whole area investi-gated so far, about 550 have been found in the site of the Shatial Bridge, on theSouthern bank of the Indus. Some inscriptions date from the 3rd-4th century,but the majority of them belongs to the 5th-6th century. Such a remarkable con-centration has raised different hypothesis42. The view that Shatial might havebeen the terminal – rather than a halting point – of one of the main trade routesleading the Sogdian merchants towards India seems to us preferable. In otherwords, we cannot be sure that the Shatial inscriptions «attestent de l’impor-tance de la présence sogdienne en Inde dès le IIIe siècle» (de la Vaissière 2002:87), at any rate not much beyond that bridge on the Indus river.

41 VOROB′EVA-DESJATOVSKAJA 1984; CANNATA 2000: 44 ff.; DE LA VAISSIÈRE 2002: 82-83.42 See SIMS-WILLIAMS 1996; FUSSMAN - KÖNIG 1997: 62-106; DE LA VAISSIÈRE 2002: 85-89.

[31] The Umamahesvara in Central Asian Art 79

Nevertheless, it would be uncautious to think that India was out of Sog-dian traders’ reach. Nothing prevents us from surmising that they could havetravelled along the route that, passing through the western Hindu Kush, led tothe North West of the Subcontinent via Bamiyan and Kapisi, that is along theitinerary which from the mid-6th century replaced the Karakorum route(Kuwayama 1987: 717-722, and forthcoming). The case of the Buddhist monkKang Seng hui, who was born in China, in the beginning of the 3rd century AD,from a Kang (Sogdian) family which had lived for generations in India(Tianzhu)43, shows that the establishment of Sogdians in India might havestarted as early as the first centuries AD, and there is some evidence, howeverslim, that they pushed on as far as Sri Lanka, or even farther east along the mar-itime trade routes (Grenet 1996).

However, it is reasonable to think that their range of action must havemainly concentrated in the North-Western regions of India44. Accustomed tothe coexistence of a mixed population, devoted to Buddhism for centuries and,since the 6th-7th century (if not earlier), theatre of a significant spread of theHindu faith, with which Buddhism apparently came to terms, this area wasprobably the side of India most familiar to Sogdian traders. There, we mayguess, they did their business and even settled, as long as they found it prof-itable, and got married. Mixed marriages are probably one of the keys for ex-plaining how Sogdia became acquainted with a Hindu devotional cult which, aswe have seen, gained popularity from Kashmir to the Hindu Kush in the 6th-8th

centuries and which, for its symbolic realm, must have been particularly dear towomen45.

Whatever the dynamics of their spread outside India, it seems that Sivaand Parvatı perfectly fitted for a Buddhist frame, preserving their own signifi-cance, as the evidence from the Tarim basin shows. On account of its icono-graphic peculiarities, in particular the cup held by the goddess as well as herunusual position, the Khotanese couples (figs. 6, 7) seem to betray a specific re-ligious bias46. At Kyzyl, on the other hand, the Umamahesvara occurs twice (inthe Maya Cave and in the Gorge Cave) in association with the Conversion ofAt

˙avika, the children-devourer yaks

˙a. In the literary version of this episode

43 The source is the Gao seng zhuan, a Chinese collection of biographies (first half of the 6th cen-tury): DE LA VAISSIÈRE 2002: 77-80; see also GRENET 1996: 65-66.

44 On the Sud (also Suda, Sudgi), supposed descendants of Sogdian (or Central Asian) traders set-tled in the Panjab, see BAGCHI 1955: 146, and GRENET 1996: 65, fn. 4.

45 We remind, in this regard, that the Skandapuran˙

a (Avantikhan˙

a, 9.48) refers to small goldenUmamahesvara images to be worn by ladies, as bearers of fortune and wealth (JOSHI 1989: 49).

46 J. Williams (1973: 144-145) links these icons to Tantric (either Buddhist or Hindu) ritual prac-tices including madhu (liquor) and maithuna (sexual union), whereas Kumar (1975: 117), who is inclinedto a Vajrayanic interpretation of the Khotanese representations of Mahesvara, thinks of a proto-yab-yumiconography.

80 Ciro Lo Muzio [32]

there is no mention of the Sivaite couple47, nor do we find it in the Gandharanreliefs depicting this episode of Buddha’s life. In the Kyzyl paintings theUmamahesvara is to be seen in the place which, in the Gandharan reliefs show-ing the same subject, is sometimes reserved to the king and the queen of At

˙avi,

that is to the right of the Buddha (Foucher 1905-51: figs. 252, 253). On the op-posite side, At

˙avika is shown as he offers the king’s son to the Buddha. The

royal couple of At˙avi, who thanks to the intervention of the Buddha gets back

its child, is here replaced and sublimated by the divine couple embodying theprotection of family and progeny.

