Minerals and Metals in Ancient India v1.c28.s8.t5: The Icon-Metallurgy of Tibet

Minerals and Metals in Ancient India v1.c28.s8.t5

Language: English

The Icon-Metallurgy of Tibet

v1.c28.s8.t5.1

We have mentioned the Magadha tradition (along with Gāndhāra and Mathurā traditions) influencing the brass-icon metallurgy of Kashmir (*vide* Table 20.7). The Pāla tradition spreaded across Nepal to Tibet starting from the beginning of the ninth century, and therefore we may consider the Tibetan scenario, at least up to the twelfth century AD, as part of the scenario of Greater India. It is true that the Newar sculptors from Nepal and the brass-casters of Kashmir helped making the Tibetan icons during the same period (Lo Bue, 1981a: 33-34; 1981b: 69-70), but the Nepalese and Kashmiri traditions were also Indian.

v1.c28.s8.t5.2

Lo Bue (1981a: 34-37) made a beautiful analysis of the huge amount of earlier data and his own data on Tibetan icons and came up to certain valid generalisations. According to him the brass-traditions in the Western and Eastern Tibet were due to the influences from Kashmir and China respectively, whereas the Central and Southern Tibet were influenced more by the Nepalese tradition of using pure copper and gilded copper in the icons. We would like to point out that the Magadha/Pāla traditions in Bihar/Bengal were the more versatile ones encompassing the uses of copper, gilded copper, brass as well as bronze. We find very little bronze or tin in the Tibetan icons probably because the production of tin in the Hazārībāgh area (referred to as *vanga*, or the metal from Bengal, in the ninth century AD *Sarveśvararasāyana*) or the import of the metal from Burma or Malay Peninsula were insufficient so that Bengal and Bihar could hardly share it with Nepal or Tibet. On the other hand there were sufficient number of working copper mines in Nepal, and significant quantities of zinc and brass were forthcoming from the Zawar Mines in Rajasthan. What the Pāla kingdom transmitted to Tibet were principally the prevalent art traditions and the technical know-how of casting through the disciples of Dhīmān and Biṭpāla, apart from sending some copper and brass.

v1.c28.s8.t5.3

Quite significantly, “brass tends to replace bronze (in the ancient Indian context) as one proceeds westward from Bengal, an area close to tin-producing

v1.c28.s8.t5.4

Table 20.9

v1.c28.s8.t5.5

Atomic Absorption Analyses of Some Tibetan/Himalayan Brass Items in British Museum Dated 11th-12th Century AD

v1.c28.s8.t5.6

| Sl. No. | Item | Cu | Zn | Pb | Sn | As | Fe | Ni | Sb | Ag | Primary Reference |
| 1. | Cleveland Buddha 98.1 cm high — W. Tibet 10th-11th century AD | 68.3 | 20.2 | 11.0 | — | — | — | — | — | — | Lo Bue (1981a:34) |
| 2. | Standing Vajrapāṇi 69.3 cm. at Musée Guimet, Paris | 87.0 | 11.7 | 1.0 | — | — | — | — | — | — | Hours (1980) |
| 3. | Mañjuśrī (Western Tibet) 12.5 cm | 65.5 | 30.4 | 1.70 | 0.30 | 0.25 | 0.20 | 0.16 | 0.05 | 0.03 | Schroeder (1981) |
| 4. | Lion of item 3 | 76.0 | 13.9 | 1.00 | 1.00 | 6.80 | 0.30 | 0.04 | 0.07 | 0.20 | — do — |
| 5. | Base of item 3 | 90.1 | 0.4 | 0.26 | 1.20 | 7.40 | 0.15 | 0.02 | 0.07 | 0.23 | — do — |
| 6. | Stūpa | 62.5 | 32.3 | 1.56 | 1.50 | 0.75 | 0.50 | 0.30 | 0.07 | 0.02 | — do — |
| 7. | Buddha and Bodhisattva 10.5 cm Azurite Pigment (Western Tibet) | 77.5 | 14.7 | 4.90 | 2.60 | 0.20 | 0.50 | 0.07 | 0.20 | 0.03 | — do — |
| 8. | Avalokiteśvara — traces of gold but no Hg on face | 98.0 | 0.015 | 0.02 | 0.40 | 0.03 | 0.40 | 0.01 | 0.01 | 0.10 | — |

