Ancient Indian Colonies in the Far East (Vol. I: Champa; Vol. II: Suvarnadvipa) v2.c21.s6: Chapter VI. Art in Bali

Ancient Indian Colonies in the Far East (Vol. I: Champā; Vol. II: Suvarṇadvīpa) v2.c21.s6

Language: English

v2.c21.s6.t1 Architecture — 45 parts
v2.c21.s6.t2 Sculpture — 53 parts
v2.c21.s6.t3 I. Palembang (Śrī-Vijaya) — 24 parts
v2.c21.s6.t4 II. Jambi ( Malayu) — 4 parts
v2.c21.s6.t5 III. The Tapanuli region — 29 parts

Chapter VI. Art in Bali

v2.c21.s6.2

¹ Unless otherwise stated, the pages, plates, and figures in the following footnotes refer to "Oudheden Van Bali" by Dr. W. F. Stutterheim (Singaradja, 1929).

Outside Java, the island of Bali is the most important region in Malayasia from the point of view of the history of art. Although very poor in architectural remains, its sculptures offer a fruitful subject of study. For a chronological study of the artistic remains of Bali, we may distinguish the following broad periods.

v2.c21.s6.4

I. The Early period, which may be called the Hindu period or Old-Hinduistic Period (eighth-tenth century A. D.).

v2.c21.s6.5

II. The Early Indo-Balinese period (tenth-thirteenth century)

v2.c21.s6.6

III. The Middle Indo-Balinese period (thirteenth-fourteenth century)

v2.c21.s6.7

IV. The modern period (fifteenth century and later)

v2.c21.s6.8

It would be more convenient to treat architecture and sculpture separately for the whole period of Balinese history.

Architecture

v2.c21.s6.t1.1

II. The Early Indo-Balinese Period

v2.c21.s6.t1.2

Very little remains of the old buildings of Bali, and there is not a single structure which may be positively dated before the eleventh century A. D. Of the monuments which have been tentatively referred to this period two alone deserve detailed consideration, a rock-cut monastery, called Goa Gaja, and the rock-hewn temples of Gunung Kavi.

v2.c21.s6.t1.3

The Goa Gaja¹ (=Skr. Guhā-Gaja or Elephant's cave) is a cave situated a little to the south of the road from Pliatan to Bedulu, and on a small rivulet, tributary to the Petanu. The name was explained by Heyting as Elephant cave by regarding the front of the cave as shaped like an elephant. But as Bosch has shown, the figure is really a Kāla-head partly damaged.

v2.c21.s6.t1.4

¹ For a detailed account and illustration of Goa Gaja, cf. Bosch in O.V., 1923, pp. 66ff ; Heyting in O.V., 1925, pp. 94ff. Pl. 23 ; N.I.O.N. 1926, pp. 333ff ; 375ff. Moojen-Bali, pp. 104ff.

v2.c21.s6.t1.5

The cave was hewn out of a large isolated block of stone. The entrance consists of an almost rectangular opening (slightly slanting upwards) about 3'-3" wide and 6'-6" high. Above this is the Kāla-head, with nose and left cheek badly damaged. Its prominent features are protruding eyes, earrings (used by females), and fingers of two hands with nails curled inside. The cave itself serves as the body of the monster. On two sides of the head, the front of the rock is carved with the so-called rock-motive, a sort of decoration such as we often find on rocks in Eastern Java. It consists of a series of intricate and intertwining curves (meant for leaves or petals) deeply undercut so as to give a free play to light and shade. In Goa Gaja this motif is improved by the addition of small panels depicting men and animals.

v2.c21.s6.t1.6

The opening runs to a rectangular passage which is at first 6½ ft. broad, and then narrows down to half this width. At the end of the passage, which is about 29¼ ft. from the entrance, lies a rectangular chamber at right angles to it, the whole structure thus having the shape of a T. This main chamber has a length of 43'-3" with breadth of 9 ft. It has eleven niches, seven on the back wall, one on each side wall and one on each side of the passage in front wall. Towards the left end, the floor of the chamber is raised to serve as a step beneath the niche. There are also four other niches, one on each side of both the wider and the narrower passage.

v2.c21.s6.t1.7

All these niches except three are now empty. A Gaṇeśa image is found in one, while another contains a part of the back of a rākṣasa which is now placed in front of the cave. The third niche, to the right end of the main chamber, contains three Lingas, each surrounded by eight smaller ones and encircled by a band at the centre.

v2.c21.s6.t1.8

Two words are inscribed on two sides of the entrance. The script of one of these may be referred to the eleventh century A. D. to which period the cave also probably belongs.

v2.c21.s6.t1.9

Next to Goa Gaja we may refer to the rock-hewn temples or Caṇḍis, of Gunung Kavi,¹ near Tampak Siring (Pl. XLVIII, Fig. 1). On the rock walls, on the two sides of the rivulet Pakerisan, have been excavated ten niches—five on each side—with a temple façade within each. These are not really temples, but merely reproduce their forms over the burial places of kings and queens, members of royal family, and one high royal official. The temple façades have all the mouldings, but are not decorated with sculptures or ornaments. The most interesting part is the high conical storeyed roof which perhaps shows a transitional stage between the art of Central and Eastern Java.² The detached towers in the temple-roofs of Central Java were, as we have seen, transformed in Eastern Java into a series of successive vertical mouldings between the horizontal bands of the roof. Here at Gunung Kavi, the towers still exist, but they regularly form successive stages in the roof.

v2.c21.s6.t1.10

¹ Krom—Kunst, II, pp. 52-53, Pl. 48. O.V., 1920, pp. 115-16; 1921, pp. 60ff, pp. 97ff; 1922, pp. 79ff. B.K.I. Vol. 76, p. 483. N.I.O.N., 1926, pp. 264ff, pp. 295ff. Djawa, 1921, Afl. 3. Moojen-Bali, pp. 104ff. Stutterheim, pp. 145ff.

v2.c21.s6.t1.11

² This is the view of Krom (Kunst, II, p. 53), but is denied by Stutterheim (p. 147 f.n. 1).