The clue provided by the Xinjiang murals leads us to suggests that theUmamahesvara should be probably interpreted as one of the main forms takenby the worship of the tutelary couple in Buddhist Central Asia. The cult ad-dressed to Pañcika and Haritı as well as to Pharro and Ardokhso was deeplyrooted in North-Western India, as witnessed in Gandharan art48, but we alsomust keep in mind that the “pattern” of the divine couple extended well be-yond Gandhara and the Buddhist domain, that it was part of a heritage com-mon to Central Asia since the Kushan epoch as late as the early medieval period(Bussagli 1984), as we can infer from the relevance of divine couples in the laterpainting of Sogdia. We can affirm that, as Bussagli suspected with regard to theDilberjin painting (loc. cit.), Siva and Parvatı adjusted themselves in that “pat-tern”. In the Tapa Skandar Umamahesvara, Skanda is not accompanied by hisvahana, the peacock, and holds none of his usual emblems (a cock or a spear).Siva’s son is here simply shown as a naked child, with bracelets and anklets,clinging to Parvatı’s dress. In other words he appears just as anyone of the chil-dren that usually gather around the Gandharan Haritı49. It was in the North-West of the Indian Subcontinent, in a land bound to be on the fringes of ortho-doxy, that Siva and Parvatı crossed the border of Brahmanism.

CIRO LO MUZIO

REFERENCES

ABDULLAEV, K. et alii (eds.) (1991) Kul’tura i iskusstvo drevnego Uzbekistana(Katalog vystavki), 2 vols., Moskva

47 An extensive account of the literary references concerning the Conversion of At˙avika is found

in SANTORO 1988: 226 ff. See also FOUCHER 1905-51: 507-512.48 FOUCHER 1905-51: I, 142-155.49 The iconographic and religious affinity link between this sculpture and the pattern of the tute-

lary couple was already suggested by TADDEI (1973: 212), while KUWAYAMA (1976: 395) stressed the com-positive similarities with the representations of Pañcika and Haritı.

[33] The Umamahesvara in Central Asian Art 81

AGRAWALA, R.C. (1956) Two Interesting Saiva Terracottas in the Bikaner Mu-seum. Artibus Asiae, XIX, 61-65

AGRAWALA, V.S. (1960) The Religious Significance of the Gupta Terracottasfrom Rang Mahal. Lalit Kala, 8, 63-71

AGRAWALA, R.C. (1995) Skanda-S˙as˙t˙hı in Gandharan Reliefs - A Review. East

and West, XLV, 1-4, 329-332AGRAWALA, R.C. - TADDEI, M. (1966) An Interesting Relief from the Swat Val-

ley. East and West, XVI, 1-2, 82-88AZARPAY, G. (1981) Sogdian Painting, Berkeley - Los Angeles - LondonBAGCHI, P. Ch. (1955) India and Central Asia, CalcuttaBANERJEA, J.N. (1956) The Development of Hindu Iconography, CalcuttaBANERJEE, P. (1969) A Siva Icon from Piandjikent. Artibus Asiae, XXXI, 1,

73-80BANERJI, R.D. (1913-14) Some Sculptures from Kosam. ASIAR 1913-14,

262-264BARRETT, D. (1957) Sculptures of the Shahi Period. Oriental Art, N.S., III,

54-59BELENICKIJ, A.M. (1973) Monumental´noe iskusstvo Pendikenta, MoskvaBERNARD, P. (1981) Un nouveau document sur les cultes hindouistes d’Asie

Centrale. Studia Iranica, 10, 2, 325-339BERNARD, P. (1987) Review of Kruglikova 1986. Abstracta Iranica, 1987,

60-62BERNARD, P. - H.P. FRANCFORT (1979) Nouvelles découvertes dans la Bactriane

afghane. Annali dell’Istituto Universitario Orientale di Napoli, N.S. XXIX,1, 119-148

BUSSAGLI, M. (1949) Due statuette di Maitreya. Nota sul simbolismo lunare nelBuddhismo. Annali Lateranensi, XIII, 355-390

BUSSAGLI, M. (1962) Note sulla divinità recante i simboli del Sole e della Lunascoperta a Pjandzikent nel 1954. Rivista degli studi orientali, XXXVII, 1-2,91-103

BUSSAGLI, M. (1984) Variabilità di valori nel «sistema» delle «coppie» in epocakus