v1.c28.s8.t5.7

| Sl. No. | Item | Elemental Percentages | | | | | | | Primary Reference |
| | | Cu | Zn | Pb | Sn | As | Fe | Ni | Sb | Ag | |
| 9. | Avalokiteśvara — Eastern India — Set with green stones | 68.5 | 26.5 | 3.00 | 0.30 | 1.20 | 0.27 | 0.04 | 0.10 | 0.03 | |
| 10. | Śiva and Pārvatī — Eastern India — gold paint on face and blue azurite on hair | 74.0 | 25.0 | 0.15 | 0.10 | 0.35 | — | 0.10 | 0.30 | 0.10 | Béguin (1977) |
| 11. | Avalokiteśvara — Eastern India — inlaid with Ag and Cu | 70.5 | 28.5 | 0.50 | 0.30 | 0.35 | 0.03 | 0.04 | 0.20 | 0.04 | — do — |
| 12. | Mañjuśrī — Eastern India or S. Tibet — inlaid with Ag and Cu; inset green stones | 72.0 | 27.0 | 0.66 | 0.10 | 0.20 | 0.20 | 0.04 | 0.10 | 0.01 | — do — |
| 13. | Samvara — E. India or S. Tibet — inlaid Ag and Cu | 71.5 | 27.0 | 0.86 | 0.10 | 0.20 | 0.18 | 0.03 | 0.10 | 0.02 | — do — |

v1.c28.s8.t5.8

countries, to Rajasthan, where zinc ores were exploited”. (Lo Bue, 1981a: 34) The above generalisation clearly emerges from the earlier works of Lal (1956), Lee (1967), Werner (1972), Sahai (1977), Uhlig (1979) etc. A survey of Tibetan and Himalayan metal statuary in the major public collections of Britain such as Ashmolean, Liverpool County, Royal Scottish and Cambridge Museums showed that out of nearly 220 Tibetan icons, 58 per cent were made of brass and 42 per cent of copper. Some were fire-gilded brass (15 per cent), fire-gilded copper (12.2 per cent), cold-gilded (2 per cent), silver-containing (5 per cent), repoussé (2 per cent) and inlaid with stones (1.5 per cent). Lo Bue (1981a: 36) The trend is also clear from the compilation of data on 121 samples (Craddock, 1981), a gist of which is displayed in Table 20.9.

v1.c28.s8.t5.9

The Newar artistes of Nepal often preferred pure copper for casting (brass was of course known in Nepal; during the second half of eighth century AD, one Nepalese king was known as ‘king of brass’). The preference for copper in Nepal was for (a) its attractive red colour, (b) the prevalent religio-aesthetic tradition connecting it to purity, (c) its advantages for mercury gilding. (Lo Bue, 1981a: 48) In the seventh century AD, Hiuen-Tsang and the text Cakradatta of 1050 AD specifically referred to Nepalese copper; the late thirteenth century AD text Rasaratnasamuccaya mentioned about its purity. However considering the trace impurities of the sample No. 8 in Table 20.9 it does not seem that highly pure native copper was used. Some copper might have been obtained from the Kamraz mines in Kashmir as mentioned in Kalhaṇa’s Rājatarangiṇī.

v1.c28.s8.t5.10

The bulk of the samples reported or analysed by Craddock (1981) are made of brass. In Table 20.9 we have not included the Tibetan icons of date later than the twelfth century AD, when the percentage of zinc in the brass went on increasing even beyond 32 per cent (vide sample No. 6). In several samples such as in No. 5 a very small quantity of zinc might have been added not for alloying but probably as a de-oxidiser. It is doubtful whether zinc ore mines in Tibet or Nepal were tapped; the metal and brass were probably imported from India.