v2.c21.s6.t1.12

According to Stutterheim nine of these temples were built over the burial places of the youngest brother of Erlangga and his eight queens; the tenth, which stands a little apart, being that of a high royal official.¹ It may be added that there is a small monastery near the temples. The use of the so-called "Kediri-script" in the inscriptions on the temples led Stutterheim to refer their construction to the eleventh century A. D., as an inscription in the same script bears the date 1077 A. D. H. T. Damste has, however, referred the foundation of the group to 959 A. D. on the strength of a Candra-Sengkala record. Stutterheim, however, sticks to the old date, taking 959 A. D. as the date of a natural bursting of the rocks, or of the foundation of the monastery.²

v2.c21.s6.t1.13

¹ PP. 145-6.

v2.c21.s6.t1.14

² B.K.I. Vol. 90, pp. 297ff.

v2.c21.s6.t1.15

A little to the south of Gunung Kavi, at the junction of the Krobokan and Pakerisan rivulets, there is a similar temple-façade within a niche flanked by two other lower chambers serving as monasteries, all hewn out of rock.³ Similar niche monasteries called Goa-Garba⁴ are found in Pura Pengukur not far from Pejeng. All these may be referred to about the same period.

v2.c21.s6.t1.16

³ PP. 146-7, Fig. 39.

v2.c21.s6.t1.17

⁴ PP. 147-8, Fig. 40.

v2.c21.s6.t1.18

III. Middle Indo-Balinese Period

v2.c21.s6.t1.19

An example of a similar rock-hewn Caṇḍi of a later period is furnished by that lying between Pejeng and Tatiapi on the river Kalebotan.⁵ But numerous examples of miniature Caṇḍis enable us to form an idea of the changes that came over the temple structures during the thirteenth and fourteenth centuries A. D. Among the more remarkable changes we may note the following⁶ :—

v2.c21.s6.t1.20

⁵ PP. 181-2, Fig. 74.

v2.c21.s6.t1.21

⁶ Cf. Figs. 69, 71, 73, 75.

v2.c21.s6.t1.22

(1) The square towers of the roof have been transformed into round vases.

v2.c21.s6.t1.23

(2) The basement has got an independent profile consisting, like classical Javanese temples, of a plinth, ogee and pointed half-round mouldings above and below the central vertical part, and the cornice.

v2.c21.s6.t1.24

(3) The cornice is furnished with ante-fixae ornaments, the fore-runner of what we find in modern Balinese temples. Occasionally we meet with Kāla-Makara ornament.

v2.c21.s6.t1.25

(4) Sometimes we find a plain vertical block of stone in the middle of the front-side of the basement, and this is perhaps now represented by the decorated piece in sacrificial altars in the modern temples in Bali.

v2.c21.s6.t1.26

The miniature Caṇḍi from the Pura Desa Pedapdapan at Pejeng shows that sometimes a temple had three subsidiary chapels on the three sides, containing images of gods.¹

v2.c21.s6.t1.27

¹ P. 181, Fig. 69.

v2.c21.s6.t1.28

Although we have no example of big structural temples dating from a period before the fifteenth century A. D., their existence can be inferred from the finds of isolated parts of temples resembling those of the miniature Caṇḍis. A stone lintel of a fairly big size has, for example, been found in the neighbourhood of Panataran Sasih,² and this undoubtedly proves the existence of a big temple.

v2.c21.s6.t1.29

² P. 181.

v2.c21.s6.t1.30

Among minor architectural objects may be mentioned the square pillars supporting lotus cushions. The specimen from Pura Santrijan³ illustrates the early form and that from Pura Desa at Bedulu⁴ shows the later development. The Nāga-heads in the later example were transformed into ante-fixae ornaments, and ultimately we have the modern ‘Kekocong’, hundreds of which can be seen on the corner-pilasters of boundary walls⁵.

v2.c21.s6.t1.31

³ Fig. 79.

v2.c21.s6.t1.32

⁴ Fig. 80.

v2.c21.s6.t1.33

⁵ 184.

v2.c21.s6.t1.34

IV. Modern period

v2.c21.s6.t1.35

There are many beautiful temples of moderate size in Bali, but all these are subsequent to the period when Javanese settled in large number in Bali after the fall of Majapahit. It would be convenient for a proper understanding of the subject if we give an idea of the general plan of temple-enclosure before describing specific examples.

v2.c21.s6.t1.36

The temple area is divided into three court-yards (Pl. XLVIII, Fig. 2), separated from one another by low walls with gate. An interesting peculiarity is that these gates are not placed in the same line of axis nor is any of them placed precisely in the centre of the wall.

v2.c21.s6.t1.37

The first court-yard is usually empty, though occasionally there is either a Balê or an offering-pillar. From the first court-yard one passes through the principal gateway to the second court-yard. Here we usually find a number of Balês and the offering-pillar. The Balê is a small square chamber where the offerings are kept. It has usually a stone-foundation which supports wooden posts and rafters carrying a simple pointed roof. The posts and rafters are decorated and the scenes from the Rāmāyaṇa and Mahābhārata are illustrated on the walls. In the Pura Desa (public temple in a locality) there is a special Balê called Balê Agung, a long rectangular chamber, which serves as a prayer hall for the village elders on the new and full-moon days. A seat for the god is placed at one end of the chamber.

v2.c21.s6.t1.38

A simple doorway leads to the third court, the holy of the holies, which contains a row of temples, sacred altars for offerings, seats of gods made of stone, small houses, offering-pillars, and occasionally also a bell-tower.

v2.c21.s6.t1.39

Of the temples, those dedicated to Śiva, Viṣṇu and Brahmā, called Meru, are the principal ones. The sanctuaries of Viṣṇu and Brahmā lie respectively to the north and south of that of Śiva. The height of the temple-roof and the number of storeys it contains clearly shows that Śiva occupied the highest place in the Trinity, and Brahmā, the lowest.

