First electronic reprint on the Internet © 1997 Volker ThewaltHomepage Volker Thewalt Verlag
Besides the petroglyphs of an earlier date and those of a more local character like hunting scenes, the rockcarvings and inscriptions of the Buddhist period form a main portion of the carvings in the Northern Areas of Pakistan.
Buddhist rockcarvings have been found mainly following the old routes of trade and pilgrimage along the Indus, the Ghizr (Gilgit) and the Hunza river (fig. 1). In the area of Chilas a tremendous concentration of Buddhist carvings and inscriptions has been discovered (fig. 2 shows the distribution of sites in Chilas area) and is being documented in subsequent campaigns [fn1] . Apart from the inscriptions written in Indian scripts (Kharosthi, Brahmi, Proto-Sarada), many others have been found to be in Sogdian, Bactrian, Tibetan, Chinese and even in Hebrew characters. These inscriptions, when read and translated fully , will help in dating and interpreting the connected rockcarvings.
The carvings themselves display a vast variety of Buddhist iconology and of the development of architectural forms and designs, especially connected with the stupa. The oldest representations of stupas very closely resemble the types in Central India (e.g., in Sanci and Bharhut) and some monumental stupas in Gandhara as well. Later on, the carvings of stupas reflect a development in stupa-architecture similar to that in Gandhara towards terraced plinths, stairways leading up to the raised pradaksinapatha and a tendency towards slender elongated forms.
Several Central-Asian types occur and even a pagoda-like structure was found during the 1982 campaign by Jettmar (fig. 3).
Many of these elaborate rockcarvings must be attributed to highly skilled craftsmen who received their artistic training in the great monasteries of Gandhara, while others are crude imitations, executed by travelling laymen or the inhabitants of the neighbouring villages, wishing to gather some spiritual merit by reproducing these sacred monuments.
Rockcarvings of the most archaic stupa-types have been discovered in Chilas II and Chilas III. Fig. 4 shows a stupa from Chilas III with a nearly quadrangular shape on a kind of a plinth or stand a votive stupa could be intended here [fn2] . The upper points of the quadrangular outline are roundish in shape, but this does not really resemble the dome of a stupa.
The next carving from Chilas III shows a regular stupa (fig. 5): the anda rests upon a cylindrical (?) drum and a flat plinth. harmika and umbrellas are crowning the structure as in the previous example; a garland decorates the dome of the monument. Quite a number of such simple stupas have been found at Chilas II as well.
A drawing from Chilas II (fig. 6) shows the next stage of the development [fn3]: the stupa raises on a rectangular plinth. It seems that corner-pilasters are indicated on the plinth. A staircase given in the simplified form of a ladder leads up to the platform; a doorway with a triangular pediment gives access to the circumambulatory platform (not to the stupa itself) , that is surrounded by a railing. The cylindrical drum and the dome of the stupa are markedly separated by a flat horizontal band; the dome is again decorated by a garland, here given in a zig-zag line. The harmika in the form of a small vedika and three umbrellas are topping the monument (Dani 1983: nos. 82, 84). To the left of the stupa two devotees are shown wearing long and heavy coats; and above them a kind of column is depicted with peculiar indications of a base (or floral leaves?) and a capital, the whole thing crowned by an animal, which cannot be determined precisely.
Such columns, standing by the side of a stupa, have been found rather often in Chilas II as well as in the later carvings from Thalpan Bridge (Jettmar 1980b: frontispiece; 1980c: fig. 2; 1982a: fig. 11 ; Thewalt 1984. fig. 3). Stylistically the column and the second person to the left are somewhat different from the main stupa and the tall devotee they might be additions by another person.
A highly advanced form of the stupa is seen in a famous carving from Chilas I (fig. 7), which was already made known to Sir A. Stein during his short visit to Chilas (Stein 1944. 20 f.). Here a big stupa together with two devotees is depicted beside a marvellous representation of the Vyaghrijataka (Thewalt 1983). The dome of the stupa rests on five storeys and a broad torus. It seems that the trapezoid element with sloping sides (similar to the umbrellas) was carved later over the fifth storey; the reason for that is unclear . The lowest storey is decorated by an alternation of beautiful floral scrolls (cf. here fig. 19) and simple pilasters with rectangular blocks indicating the bases and capitals. The second storey shows three tribolate niches; these niches are empty as is the niche decorating the anda; presumably the niches were left empty due to the rather small size of the niches and the fact that the carving was obviously executed by means of a coarse stone implement that did not allow for extremely fine lines. The stupa is crowned by a harmika in form of an inverted stepped pyramid and by nine umbrellas, and on top of them there is a crescent and a round disk representing the sun [fn4]; streamers in a stiff stylized form and small bells hang down from the top and from the lowermost umbrella.