v1.c28.s8.t5.11

Some of the terms for zinc ore, zinc and brass in the ancient world were:

v1.c28.s8.t5.12

Persian: tūtiyā (calamine), tuzia or tutty (impure zinc oxide), tutenāga (zinc)

v1.c28.s8.t5.13

Chinese: tutenāga (zinc), teou-shih (brass or copper alloy in general, the term used by Hiuen-Tsang)

v1.c28.s8.t5.14

Tibetan: ti-tsha (zinc), tikzā (zinc mineral), red li (copper), white li (high zinc brass)

v1.c28.s8.t5.15

Lo Bue (1981a: 47-48, 50-51 and 65-66) explained in great detail that li was a general term in Tibet for copper alloys. White *li* or high (around 25 %) zinc brass was the constituent of many icons such as No. 10 with strong Magadha influence. He quoted *Padma-dkar-po* (1973 reprint of *Collected Works*, Darjeeling) writing on “images of Magadha made out of white *li*, of indisputable quality with plump face and a great share of grace, with perfect copper lips and silver eyes”. Béguin (1977: 11-12 and 70) indentified No. 10 as a replica of an original Indian Umā-Maheśvara icon. It is quite possible that the Pāla artistes introduced in the ninth century AD. Tibet this tradition of brass icon (24-28 % zinc, rest copper) inlaid with copper and silver. Samples such as Nos. 2, 4 and 7 contain 11-15 per cent zinc and may be called, in Tibetan terminology, yellow *li*; the bulk of the icons however contained 25-32 per cent zinc (white *li*). Still higher percentages of zinc are noted in the icons produced after the twelfth century AD e.g. sixteenth century AD. Maitreya analysing 40.7 per cent zinc. (Craddock, 1981: 29, sample 98)

v1.c28.s8.t5.16

Tin percentages were well below one in most cases, where a little scrap bronze might have been used. *Rasaratnasamuccaya* recipe of ‘eight parts of copper and two parts tin’ i.e. 20 per cent tin bronze could not be used in Nepal or Tibet for the scarcity of tin. Significantly two bells from ninteteenth century Tibet (Craddock, 1981: 27-28, Nos. 47 and 80) assay 20.8 and 19.8 per cent tin respectively.

v1.c28.s8.t5.17

Lead is a low melting point metal insoluble in copper. Up to 2 per cent lead in molten copper improves fluidity and thus it clearly aids in casting. At higher percentages the globules form ‘lakes’ of lead and the lead-copper/bronze interface causes weakness and cracking if hammered. Most Tibetan images contain only moderate quantities of lead which is at variance from some of the Kashmir and North Indian examples. Cleveland Buddha (sample No. 1) contains 11.0 per cent lead. The brass statue of Sūrya from Kashmir (Lee, 1967) contains 2.75 per cent lead, 18.7 per cent zinc and 78.1 per cent copper. Uhlig (1979: 122) has described a brass Buddha containing 10. 37 per cent lead, 13.2 per cent zinc and 73.1 per cent copper. The Newar sculptors preferred far less quantities of lead which are soluable in mercury and thus hinder the process of gilding using mercury. It is noteworthy that many fire-gilded brass statues and copper icons contain negligible quantitites of lead (*vide* Nos. 5, 8, 10-13 in Table 20.9).

v1.c28.s8.t5.18

Tibetan icons contain detectable quantities of arsenic. Some weathered *fahlerz* ores of copper, containing not only arsenic but also some antimony, nickel and silver, might have been used for smelting. (Tylecote, 1962) The items 4 and 5 (Table 20.9) viz., the lion and the base of the brass Mañjuśrī however contain higher proportions of arsenic and much less of the other trace elements and therefore suggest deliberate arsenic-alloying. This 12.5 cm icon has been described by Schroeder (1981), and elemental analysis for the icon, lion and base published by Craddock (1981: 27, No. 42). La Niece *et al.* (1989) showed that like the third century BC Libyan ‘white copper’ coins, the base of the Mañjuśrī statuette (No. 5 in Table 20.9) contains high per cent of arsenic (7.4) in the bulk and even higher (14.0) per cent of the element on the surface. X-ray fluorescence and scanning electron microscopoy with EDAX analysis revealed that the arsenic-rich layer was formed due to inverse segregation which occurs in alloys freezing over a long temperature range and contracting during solidification. (La Niece *et al.*, 1989: 12) Evidently the arsenical casting alloy was chosen for the visual effect of a ‘golden’ statue on a white ‘silvery’ base. The eleventh century text *Rasahṛdaya* (18.23-69) actually described the process of “heating of copper with an equal weight of *tāla* (orpiment, As₂S₃) and *vaṅga* (tin) to convent this into silver”.