v2.c21.s6.t1.40

Having thus described the general plan of temple we shall now describe the different parts of it in some detail. In essence, the plan of a Meru resembles that of Javanese temples, and rock-hewn Caṇḍis of Bali. A square basement with mouldings, a square chamber constituting the cella, and a conical pyramidal roof consisting of a series of gradually diminishing storeys ending in a finial representing linga or ratna,—these have been the essential features of Balinese temples in all ages. There have been no doubt modifications, mainly due to materials used. At first perhaps the shrine was closed by wooden walls and the roof was made of wooden frame with thatched cover. Gradually stone replaced wood in all these parts. Pl. XLIX illustrates the chief temple of Besakih which preserves the wooden forms. The next stage, when the walls of the sanctuary were made of stone, the roof remaining the same, is illustrated by the temple of Kesiman (Pl. L). The final evolution of a temple completely made in stone is illustrated by the temple at Tonjo (Pl. LII, Fig. 1). It is remarkable how the roof has an altogether different appearance although the essential features remain the same.

v2.c21.s6.t1.41

Next to the temples, the gateways demand our chief attention. They are either open or closed at the top. The former type is chiefly found in North Bali at the entrance of the first court-yard. It is virtually a large opening flanked by two pilaster-like constructions, having neither doors nor any roof.

v2.c21.s6.t1.42

There is no doubt that the closed gateway was modelled as a miniature temple, and the chief gateway in North Bali with two side-gates exactly correspond to the three Merus.

v2.c21.s6.t1.43

Pl. LI illustrates the open and closed types of gateways at Singraja. The beautiful inner gate with a roof of five storeys, exactly resembles a Meru, and shows the scheme of decoration which plays such a prominent part in Balinese temples. The open or split doorway on the right-hand side of the figure is obviously derived from the other type.

v2.c21.s6.t1.44

The Balinese display their artistic skill in lavishly decorating these gates and sometimes colour is added to heighten the effect.

v2.c21.s6.t1.45

The smaller adjuncts of the temple such as offering pillar with niches (Pl. LII, Fig. 2), sacred altars for keeping offerings (Pl. LV, Fig. 1), and Bell-towers (Pl. LIII) (see later) are also sometimes done with elaborate details.

Sculpture

v2.c21.s6.t2.1

1. The Hindu Period

v2.c21.s6.t2.2

The earliest sculptures in Bali are the Buddhist figures engraved in relief on small clay tablets. These are found with small clay stūpas containing small clay seals inscribed with the well-known Buddhist formula ‘ye dharmā hetuprabhavā etc.’ The purely Indian alphabet of these inscriptions may be referred to the eighth century A.D. The relief figures of the Buddhas and the Bodhisatvas may also be approximately referred to the same, or perhaps a little later period, and show a purely Indian style. Stutterheim describes this style as ‘internationally Buddhistic’ and says that such figures could as well have been found in the mountains of Tibet or the hills of China. But he has not adduced any evidence in support of his assertion. The style is, however, predominantly Indian, and these figures might well have been mistaken for Indian sculptures of the late Gupta period. This can only be explained by a predominant Indian element in the population of Bali who had carried their art traditions direct from India, and had not yet been subject to indigenous influence, at least to any considerable degree. The figures were presumably impressed upon clay by means of bronze stamps, and the tablets were perhaps regarded as sacred objects or amulets with magic properties¹.

v2.c21.s6.t2.3

¹ See below, Chap. IX. Also cf. T.B.G. 1935, pp. 133-4 for the evidence that the clay-seals were made in Bali and not imported from outside.

v2.c21.s6.t2.4

But the influences of art-traditions of Central Java at about the same period (c. eighth to tenth cent. A.D.) are proved by the image of Śiva from Bedulu illustrated in Pl. LIV. Although somewhat damaged, this fine figure is undoubtedly reminiscent of the flourishing period of Central Javanese art. The political influence of Mataram dynasty² over Bali may account for this. Reference may also be made to a remarkable head of a Buddhist (?) image (Pl. LVI, Fig. 1) which shows the peculiar smile of Cambodian figures. Its high cylindrical head-dress contains a lion's head in front, and on the whole it shows striking difference from Javanese style.

v2.c21.s6.t2.5

² See above, Vol. I, pp 419ff.

v2.c21.s6.t2.6

To judge from these and other similar sculptures which may all be dated between the eighth and tenth centuries A. D., we must regard this as the earliest period in the evolution of art in Bali during which the Indian influence was dominant and indigenous element was conspicuous by its absence.

v2.c21.s6.t2.7

II. Early Indo-Balinese Period

v2.c21.s6.t2.8

During the succeeding period, however, the native element comes to the forefront, and we find a distinctive Balinese style. But gradually this Balinese style is influenced by Javanese, evidently on account of the close political relations subsisting between the two islands.

v2.c21.s6.t2.9

The earliest specimens of this art are furnished by two groups of images found at Pura Sukavana on the Genung Panulisan¹ (Pl. LV, Fig. 2). Each group consists of two figures, a man and a woman, in one case standing, and in the other case, seated. The standing images wear clothes of variegated patterns, while the seated ones have plain clothes and wear the sacred threads across the upper part of the body. Both groups have a small round object in their hand. The figures appear to be those of a king and a queen, and both groups may represent the identical royal pair. The date 1101 A. D. engraved on the back makes it not unlikely that they were the parents of Airlangga, *viz.*, Udayana and Mahendradattā. This is, however, by no means certain.

v2.c21.s6.t2.10

¹ Figs. 19, 20.

v2.c21.s6.t2.11

About this class of images Stutterheim has observed as follows :—"Some figures of deified kings and queens, very often depicted together standing on one base, show such pronounced Indonesian features that no doubt can remain about the drastic departure from Hindu styles. The bodies have no more the slender grace and movement of the Hindu gods. Ridgily closed, they stand like mummies. The hands and feet are heavy and crude and the faces larger than the proportions of nature would allow for”.¹

v2.c21.s6.t2.12

¹ I. A. L., Vol. VI, pp. 2-3.

v2.c21.s6.t2.13

Stutterheim has very rightly drawn attention to the fact that some of the peculiar features of this style may perhaps be due to the circumstance that the statues were intended to represent deceased kings and queens, and belonged to a special category of art, which may be called the ‘art of the dead’.² That the Balinese art in this period was capable of producing more graceful figures is proved by the statues found near the cave Goa Gaja, to which we next turn.