Two more highly interesting features are to be noted in this rockcarving of a stupa presumably of the 6th or 7th century: the umbrellas are connected with each other by means of small oblique supporting elements and the lowermost umbrella is fixed with longer staffs to the dome, that has a special foothold for that purpose on its shoulders. This and other carvings showing the same details may give important hints at how to reconstruct actual votive or monumental stupas. The second unusual feature about this carving are the two figures of soldiers in long robes, standing on top of the anda besides the harmika, holding Iong spears in their hands. This is motif is uncommon in the art of Gandhara and is known till now only from much later wallpaintings in the Alchi monastery.
The next rockcarving from Chilas II shows a more simple form of the stupa (fig. 8). The stupa rests on three platforms of diminishing size and presumably quadrangular groundplan, carrying a round (?) drum that is smaller in diameter than the crowning dome of the stupa. Harmika, umbrellas and the other topping elements are given in the usual manner.
Another stupa from Thalpan Bridge (fig. 9), executed with a sharp metal chisel in very delicate lines, shows the same scheme of construction. The topping elements above the umbrellas are still more elaborate than in the previous examples; the streamers are shown flowing symmetrically in wavy lines from the top to both sides. The posts supporting the umbrellas are given here in a somewhat different manner: the lower posts resting on the anda show indications of bases and capitals, while the smaller ones between the umbrellas are indicated just by round dots, and in the same manner the shaft carrying the umbrellas is stylized. Is this just a decorative form of representation, or did the artist no longer understand the construction of such a monument ?
A very small-sized carving from Thalpan Bridge (fig. 10) shows three stupas on a common base that is decorated by trilobate niches. To each of the stupas leads a staircase. Even these miniature stupas show the beams supporting the umbrellas and the waving streamers. The carving is very finely done, obviously with a metal chisel too.
Not completely finished are three other stupas from Thalpan Bridge (fig. 11). The left stupa with a beautifully decorated plinth is lacking the usual streamers, as is the case with the smaller stupa in the middle. The stupa on the right shows only one streamer , but here the bells are missing and the decoration of the plinth is not finished. All three stupas are very similar in design. The plinth consists of 5-6 storeys, the anda shows a trilobate niche and above the dome rises the harmika in form of an inverted stepped pyramid; the umbrellas are supported by beams. While the umbrellas at Chilas I were of a trapezoid shape with pointed edges, here the edges of the umbrellas are rounded.
The next beautiful stupa is also from Thalpan Bridge (fig. 12). The three lower storeys are again decorated with floral ornaments. The dome shows three tribolate niches, two of them in a lateral view and the central one in frontal view . Of course these niches on the dome of the stupa do not indicate openings in the structure, but they resemble the false gables depicted in many of the stupas from Gandhara[fn5] . The front niche shows a Buddha, sitting crosslegged in dharmacakramudra, the gesture of preaching. In the two lateral niches the Buddha figures are hardly recognizable. The dome of the stupa is topped by the usual elements of harmika, umbrellas, bells and streamers, crescent and disk. Very similarly to the big stupa at Chilas I (fig. 7), three persons are shown standing on top of the anda but obviously these persons are not meant to be guards since they are not holding any weapons at all. They could represent figures of Buddhas or Bodhisattvas.
Two more stupas from Thalpan Bridge show a different and more simple design (fig. 13). The dome rests on two or three storeys. While these storeys are left blank in the case of the left building, the three storeys of the second stupa show decorations consisting of commonly known architectural elements: the first storey is decorated by trilobate niches of an unusual shape alternating with pilasters; the second storey has an arcade consisting of simple vaulted niches; and the third storey is decorated by a row of pilasters only. The harmika is missing in both carvings and the curvilinear umbrellas seem to consist of a wooden structure decorated with many bells very similar to the representation of a big stupa at Shatial (Jettmar 1980: 6; 1980d: pl. 3. Dani 1983. no. 54).
A simplified stupa from Chilas I (fig. 14) consisting of a triple plinth, the anda and five umbrellas (Stein 1944. 21 pl. Vb) depicts a type that has been repeated over and over again along the ancient routes by those travellers who had no artistic training but just depicted simple imitations of the more sophisticated prototypes. The rockcarvings of this kind are all done with simple stone implements, leaving aside all the fine details. Even the streamers are just given as a long wavy line.
On a kind of stamped clay medals or seals from Skardu (fig. 15) the common stupa forms are repeated. Over a multistoried substructure, accessible by a staircase, raises the anda, which is surmounted by the umbrellas and the streamers. Miniature stupas of a similar construction have been found in the vicinity of Gilgit too, as well as stamped seals containing the Buddhist creed (Fussman 1978. 5 ff.).
Quite another type of building is depicted in rockcarvings from Chilas III and Chilas IV (figs. 16, 17). On top of a multistoried structure very often a kind of triangular roof is shown, in many cases crowned by a trisula, streamers and other decorative elements. The several storeys as well as the roof very often are depicted as a kind of framework with wooden layers strengthening the stone construction. Whether the triangular structure on top represents a roof or a decoration of symbolical meaning derived from the umbrellas of the stupa, is still an open question. In many cases devotees (fig. 16) or vases of abundance (fig. 17) are combined with this type of monuments, which show no traces of doors or windows.