v1.c28.s8.t5.19

Considering that many copper and brass/bronze samples from India (such as the Nālandā samples) are high in iron content, it is surprising that the Tibetan icons are so low in iron content (*vide* Table 20.9). Out of 121 samples analysed by Craddock (1981: 26-30), only six contain more than 1 per cent iron. Possibly this indicates that the source of copper was Nepalese which involved a not-too-high smelting temperature precluding iron contamination.

v1.c28.s8.t5.20

The icons normally contained 100-400 ppm silver. Higher proportions of silver (items 4 and 5 in Table 20.9) were associated with high proportions of arsenic and might have been derived from *fahlerz* ore as explained before. Craddock (1981: 20) reported that some icons contain 0.1-0.4 per cent gold in the bulk, unrelated to gilding on the surface; thus specific efforts might have been made to employ the commonly processed placer deposit of Tibetan gold in alloying. The major use of gold was however in the process of gilding.

v1.c28.s8.t5.21

It is well-known that gilding can be done by one of the three methods: (1) gold-mercury amalgam is prepared and applied over the baser metal, and then the object is heated, (2) gold-leaf is pressed on the amalgamated surface and then the object is heated, (3) gold may be applied as a paint with a binder (‘cold gilding’). The first two methods known as ‘fire-gilding’ involve the use of mercury, a trace of which remains on the surface. X-ray fluorescence spectroscopy and emission spectrography have been utilised to detect gold and mercury on the surface. Craddock (1981: 26-30) found that out of 121 Tibetan icons/samples, 45 had been fire-gilt, but none of the 45 are dated earlier than the fourteenth century. The art of gilding was transmitted from India to Sri Lanka and Nepal around ninth-tenth century and reached Tibet a few centuries later.

v1.c28.s8.t5.22

The earliest textual reference to gilding in India is to be found in Kauṭilya’s *Arthaśāstra* (2.13.44-47) in which *tvaṣṭrkarmaṇaḥ* (gold-plating) on copper is recommended through the use of *vālukāhiṅgulukasya rasa* (the juice of cinnabar or mercury). We have mentioned about the first century BC thickly gilt inscribed brass urn from Wardak, Kabul.

v1.c28.s8.t5.23

The practice of gilding in India was firmly established around the Gupta Period. The 84 cm tall standing Mañjuśrī from Mahasthan, Bogra district, Bangladesh was described by Majumdar (1926) and Saraswati (1962) as gold-plated, probably fire-gilt and ‘of definitely Gupta workmanship’. We have earlier mentioned the Buddhist bronze images from sixth-seventh century AD Sirpur, Madhya Pradesh, one of which was gilt.

v1.c28.s8.t5.24

Ray (1956: 115) referred to the sixth century AD Kubjikā Tantra manuscript found in the collections of the Maharaja of Nepal referring to ‘the transmutation of copper into gold with the aid of mercury’ which Lo Bue (1981b: 82) interprets as an allusion to the fire-gilding process. Quoting Subbarayappa (1971: 309-38), Oddy et al. (1981: 89-90) also supports the idea of linking Indian alchemy, initiated during the sixth-seventh century AD, with the tradition of fire-gilding or ‘converting the base metal into gold using mercury’.

v1.c28.s8.t5.25

Several Himalayan icons analysed by Uhlig (1980) and two Pāla period (AD 750-1150 AD) figures reported by Schroeder (1981) have been found to be fire-gilt. Oddy et al. (1981: 88) have quoted T.G. Padley of the British Museum who undertook X-ray fluorescence spectrometry and emission spectrography work and detected mercury on the gilt surfaces of the following icons:

v1.c28.s8.t5.26

v1.c28.s8.t5.27

The practice of fire-gilding became common and popular in Tibet during fourteenth century AD.

v1.c28.s8.t5.28

Theophilus knew in the twelfth century that gilding is better done on unalloyed copper rather than on brass or leaded brass. (Oddy et al., 1981: 92) The presence of lead and zinc on the surface causes ‘white spots resistant to even drying’, there being no such problem on copper surface. The Indian and Nepalese artistes must have known this phenomenon through practical experience. This is confirmed by Schroeder’s (1981) observation that of the ten Pāla period images, the eight ungilded ones were made of brass, bronze or leaded copper while the two gilded ones were cast of almost pure copper. Analysing pre-seventeenth century Himalayan images, Oddy *et al.* (1981: 94) observed that the fire-gilded icons mostly contain less than 5 per cent zinc, while there is a large group of ungilded objects with higher zinc content (12 to 32 per cent).