˙an

˙a in Asia Centrale. In: L. Lanciotti (ed.), Incontro di religioni in Asia

tra il III e il X secolo d.C., Firenze, 115-146CANNATA, P. (2000) Sulle relazioni tra India e Asia interna nelle testimonianze

cinesi (Rivista degli studi orientali, LXXIII, Suppl. 1), RomaCHUVIN, P. (2002) (ed.) Le arti in Asia Centrale, Milano (1st ed. Paris

1999)COLLINS, Ch. D. (1988) The Iconography and Ritual of Siva at Elephanta, New

YorkCRIBB, J. (1997) Shiva Images on Kushan and Kushano-Sasanian Coins. In:

K.Tanabe, J.Cribb, H.Wang (eds.), Studies in Silk Road Coins and Culture.

82 Ciro Lo Muzio [34]

Papers in Honour of Professor Ikuo Hirayama on his 65th Birthday, Cam-bridge: 11-66

FITZSIMMONS, T. (1996) Chronological Problems at the Temple of the Dioscuri,Dil’berdzin Tepe (North Afghanistan). East and West, XLVI, 3-4,271-298

FOUCHER, A. (1905-51) L’art gréco-bouddhique du Gandhara, 3 vols., Paris-Hanoi

FUSSMAN, G. (1978) Chroniques des études kouchanes (1975-1977). JournalAsiatiques, 419-436

FUSSMAN, G. (1987) Chroniques des Études Kouchanes (1978-1987). JournalAsiatique, CCLXXV, 3-4, 333-400

FUSSMAN, G. (1989) The Mat˙

devakula: A New Approach to Its Understanding.In: D.M. Srinivasan (ed.) Mathura. The Cultural Heritage, New Delhi,193-199

FUSSMAN, G. - D., KÖNIG (1997) Die Felsbildstation Shatial (Materialen zur Ar-chäologie der Nordgebiete Pakistans, 2), Mainz

GADZIEV, M.S. - A.A. KUDRJAVCEV (2001) Steinmetzzeichen des 6. Jahrhundertsn. Chr. in Darband. Archäologische Mitteilungen aus Iran und Turan, 33,357-390

GAULIER, S., R. JERA-BEZARD, M. MAILLARD (1976) Buddhism in Afghanistan andCentral Asia (Iconography of Religions, XIII, 14), 2 pts., Leiden

GNOLI, Gh. (1963) The Tyche and the Dioscuri in Ancient Sculptures from theValley of Swat. East and West, XIV, 1-2, 29-37

GÖBL, R. (1984) System und Chronologie der Münzprägung des Kusanreiches,Wien

GOETZ, H. (1965) A Kashmiri Lingam of the 10th Century. Artibus Asiae,XXVII, 3, 275-279

GRENET, F. (1987) L’Athéna de Dil’berdzin. In: F. Grenet (sous la direction de)Cultes et monuments religieux dans l’Asie Centrale préislamique, Paris,41-45

GRENET, F. (1994) The second of three encounters between Zoroastrianism andHinduism: plastic influences in Bactria and Sogdia, 2nd-8th c. AD. In: G.Lazard, D.R. SarDesai (eds.) James Darmesteter Memorial Lectures (Jour-nal of the Asiatic Society of Bombay, 69), Bombay, 41-57

GRENET, F. (1996) Les marchands sogdiens dans les mers du Sud à l’époquepréislamique. In: Inde-Asie Centrale. Routes du commerce et des idées(Cahiers d’Asie Centrale, 1-2), Tachkent – Aix-en-Provence, 65-84

GRÜNWEDEL, A. (1912) Altbuddhistische Kultstätten in Chinesisch Turkistan,Berlin

GRÜNWEDEL, A. (1920) Alt-Kutscha, BerlinGUPTA, P.L. - D.C. SIRCAR (1973) Uma-mahesvara image inscription from Skan-

[35] The Umamahesvara in Central Asian Art 83

dar (Afghanistan). Journal of Ancient Indian History, VI, Parts 1-2 (1972-73), 1-4

HÄRTEL, H. (1989) A Siva Relief from Gandhara. In: K. Frifelt, P. Sørensen(eds.), South Asian Archaeology 1985, Copenhagen – London, 392-396

HOWARD, A.F. (1991) In Support of a New Chronology for the Kizil MuralPaintings. Archives of Asian Art, XLIV, 68-83

HUMBACH, H. (1975) Vayu, Siva und der Spiritus Vivens im ostiranischenSynkretismus. Acta Iranica. Monumentum H.S. Nyberg, I, Leiden, 397-409