v2.c21.s6.t2.14

² In a recently published work Stutterheim maintains “that almost everything that remains from the ancient art of Bali and Java pertained to this cult of deceased kings. Hardly any of the innumerable images found or excavated ever served for the direct worship of a god in the manner customary in India previously and now........In brief, nearly all the pieces of any significance, fashioned of natural stone and thus able to resist the wear of ages, belonged in the past to the cult of dead kings.” ( Indian influences in Old-Balinese Art—pp. 23-4 ) It is impossible to accept this highly exaggerated value attached to the cult of the dead kings. The absurdity of this view will appear from the account of images given in this Chapter.

v2.c21.s6.t2.15

The images found in the cave, and those that are now set up in front of it, were evidently brought from some other places in the neighbourhood. Among these are several male and female figures, including an image of Gaṇeśa, which served as water-spouts. The female figure, illustrated in Pl. LV, Fig. 3, is a fine piece of sculpture. The modelling is graceful and shows a developed technique and naturalistic style allied to Indian art. The expression is full of serenity and beauty. Another group, Hāriti surrounded by seven children, though less successful as work of art shows a fair degree of skill. Stutterheim has drawn attention to the fact that the style of these works resembles that of Belahan, and he refers them to the reign of Airlangga.¹ To the same group belong other sculptures which have a more pronounced indigenous element.²

v2.c21.s6.t2.16

¹ pp. 123-4.

v2.c21.s6.t2.17

² Fig. 23.

v2.c21.s6.t2.18

The third group of images, called Kutri-group, shows the same style as we meet with in Goa Gaja. The best specimen of the group, Durgā-Mahiṣāsuramardini,³ found at Kutri, is now very badly damaged, but enough remains to show that it was one of the finest images in Bali. The goddess stands over the buffalo and holds various weapons in her eight hands. A similar figure from Pura Puseh in the neighbourhood offers a strong contrast by its heaviness and lack of grace.⁴ Midway between the two stands the figure of the eight-armed Amoghapāśa.⁵

v2.c21.s6.t2.19

³ Fig. 27.

v2.c21.s6.t2.20

⁴ Fig. 28.

v2.c21.s6.t2.21

⁵ Fig. 29.

v2.c21.s6.t2.22

The fourth group of images, called Gunung Kavi group, found over a wide area from Gunung Panulisan to the southernmost part of the island, is characterised by the presence of a cursive writing known as Kediri script. There is no uniformity of style in these images. Some of them indicate an independent style, but others resemble more or less that of the Kutri-group. Among the former may be mentioned the fine figure of a young lady found in Pura Sukavana⁶ on the Gunung Panulisan. The ornaments are different in character, but simpler and more finely worked. The face is badly damaged, but the features are on the whole graceful. An inscription on the back, dated 999 Śaka, relegates it to the last quarter of the eleventh century A. D.

v2.c21.s6.t2.23

⁶ Fig. 31.

v2.c21.s6.t2.24

Before closing the Early Balinese period we may make a few general remarks. As we have seen above, purely Balinese traditions prevailed at the beginning, but these were gradually modified by the art-traditions of Java. This can be easily explained by the close contact between the two islands during the period of Airlangga. It is, however, also very probable that the accession of this Balinese prince on the throne of Java introduced Balinese art-traditions in Java, and this may explain the characteristic and distinguishing features of East Javanese art as opposed to that of Central Java. The somewhat sudden growth of a new style of architecture and sculpture in Eastern Java forms an interesting problem, and the possiblity that the Balinese influence might have been an important factor in this transformation cannot altogether be ignored. Unfortunately sufficient data have not yet been collected for the purpose.

v2.c21.s6.t2.25

III. Middle Indo-Balinese period

v2.c21.s6.t2.26

From the point of view of art and architecture the two following centuries, the thirteenth and the fourteenth, may be regarded as the Middle Indo-Balinese period. We give below a short description of some of the representative specimens of the period.

v2.c21.s6.t2.27

(a) Pl. LVI, Fig. 2. illustrates an aristocratic lady. The purely formal representations of dress and the attitude of frontality, together with the coarseness of the legs and hands, show a degraded form of art. The face is, however, vigorously executed, and special attention may be drawn to the terraced head-dress, the long hanging ear-ornaments, the peculiar shape of the halo, and the heavy ornaments.

v2.c21.s6.t2.28

The figure shows the gradual dominance of Indonesian over Indian element which is also a characteristic feature of late East-Javanese art. As the figure contains a short record, with a date equivalent to 1342 A. D., it is earlier in date than the known Javanese specimens of that style. It is conceivable, therefore, that this assertion of indigenous spirit might have exercised considerable influence in the later development of Javanese art.

v2.c21.s6.t2.29

Stutterheim thinks that it was a memorial image of a dead queen, perhaps the mother of king Valajayakṛtaningrat.

v2.c21.s6.t2.30

(b) Several sculptures represent a combination of four figures facing the four directions. A good example is furnished by Pl. LVI, Fig. 3. The head-dress and the ornaments resemble those of the preceding one. Each figure has got four hands and a third eye, and holds a *śaṃkha* (conch-shell) and a book in the upper hands. It seems therefore that each figure represents the Hindu Trinity, the third eye, the *śaṃkha*, and the book being respectively the attributes of Śiva, Viṣṇu and Brahmā.

v2.c21.s6.t2.31

(c) A horrid but vigorously executed example of the above type is furnished by Fig. 1, Pl. LVII. It illustrates four dancing figures with masks in the form of shields covering their faces.

v2.c21.s6.t2.32

(d) Sometimes the sculptures are realistic and even characterised by individuality, as we find in the Sukuh-group in Eastern Java. The best specimen is furnished by the figure of a man with child serving as a spout.¹ Another good specimen is the figure of the so-called ascetic Arjuna serving as a spout.² The features are fairly modelled, and the dress and ornaments are of the simplest kind.

v2.c21.s6.t2.33

¹ Fig. 64.

v2.c21.s6.t2.34

² Fig. 61.

v2.c21.s6.t2.35

(e) The fanciful and bizarre figures like those of Eastern Java are also not wanting³.

v2.c21.s6.t2.36

³ Fig. 62.