v1.c28.s8.t5.29

Nieuwenhuysen *et al.* (1984) have reported similar trend on the basis of their analysis of twenty-three Himalayan statues of the eighteenth-nineteenth century period kept in the Antwerp Museum. Nine of the 13 gilded statues examined contain less than 0.2 per cent lead; three gilded statues, containing more than 1 per cent lead, have lost most of their gold. Of the six ungilded statues examined, 4 contain more than 0.2 per cent lead and 5 contain more than 20 per cent zinc. Lo Bue (1981a: 39) aptly observes that the modern tradition in the Himalaya, resembling the ancient tradition of India and Nepal, seems to be that figures for gilding are cast in almost pure copper.

v1.c28.s8.t5.30

Lo Bue (1981a: 54-57, 59-61 and 67) has reviewed the occurrence of the mercury ore cinnabar, silver and gold in Tibet. Cinnabar desposits have been available in the central, eastern and south-eastern Tibet. There are reports of Tibetan castings of silver being presented to China in 824 AD. Though averse to sub-surface mining, the people of Ladakh and Tibet always recovered gold washings from river and other placer deposits ever since the era of Herodotus’s tale of ‘gold-digging ants’. When the Tibetan students used to visit the Buddhist monasteries in Bihar and invite the celebrated teachers like Atīśa Dīpaṅkara to visit Tibet, one of the commonest presents they offered was gold (dust or bar). Thus it seems that the Tibetans never lacked in the raw materials for gilding; what they received from the Nepalese and Indians was the necessary technology.

v1.c28.s8.t5.31

It is reported that even before the fourteenth century AD when fire-gilding was introduced in Tibet, the icons there used to be cold-gilded by the application of paints containing gold and binders. For example Zwalf (1981: 107) has reported that the Avalokiteśvara of twelfth century AD (No. 8 in Table 20.9) made of almost pure copper shows traces of gold but no mercury on the face; this suggests that gold-painting or cold-gilding without mercury might have been attempted (as in No. 10). Table 20.9 further shows the use of inlaid silver (eyes), copper (lips) and green stones in the icons (Nos. 9 and 11-13). Blue azurite pigment was used in several cases (Nos. 7 and 10). In some Tibetan icons of later dates (14th-17th century AD) red lead and cinnabar paints were noticed. (Oddy *et al.*, 1981: 100-101)

v1.c28.s8.t5.32

The Nepalese-Tibetan scenario has been presented by us in some detail for two reasons. The work at the British Museum (Occasional Paper No. 15 on Tibetan Metallurgy) is an outstanding example of the application of modern analytical facilities to the works of ancient art. Secondly, the data amply illustrate how the metal technology of ancient India moved across the international borders. Of course there was a reciprocal exchange of materials and ideas. India received during this era (up to 12th century AD) naipālī (Nepalese) copper, Tibetan gold and silver, some cinnabar and borax. Tibet also transmitted to India the Chinese knowledge of making cementation copper (precipitation of copper from mine-water by iron powder). We would deliberate on a reference to this process (made in Nalina’s Dhātuvāda) in the chapter 8 of our Vol. 2. Tibet rendered another singular service to India. During the onslaught of invaders starting in the tenth century AD many ancient texts and records were brought to Tibet for safe custody and translation. Thus today we are able to re-construct a part of the cultural and scientific heritage of ancient India for the period of seventh-twelfth century AD through the studies on Nepalese/Tibetan literatures and icons.

Linked texts (5)

Rājataraṅgiṇī — Kalhaṇa, Sanskrit
Kauṭilīya Arthaśāstra — Kauṭilya, Sanskrit
Cakradatta (Cikitsāsaṅgraha) — Cakrapāṇidatta, Sanskrit
Rasaratnasamuccaya — Vāgbhaṭa (alchemist), Sanskrit
Rasahṛdaya — Govinda, Sanskrit