INVERNIZZI, A. (1968-69) A relief in the Style of the Gandhara School fromChoche. Mesopotamia, III-IV, Torino, 145-158

JOSHI, N.P. (1989) Brahmanical Sculptures in the State Museum, Lucknow. PartII, I (Siva and Visnu), Lucknow

KLIMBURG-SALTER, D. (1989) The Kingdom of Bamiyan. Buddhist Art and Cul-ture of the Hindu Kush, Naples - Rome

KLIMKEIT, H.J. (1983) The Sun and Moon as Gods in Central Asian Texts.South Asian Religious Art Studies Bulletin, 2, Reading, 11-23

KOSELENKO G.A. (1966) Unikal´naja vaza iz Merva. Vestnik Drevnej Istorii, 1,92-105

KRAMRISCH, S. (1981) The Manifestations of Siva, PhiladelphiaKREISEL, G. (1986) Die Siva-Bildwerke der Mathura-Kunst. Ein Beitrage zur früh-

hinduistischen Ikonographie (Monographien zur Indien Archäologie,Kunst und Philologie, 5), Stuttgart

KRUGLIKOVA, I.T. (1974) Dil′berdzin I. Raskopki 1970-1972 gg., MoskvaKRUGLIKOVA, I.T. (1976) (ed.) Drevnjaja Baktrija. Materialy sovetsko-afganskoj

ekspedicii 1969-1973 gg., I, MoskvaKRUGLIKOVA, I. (1977) Les fouilles de la Mission Archéologique soviéto-afghane

sur le site gréco-kushan de Dilberdjin en Bactriane (Afghanistan). ComptesRendus de l’Académie des Inscriptions & Belles-Lettres, 407-427

KRUGLIKOVA, I.T. (1986) Chram Dioskurov, MoskvaKUMAR, K. (1975) A Dhyana-Yoga Mahesamurti, and Some Reflections on the

Iconography of the Mahesamurti-Images. Artibus Asiae, XXXVII, 1/2,105-120

KUWAYAMA, Sh. (1976) The Turki Sahis and Relevant Brahmanical Sculpturesin Afghanistan. East and West, XXVI, 3-4, 375-407

KUWAYAMA, Sh. (1987) Literary Evidence for Dating the Colossi in Bamiyan. In:G. Gnoli, L. Lanciotti (eds.), Orientalia Iosephi Tucci Memoriaes Dicata, 2,pp. 703-727

KUWAYAMA, Sh. (forthcoming) Two Itineraries Concerning the Decline of

84 Ciro Lo Muzio [36]

Gandhara: Chinese Evidence. The Buddhism of Gandhara: An Inter-disciplinary Approach, Toronto

LO MUZIO, C. (1995-96) OHIcO: a Sovereign God. Silk Road Art and Archaeol-ogy, IV, 161-174

LO MUZIO, C. (1999) The Dioscuri at Dilberjin (Northern Afghanistan): Re-viewing their Chronology and Significance. Studia Iranica, 28/1, pp.41-71

MARSHAK. B.I. (1995-96) On the Iconography of the Ossuaries from Biya-Naiman. Silk Road Art and Archaeology, IV, 299-321

MARSHAK, B. A. (1996) New Discoveries in Pendjikent and a Problem of Com-parative Study of Sasanian and Sogdian Art. In: La Persia e l’Asia Centraleda Alessandro al X secolo (Atti dei Convegni Lincei 127), Roma,425-438

MARSAK, B.I. - V.I. RASPOPOVA - V.G. SKODA (1993) Novye issledovanija sogdij-skoj kul’tury v Pendzikente. Archeologiceskie Vesti, II, 91-102

MAXWELL, T.S. (1983) Sun/Moon and Aditya/Rudra Oppositions in Visvarupa.South Asian Religious Art Studies Bulletin, 2, 41-55

MODE, M. (1992) Zum sogdischen Mahadeva. Archäologische Mitteilungen ausIran, n.s., XXV, 329-331

Oxus (1993) Oxus. Tesori dell’Asia Centrale, RomaPAUL, P.G. (1986) Early Sculpture of Kashmir, Doctoral Thesis, University of

LeidenRAO, T.A. GOPINATHA (1968) Elements of Hindu Iconography, New York (1st

ed. Madras 1914)ROSENFIELD, J.M. (1967) The Dynastic Arts of the Kushans, Berkeley-Los