v2.c21.s6.t2.37

(f) A composite scene is depicted on a sacred water-pot found in Pura Puser at Pejeng⁴ (Pl. LVII, Fig. 2). It is about 2½ ft. high, drum-shaped, and open at the top. The whole outer surface is minutely carved. The principal figures are those of eight snakes in four pairs on four sides. The hoods of each pair are interwined and crowned by a Śivaliṅga. The tail of each snake is interlaced with that of the neighbouring snake in the next group, and the two ends of each pair of tails support a lotus cushion on which a god is seated. The body of the snake is carried by gods, richly dressed, and bending under the heavy burden. They are evidently running over the sea, the representation of which covers the lower part. The upper part is filled with landscape, plants, trees and birds, and a few figures. The upper and lower rims are covered respectively with lotus petals and conventional floral designs.

v2.c21.s6.t2.38

⁴ Pp. 169 ff. Figs. 57-60.

v2.c21.s6.t2.39

The whole scene is overcrowded with details and indicates a degraded form of art such as we meet with in Eastern Java. It has undoubtedly a mythological import, the exact nature of which is, however, difficult to understand. There is a short record of date in Candra Sengkala. It has been interpreted by Stutterheim as 1251 (=1329 A. D.) but the reading is doubtful.

v2.c21.s6.t2.40

But by far the most notable monument of the period is found at Jeh Pulu¹ less than a mile to the south of Desa Bedulu. It is a chamber hewn out of the side of a small hillock. The front is open, but there were three pilasters of which one has disappeared. This chamber probably once served as a monastery, but is now empty. The most interesting part of the monument is a long series of reliefs on the left covering a space of more than 80 ft. along the hill-side. The height of the reliefs varies, but may be roughly taken as 6'-6".

v2.c21.s6.t2.41

¹ O. V., 1925, pp. 96-8, Pll. 24-27.

v2.c21.s6.t2.42

This series of sculptured reliefs (Pll. LVIII-LIX) is undoubtedly the best so far discovered in Bali, and in point of grouping, composition and phonetic value, occupies a very high place indeed in any critical estimate of Indonesian art outside Central Java. A clear conception of the subject-matter, concentration on essential features avoiding unnecessary details, absence of over-crowding, and a general sense of rhythm, symmetry, and harmony are the chief characteristics of these reliefs. The execution of the individual figures also sometimes reaches a high level. The pose, modelling, and expression of the first three figures on the left indicate a high artistic skill, while the spirited hunting scene in the centre, and the figures of a horse and rider on the extreme right show that the artists were equally capable of executing vigorous and animated movements. Among the scenes illustrated is a representation of a dwelling house, with a female figure standing near the doorway. The house closely resembles an ordinary thatched cottage in Bengal, with four triangular roofs meeting in a point at the top. On both sides of the door the walls have sculptured decorative motives such as scroll, and reel and bead border. The so-called rock-motives, used to fill the vacant space, are somewhat different in design, and have served as a clue to the dating of this record. Dr. Bosch concludes that the sculptures should be referred to the Majapahit period.

v2.c21.s6.t2.43

Before we conclude the study of sculptures we may refer to two characteristic specimens which cannot be definitely referred to any specific period. The first is a Śivaliṅga surrounded by eight busts and eight seated figures in two rows.¹ The upper row of busts is enclosed within lotus petals, and the figures in the lower row are seated on lotus cushions.

v2.c21.s6.t2.44

¹ Fig. 85.

v2.c21.s6.t2.45

The next is a representation of Narasiṁha incarnation (Pl. LX, Fig. 1). The style is very peculiar, almost baroque in character.²

v2.c21.s6.t2.46

² Fig. 86.

v2.c21.s6.t2.47

IV. Modern Period

v2.c21.s6.t2.48

The sculpture of the latest period is best seen in the decoration of the hundreds of temples scattered over the island. The Balinese, not rich or resourceful enough to raise imposing structures, lavished their skill and energy in the exuberant and excessive decoration of temples, specially their gates and walls, by means of sculpture and painting. Among purely decorative elements we find a rich variety of foliage, flowers and geometrical patterns, besides birds, legendary animals, Kāla heads, and figures, which are often of grotesque, fantastic and demoniac character. Reliefs, depicting stories from old Hindu scriptures, adorn the walls. Among the figures in the round we have not only images of benevolent gods and mythical beings such as Garuḍa and Kinnara, but also gods of terrific appearance, dreadful dvārapālas (temple-guards), and various types of demons and monsters. In spite of high technical skill and some amount of graceful beauty, the art is marked by a highly conventionalised style and the general tendency seems to be a leaning towards the representation of the weird, the grotesque, and the terrible in a somewhat Baroque style. This will be best illustrated by Pll. LIII, LV (Fig. 1), and LX. It must be remembered, however, that the weird and grotesque representations are dictated by religious considerations, and used for the definite purpose of warding off evil spirits who are continually threatening the sacred places with destruction. The modern Balinese art is a living one, and an adequate idea of its nature and varied character can only be given by a large number of illustrations which are impossible within the present scope of this work.

v2.c21.s6.t2.49

Conclusion

v2.c21.s6.t2.50

A general review of the architecture and sculpture in Bali enables us to mark several well-defined stages in the evolution of its art. At first it was dominated by purely Indian traditions. From the tenth century A. D. the art was gradually modified by the working of indigenous element, and it was further subjected to influence from Java. The result was that during the same period we find different styles and levels of excellence in proportion to the presence or absence of Indian and Indo-Javanese tradition and influence. This may be called the truly Balinese period as distinguished from the purely Indian period preceding it. Following Stutterheim's classification we have divided this period into Early Indo-Balinese and Middle Indo-Balinese periods. In general it is not always easy to distinguish between the two, though on the whole the later period is marked by a gradual lessening of Indian influence, the growing influence of late East-Javanese style, the marked dominance of Tāntrik ideas, a predilection for the bizarre, fanciful, repulsive, and uncouth figures, the exuberance of ornaments, and some peculiarities in dress, head-gear, and ear-ornaments.

v2.c21.s6.t2.51

The final Javanese conquest of the island of Bali in the fourteenth century A. D. led to a sudden end of the normal evolution of art in that island. For some time the artistic instinct lay dormant, and it appeared as if it vanished altogether as it did in Java after the Muslim conquest. But with the infusion of new blood as a result of the large-scale immigration of the Javanese into the island we find a gradual revival of art, specially temple-architecture. Thus was ushered in the modern period of Balinese art which continued down to the Dutch conquest of the island. Although considerably influenced by the Javanese ideas, the modern art is based on the traditions of the earlier Balinese art. Whether it would continue to flourish under the present conditions is more than can be said at present.

v2.c21.s6.t2.52

ART IN SUMATRA

v2.c21.s6.t2.53

The archaeological remains in Sumatra are very poor, and it is not possible to write a connected history of its architecture and sculpture. The most convenient way of dealing with the extant remains is to treat them separately according to some well-defined regions.