A very late stylized form of stupas has been discovered at Chilas-New Colony (fig. 18). The anda here makes a perfect circle and rests on a substructure of several platforms with roundish outlines. Anda and umbrellas are covered by a kind of veil; and right on top the uppermost part of the shaft and two short stiff streamers are depicted (Dani 1983: 88, no. 66) [fn6].
At Thalpan Bridge a marvellous carving of a floral ornament has been discovered (fig. 19), resembling the floral decorations on some of the stupa bases (figs. 7, 11, 12). Of course this small carving does not depict a conch-shell [fn7]. Similar carvings have been found at different sites they have not yet been published.
Beautiful carvings of horses have been found at Thalpan Bridge, one of them (fig. 20) (Colour photograph [231KB] added 17 June 2000) with saddle-cloth and bridles, the other with the same special form of bridles only (fig. 21) (Colour photograph [151KB] added 15 June 2000). This kind of bridles the cavesson can be traced in Sogdian wall-paintings as well as in Sasanian works of art (Thewalt 1984); the murals of Alchi show a slightly modified form of such bridles. Except the tail of the first horse, there is nothing in them to compare to the horses of the T'ang dynasty , as Dani did ( 1983 : 238 , nos. 194 , 195).
Both horses are depicted without a rider in an imposing movement.
Besides the representations of architectural features, of animals and decorative designs, many carvings have been discovered depicting the Buddha himself and several Bodhisattvas (Jettmar 1980a: pls. 2-4; 1980b: 7; 1980c: pl. 2; 1983: fig. 1 ; Dani 1983: 129 ff.) as well as a number of jataka-scenes (Thewalt 1983) and some stories of the life of the Buddha.
In Thalpan Bridge for example, the First Sermon of the Buddha in the deer-park at Benares is depicted (fig. 22) (Colour photograph [91KB] added 15 June 2000) (Jettmar 1983: fig. 7; Dani 1983: 142, pl. 6). The Buddha is sitting crosslegged among his first disciples, holding the hands in dharmacakramudra, the gesture of teaching. Both his shoulders are covered and the folds of his robe are rendered in circular parallel lines. Below the Buddha the dharmacakra is shown on a column and two deers on both sides of the wheel indicate the site where the First Sermon took place.
Another rockcarving from Thalpan Bridge shows the legend of the Buddha's temptation by the daughters of Mara (fig. 23) (Colour photograph [112KB] added 15 June 2000). Unfortunately, this picture is very dim and worn-off by the erasing force of the daily sandstorms. In the centre of the composition the Buddha is depicted, sitting in a crosslegged fashion in a rocky landscape, indicated by angular meandering lines (Thewalt 1983: fn. 7, fn. 34). He is shown in a meditating pose, the left hand being covered by the garment and the right hand touching the earth in bhumisparsamudra, while two half-naked daughters of Mara are shown dancing in seducing and lascive postures on both sides of the Buddha who, of course, is completely unimpressed by these temptations. In this carving the right shoulder of the Buddha remains uncovered, while the rendering of the folds is very similar to the scene of the Buddha's First Sermon and a number of other representations of the Buddha (Dani 1983: nos. 111, 113, 116), but these features need further investigation, which will be carried out in future .
While representations of stupas and other monuments of different faiths abound in nearly all the sites where carvings have been discovered, figures of the Buddha and Bodhisattvas have been found nearly exclusively in the sites of Chilas I and Thalpan Bridge with a few exceptions at Thor-Nord, Chilas II and Shatial. The distribution of the various types of carvings and of the different themes will be shown later on in the maps of the sites. But this is a future task and cannot be undertaken before the rockcarvings and inscriptions in the Northern Areas of Pakistan are completely documented.
 From 1979 through 1983 research has been made possible by the generous financial aid of the German Research Society (DFG) and the foundation Volkswagenwerk; in future the project will be promoted by The Academy of Arts and Sciences in Heidelberg.
 Some stupas from Chilas II also resemble very closely the shape of votive stupas (cf. Dani 1983, nos. 86, 87).
 For other examples, see Jettmar 1980a: pl. 5.2; 1980b: frontispiece; 1980c: pl. I; 1982a: pl. 5.2; Dani 1983: 91 f'f.; Thewalt 1984: figs. 2-4.
 The various crowning elements deserve further investigation.
 Cf . Foucher 1905: 70, 72, 77, figs. 3, 4: Ingholt 1957: fig. 471; Concerning the shape of the niche , cf. Thewalt 1982: 171 ff .
Dani's description makes no sense at all e.g., no trisula is visible in this particular rockcarving, and stupas of a very similar type, with veiled anda and umbrellas, are depicted in the murals of Alchi (c.f.. Goepper 1982; Pal 1982).
 As it was interpreted by Dani (1983. 224 , no. 18 1); but it seems to be a kind of exercise of a travelling artist. Similar carvings have also been discovered in other places they will be published later on.