AngelesROWAN, D.P. (1985) Reconsideration of an Unusual Ivory Diptych. Artibus

Asiae, XLVI, 4, 251-304RTVELADZE, E.V. ET ALII (EDS.) (1991) Drevnosti juznogo Uzbekistana (Antiqui-

ties of Southern Uzbekistan), SokaSANTORO, A. (1988) Note di iconografia gandharica, III. A proposito della con-

versione di At˙avika. Annali dell’Istituto Universitario Orientale, n.s.,

XLVIII, 219-242SCHLUMBERGER, D., LE BERRE, M. & FUSSMAN, G. (1983) Surkh Kotal en Bactri-

ane, I. Les temples: architecture, sculpture, inscriptions (MDAFA, XXIV), 2vols., Paris

Sérinde (1996) Sérinde, Terre de Bouddha, ParisSHERRIER, J. (1993) Siva in Gandhara. In: A. Gail, G.J.R. Mevissen (eds.), South

Asian Archaeology 1991, Stuttgart, 617-624SIMS-WILLIAMS, N. (1996) The Sogdian Merchants in China and India.

[37] The Umamahesvara in Central Asian Art 85

In: A. Cadonna, L. Lanciotti (eds.), Cina e Iran. Da Alessandro Magnoalla dinastia Tang, Firenze, pp. 45-67

SIMS-WILLIAMS, N. - J. CRIBB (1995-96) A New Bactrian Inscription of Kanishkathe Great. Silk Road Art and Archaeology, 4, 75-142

SKODA, V.G. (1992) Ein Siva-Heiligtum in Pendzikent. Archäologische Mit-teilungen aus Iran, n.s., XXV, 319-327

STAVISKIJ, B.Ja. (1973) The Capitals of Ancient Bactria. East and West, 23,265-277

STEIN, M.A. (1907) Ancient Khotan. Detailed Report of Archaeological Explo-rations in Chinese Turkestan, 2 vols. Oxford

TADDEI, M. (1964-65) A Linga-shaped Portable Sanctuary of the Shahi Period.East and West, XV, 24-25

TADDEI, M. (1971) On the Siva Image from Kuh˘ah, Mesopotamia. Annali del-

l’Istituto Universitario Orientale, 31, 548-52TADDEI, M. (1973) The Mahisamardini image from Tapa Sardar, Ghazni. In: N.

Hammond (ed.), South Asian Archaeology, London, 203-213TADDEI, M. (1985) A New Early Saiva Image from Gandhara. In: J. Schots-

mans, M. Taddei (eds.), South Asian Archaeology 1983, Naples, II,615-628

TADDEI, M. (1996) Bhiks˙at˙anamurti or Harihara? East and West, 46, 3-4,

453-456TANABE, K. (ed.) (1986) Gandharan Ladies & Toilet-Trays from Japanese Collec-

tion (Tokyo, The Ancient Orient Museum, March 15th - May 6th),Tokyo

TANABE, K. (1991-92) OHIcO: Another Kushan Wind God. Silk Road Art andArchaeology, 2, 51-71

DE LA VAISSIÈRE, É. (2002) Histoire des Marchands Sogdiens (Bibliothèque del’Institut des Hautes Études Chinoises, XXXII), Paris

VERARDI, G. (1982) Osservazioni sulle sculture in argilla e su alcuni ambientidei complessi templari I e II di Pendzikent. Annali dell’Istituto Universi-tario Orientale, 42, 247-304

VOROB′EVA-DESJATOVSKAJA, M. I. (1984) Indijcy v Vostocnom Turkestane vdrevnosti (nekotorye sociologiceskie aspekty). In: B.A. Litvinskij (ed.),Vostocnyj Turkestan i Srednjaja Azija. Istorija. Kul´tura. Svjazi, Moskva:61-96

WHITFIELD, R. (1984) The Art of Central Asia. The Stein Collection in the BritishMuseum, III, Tokyo

WILLIAMS, J. (1973) The Iconography of Khotanese Painting. East and West,XXIII, 109-154

YALDIZ, M. (1987) Archäologie und Kunstgeschichte Chinesisch-Zentralasien(Xinjiang), Leiden

86 Ciro Lo Muzio [38]

YALDIZ, M. et alii (s.d.) Magische Götterwelten. Werke aus dem Museum für In-dische Kunst, Berlin, Berlin

ZEYMAL, E.V. (1997), Visha-Shiva in the Kushan Pantheon. In: R. Allchin et alii(eds.), Gandharan Art in Context. East-West Exchanges at the Crossroads ofAsia, New Delhi, 245-266

ZIMMER, H. (1974) Myths and Symbols in Indian Art and Civilization, ed. by J.Campbell, Princeton (1st ed. 1946)