I. Palembang (Śrī-Vijaya)

v2.c21.s6.t3.1

¹ The account is based on the following :— Perquin in O. V., 1928, pp. 123ff. Bosch in O. V., 1930, pp. 152ff. Krom in A. B. I. A., 1931, pp. 29-33. As regards the origin of the art of Śrī-Vijaya I have discussed the question in an article in the Journal of the Indian Society of Oriental Art, June, 1935. The article also contains a detailed discussion of the sculptures of Palembang, referred to below.

v2.c21.s6.t3.2

We may begin with Palembang, generally identified with the capital of the far-famed kingdom of Śrī-Vijaya.² It is a matter of considerable surprise that it has not yet yielded any architectural remains of any importance. A brick terrace or floor of a Buddhist foundation on the Seguntang hill outside the city of Palembang, and similar ruins at Caṇḍi Angsoka within the city itself practically exhaust the list. Dr. Bosch has truly observed that "unless we presume that only perishable materials were used or that there has been a general subsidence of the soil, we must conclude that the extant remains give no support to the view that the capital of Śrī-Vijaya occupied the site of the present Kota Palembang.³"

v2.c21.s6.t3.3

² On this point, see Part I., pp. 209ff.

v2.c21.s6.t3.4

³ O. V., 1930, pp. 155-6.

v2.c21.s6.t3.5

We are, however, more fortunate in regard to sculptures. For, although their number is not very large, they are of considerable interest.

v2.c21.s6.t3.6

A Buddha statue found on the Seguntang Hill (Pl. LXI-LXII) may be regarded as belonging to the earliest period of Śrī-Vijayān art.¹ The most striking feature of the figure is the drapery with prominently marked fold-lines. It is unlike anything hitherto discovered in Sumatra and Java, and must be traced directly to Indian influence. Originally a Gandhāra origin was postulated, but the nature of the fold-lines and the threefold vertical border at the back definitely militate against this view. In a recent article Krom traces its origin not to Gandhāra school but to ‘some Indian school of art in which the Graeco-Buddhist principles are found in a modified and Indianized form e.g. that of Amarāvatī.’² This is more acceptable, although we need not confine our attention exclusively to Amarāvatī school. For drapery like this was current in Mathura, and the Mathura school exerted its influence all over Northern India. Similar drapery is found in images as far apart as Bihar and Ceylon.

v2.c21.s6.t3.7

¹ By piecing together the loose fragments which were found on the Seguntang Hill it was possible to reconstruct the torso only. For illustration cf. A. B. I. A., 1931 Pl. XI, a, b. The Head of the statue was subsequently traced by Schnitger in the Museum at Batavia (Oudheidkundige vondsten in Palembang-Bijlage B, 1st Ed., p. 4, 2nd Ed., (Palembang, 1935) pp. 1, 4.

v2.c21.s6.t3.8

² A. B. I. A., 1931, p. 32.

v2.c21.s6.t3.9

Another remarkable stone statue is that of Avalokiteśvara found near the village of Bingin in the Palembang Sub-division³ (Pl. LXXI, fig. 1). Its vigorous expression gives it a peculiar stamp of originality and indicates a high degree of technical skill. A torso found at Mangkubumi, and the fragment of a Bodhisatva head found at Seguntang are also good specimens of art⁴.

v2.c21.s6.t3.10

³ Ibid, Pl. XII, d.

v2.c21.s6.t3.11

⁴ Schnitger, op. cit., pp. 5, 6, 10.

v2.c21.s6.t3.12

Special importance attaches to three bronze images, found in course of dredging the river Komering, about 1100 yds. from the place where it flows into the river Musi. According to Bosch these finds indicate it to be the probable site of the old capital city, which thus lay to the south of modern Palembang in a territory now fully covered by morass.

v2.c21.s6.t3.13

The images (Pl. LXIII, figs. c, d, e) are those of Avalokiteśvara, Buddha, and Maitreya.¹ The most striking feature of these figures is the style. It is difficult to agree with Krom that these ‘might easily be taken for Javanese Bronzes’.² As Bosch has already observed, the style is very different from that of the stone Buddhist images of pure Central-Javanese type found at Bukit Seguntang. The same scholar has further remarked that these images furnish the only known specimens in Śrī-Vijaya of a style most closely related to the Śailendra art of Java³.

v2.c21.s6.t3.14

¹ Illustrated in O. V., 1930, Pl, 45. A. B. I. A., 1931, Pl. X, c. d. e.

v2.c21.s6.t3.15

² A. B. I. A., 1931, p. 31.

v2.c21.s6.t3.16

³ O. V., 1930, p. 156.

v2.c21.s6.t3.17

In my opinion these images bear a far closer affinity to Indian art. The style of the bronze image of Buddha is clearly reminiscent of the Gupta period, and that of Maitreya probably belongs to the same school. The fine image of Avalokiteśvara probably represents a later stage of the same art. These three images may be regarded as the best specimens of sculpture discovered in Sumatra. Taking them along with the Buddha image of Seguntang we may infer that a direct stream of artistic influence flowed from India to Śrī-Vijaya down to the sixth and seventh centuries A. D.⁴

v2.c21.s6.t3.18

⁴ Cf. my article referred to in f. n. 1, p. 323.

v2.c21.s6.t3.19

Reference may be made to a peculiar bronze head of Buddha.⁵ (Pl. LXIII, figs. a-b). A fillet, decorated with rosettes, surrounds the head,—a feature not met with elsewhere. In this case also we cannot postulate a Javanese origin.

v2.c21.s6.t3.20

⁵ A. B. I. A., 1931, Pl. X, a-b.

v2.c21.s6.t3.21

The high degree of development of bronze sculpture in Sumatra is also evidenced by the images of Buddha and Vairocana⁶ described by Schnitger, and a very fine bronze elephant with a female driver and a man sitting behind her (Pl. LXIV ) which was found in the river by Kg. Penjabungan, in the Mandabilang Sub-division.¹ A bronze image of Śiva, of the style of Central Java, and three others of Śiva, Viṣṇu and Brahmā of the fourteenth century have been found near Kota Palembang. The latter stand on their riders, lion, Garuḍa and swan².

v2.c21.s6.t3.22

⁶ Op. cit. pp. 7-9.

v2.c21.s6.t3.23

¹ O. V., 1920, p. 62, Pl. facing p. 64.

v2.c21.s6.t3.24

² Jaarboek, Vol II, 1934 (published by Royal Batavian Society) Pll. 2-3. A stone image of a princess in the Majapahit style is figured in Pll. 5-6.

II. Jambi ( Malayu)

v2.c21.s6.t4.1

Next to Palembang we may refer to Jambi, the capital of the famous kingdom of Malayu. The most important ruins in this region are those of Solok to the west of Jambi. A big structure must have once stood there. Nothing now remains of it except a standing Buddha image and four great Makaras ³ (Pl. LXV, Fig. 1) which no doubt served as the decorative ornaments of stair-curbs. Ruins of temples and a few images have also been found in other localities in the neighbourhood. One of the Makaras at Solok bears a date 986 (=1064 A.D.), and in style resembles the Javanese. This is somewhat perplexing as the political domination of Java over Jambi dates only from the thirteenth century A. D. It appears that Jambi imbibed the culture of Java more than other parts of Sumatra.

v2.c21.s6.t4.2

³ Dr. C. J. Neeb describes and illustrates the four Makara heads found near Jambi in T. B. G. Vol. 45 ( 1902 ), pp. 120-27. Dr. Brandes ( Ibid. pp. 128ff ) gives a critical description of the Makara heads and indicates their great difference from the Central Javanese art. He also gives a list of some other antiquities of Jambi.

v2.c21.s6.t4.3

The Javanised art of Jambi spread all over the ancient kingdom of Malayu, as far as the Padang Highlands⁴. It must be noted, however, that the four Makaras at Solok offer rich varieties of style, and differ, in many respects, from that of Central Java. These Makaras display a high degree of technical power and are fairly good specimens of Sumatran art. It is possible that the art of Jambi was an Indo-Sumatran art subjected to considerable influence from Java.

v2.c21.s6.t4.4

⁴ A. B. l. A., 1931, p. 30.

III. The Tapanuli region

v2.c21.s6.t5.1

Almost all the old structures of Sumatra have been so completely ruined that it is not possible to discuss their architectural style. The one structure of importance, that is in a fair state of preservation is Biaro (Bihāra) Bahal I in Padang Lawas (Tapanuli)¹ (Pl. LXV, Fig. 2., Pl. LXVI). It stands on two square basements, and is itself square in shape with shallow projections on each side. The basements have a number of sunken panels decorated by sculptures. The body of the temple is intersected by a number of horizontal courses. A domical stūpa stands on the square roof of the temple and forms its śikhara, as it were. A garland-decoration surrounds its central part.

v2.c21.s6.t5.2

¹ O. V., 1920, p. 68 ; 1925, p. 13, Pl. 5 ; 1926, pp. 26, 142 ; 1930, pp. 136ff. Pll. 32-36. Plans VI-XIV.

v2.c21.s6.t5.3

The Buddhist character of the temple is indicated by an image of Bohisattva inside it, and the remains of a stūpa (Pl. LXIX, Fig. 1), and of a pillar of the form of an elongated stūpa in the neighbourhood.

v2.c21.s6.t5.4

A second temple, Biaro Bahal II,² (Pl. LXVII) had probably the same general features, though the stūpa-śikhara is missing, and only the lower part now survives. It contained an image of Heruka, and there were also three heads of Buddha (Pl. LXVIII).

v2.c21.s6.t5.5

² O. V., 1930, p. 139, Pll. 40-42.

v2.c21.s6.t5.6

Remains of other temples have also been found in Padang Lawas, notably Biaro Si Topajan³ and Biaro Si Mangambat¹, but they are too ruined for any general description. Their ruins have, however, furnished important clues for the origin and antiquity of the Sumatran art.

v2.c21.s6.t5.7

³ O. V., 1920, p. 65 ; 1925, p. 11 ; 1926, pp. 25ff ; 1930, p. 135, Pll. 29ff, Plans I-V. On p. 243 (Corrigenda et Addenda) of O. V. 1930, Goris suggests that the second inscription contains a date Śaka 1157 which then must be the date of the temples, at least of Si Topajan.

v2.c21.s6.t5.8

¹ O. V., 1930, p. 134, Pl. 29a. According to Bosch the ruins clearly betray the influence of Central Java, and the temple probably belongs to the Śailendra period.

v2.c21.s6.t5.9

The date of the structures can be approximately ascertained from epigraphic data. An inscription in the Biaro Si Topajan bears the date 1157 Śaka=1235 A.D. The Biaro Bahal I and II may also be ascribed to the same period².

v2.c21.s6.t5.10

² Bosch at first assigned somewhat later date but changed his view on the correct reading of the date of Si Topajan Inscription.

v2.c21.s6.t5.11

The sculptures of Biaro Bahal I such as the dancing rākṣasa figures (Pl. LXVI, Fig. 2), the crouching lions, the temple watchers, and the Makara ornament do not belong to any primitive school of art, but show a high degree of technical power. But their artistic conception cannot be regarded as very high. The same thing may be said of the decorative motives.

v2.c21.s6.t5.12

As regards sculptures in the round, the stone image of Bajrasatva³ and the Bronze female figure⁴ in Biaro Bahal I deserve special attention (Pl. LXIX, Figs. 2, 3). Longhurst regarded the former as being in the same style as an image he found at Salihundan in the Ganjam district, not far from the old port of Kalingapatam, and belonging to the sixth century A. D. Bosch rejects this view and regards it as belonging to the Pāla school of sculpture. The latter seems to be the right view. The bronze female figure is undoubtedly of south Indian character. Longhurst referred it to the fifteenth century A.D., but Bosch places it about five centuries earlier. Here again we are inclined to agree with the Dutch scholar. Prof. K. A. N. Sastri is inclined to refer the image to a still earlier period, about seventh century A. D. ⁵.

v2.c21.s6.t5.13

³ O. V., 1930, Pl. 37a.

v2.c21.s6.t5.14

⁴ O. V., 1930, Pl. 38.

v2.c21.s6.t5.15

⁵ J. G. I. S., Vol. III, pp. 104ff.

v2.c21.s6.t5.16

The Heruka image⁶ in Biaro Bahal II (Pl. LXVIII, Fig. 2) is very important from iconographical point of view, as it closely corresponds to the Indian representation of the same figure, which is very rare. It is badly damaged and calls for no special remark from artistic point of view.

v2.c21.s6.t5.17

⁶ O. V., 1930, Pl. 41.

v2.c21.s6.t5.18

It is obviously risky to form any general conclusion about the art of Sumatra from such incomplete data. Dr. Bosch, who has studied the question more thoroughly than others, has, however, drawn some general conclusions from a critical study of the temples. The following observations sum up his views on the subject¹.

v2.c21.s6.t5.19

¹ O. V., 1930, pp. 144ff.

v2.c21.s6.t5.20

“The history of Buddhism in Tapanuli begins with the well-known bronze image of Bhaṭāra Lokanātha of the year 1024. As Krom has remarked, the scripts resemble the Old-Javanese alphabets of Central Java, and the style is closely related to that of Central Java.

v2.c21.s6.t5.21

“In later times a Buddhist art developed in Padang Lawas, which is sharply differentiated from Indo-Javanese by many characteristics. This may be attributed to foreign influence, particularly that of South India, which is proved by the inscription of Porlak Dolok and the bronze female figure in South Indian style found within the enclosure of Biaro Bahal I. But a closer examination negatives this theory. For neither the architectural nor the sculptural style of South India is reflected in the remains so far discovered in Padang Lawas. Thus a direct influence of South India falls to the ground. On a closer comparison of the Biaros with the Indo-Javanese temples, it appears that both have the same essential plan, and the differences are merely due to the different manners of execution. The essential elements of Bahal I and II,—the square basement with passage for circumambulation and balustrade, the stair-curbs ending in makaras, the square body of the temple with projections on the sides, the square roof with octagonal upper part and niches (Bahal II) ending in the Stūpa-form Śikhara,—all these belong to Central Java, though they have undergone in Batakland more or less radical changes and transformations, which give an altogether different appearance to these temples. The stone is replaced by brick. The relations of the lower parts to one another and to the whole are different. Ornaments, though using the same motive, make an impress of a different character; cf. *e.g.* the style of the Siṃha and the Rākṣasa figures in Bahal I. It is notable that although the main parts of the Biaros are essentially Central-Javanese in character, there are additional elements which seem to be derived from the East-Javanese art. Take for instance (1) the unsystematic plan of the complex, in which the principal temple indeed occupies a central position in respect of the surrounding wall, but the other structures are irregularly distributed in the temple area; (2) the absence of semi-circular and ogee mouldings in the basement where the rectangular courses in the base and the cornice correspond to each other; (3) the placing of temple-guards to the two sides of the entrance of the main temple; and (4) lastly, the standing position, the dress (Kain Pandjang) and the raised left fore-finger of these figures. The common characteristics of the Biaros and the Indo-Javanese architecture—which may be still multiplied—strengthen the view resulting from the absence of South Indian characteristics, *viz.* that the development of the art in Lawang Padas is not to be attributed to external influence. This art must have developed in or shortly after the Śailendra period (before the Lokanātha image of 1024 A. D.) during which an art tradition was formed directly or indirectly under the Central Javanese influence. This tradition was maintained in the following centuries with an additional influence from East Java, and subjected to a strong native influence. Thus was produced an art which clearly betrayed its Javanese origin, but nevertheless presents special character of its own.”

v2.c21.s6.t5.22

The general conclusions of Bosch do not, however, carry immediate conviction. The peculiar characteristics of the Sumatran art, as distinguished from Indo-Javanese, entitle us to regard them as Indo-Sumatran, and many of the resemblances between the two may be attributed to their common origin from Indian art tradition. We need not necessarily look to South India as the home of the parent style,—an obsession which accounts for much misconception and confusion of ideas—but we may hold that the Indo-Sumatran style owed its origin to Indian colonists who settled in this region of Sumatra. It is the result of Indian art traditions modified in Sumatra, and perhaps to some extent influenced by Java.

v2.c21.s6.t5.23

Krom has expressed this view in connection with the ruins of Muara Takus, on the upper Kampar river, and Tanjung Medan (Lubuk Sikaping)¹. There are remains of six brick monuments in the former place, including three stūpas. One of these, a high, slender, towerlike structure is a fairly good specimen of Sumatran art. The stūpa at Tanjung Medan is similar in design. Krom regards these as well as the Biaro Bahal described above as originating from Indian colonists who settled along the rivers Kampar, Rokan, Panei, and Barumun which flow to the sea on the eastern coast of Sumatra.

v2.c21.s6.t5.24

¹ Geschiedenis²—pp. 132-3; Kunst—II, pp. 412-3. T. B. G. Vol. 35 (1893) pp. 48-74.

v2.c21.s6.t5.25

Indirect evidences of the settlement of Indian colonists in these regions are not altogether wanting. One of the stūpas at Muara Takus contained, in its relic-chamber, a golden plate and a piece of stone, both engraved with Vajra and the mystic syllables in Indian scripts. A golden plate at Tanjong Medan also contained the mystic syllables and the names of Dhyāni Buddhas written in Indian script. Inscriptions in similar scripts have also been discovered in the Karimun islands, not far from the mouth of the Kampar river.

v2.c21.s6.t5.26

Krom is inclined to associate these ruins and inscriptions with the Śailendras. He is no doubt largely influenced by the prevalent belief that Śri-Vijaya was the chief seat of Śailendra power. Bosch refers the inscriptions of Muara Takus, and also the *stūpas*, to a period not earlier than the twelfth century A. D.¹. But whatever the date may be, these inscriptions indicate a direct stream of colonisation from India and the architectural remains also seem independently to bear the same testimony.