Pramāṇa · Search · About · API

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

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

Language
English

Contents

Book VI. Art in Suvarnadvipa

v2.c21.2

the Lacedaemonians, who had been waiting for the signal, now attacked the city. The Plataeans, seeing the enemy already inside and their own men falling, and knowing nothing of the plot, were seized with panic and, thinking that the Thebans were far more numerous than they really were, surrendered at discretion. But when they discovered that the Thebans were few, they took heart and, resolving to attack them, closed the gates and fell upon them.

v2.c21.3

The Thebans, finding themselves in a trap, began to defend themselves. Twice or thrice they repelled the attacks of the Plataeans, but when the women and slaves began to shout and throw tiles and stones from the houses, and a heavy rain began to fall, the Thebans, confused and blinded, broke and fled through the city. Most of them, not knowing the way in the darkness and rain, fell into the drains or were killed by the Plataeans, who knew the streets. The gate through which they had entered had been shut by a Plataean who had found a javelin-head and used it as a bolt, so that there was no way out. Pursued through the city, some climbed the wall and jumped down, most of them perishing; others reached the gate and, finding it shut, were cut down by the women and slaves. A small party, seeing a large building with an open door, rushed in, thinking it was a gate, and were captured.

v2.c21.4

Meanwhile, the main body of the Thebans, who were to have arrived at daybreak to support the first party, had been delayed by the rain and the swollen river Asopus. When they heard of the disaster, they hurried to the city, intending to seize any Plataeans who might be outside the walls. The Plataeans, suspecting their intention, sent a herald to them, warning them not to harm the prisoners they had taken, and threatening to kill them if they did not withdraw from their territory. The Thebans, upon this, retired.

v2.c21.5

The Plataeans, having thus cleared the city, collected their dead and the prisoners. The prisoners, numbering about one hundred and eighty, were executed the next day, despite a message from the Thebans claiming that they had surrendered on terms. The Plataeans denied this, asserting that they had promised to spare them only if they did not commit further acts of violence.

v2.c21.6

Thus ended the attempt of the Thebans upon Plataea, an event which was the immediate cause of the war between Athens and the Peloponnesians. The Plataeans, having secured their city, sent a messenger to Athens with the news, and the Athenians, fearing for the safety of the Plataeans, sent a garrison to their aid and removed the women and children to Athens.

v2.c21.7

BOOK VI

Chapter I. The Architecture of Central Java

§ 1. General Classification

v2.c21.s1.t1.1

Apart from the pre-historic artifacts described above in Bk. I Ch. III, the earliest monuments in Java belong to its central region and may be referred to the period during which that region played the supreme political rôle in the island. It would be wrong to infer from this fact that art did not develop in the earlier Hindu colony of Western Java over which Pūrṇavarman ruled nearly three centuries before. As regards Eastern Java, we have positive evidence that it possessed temples which were older than those of Central Java.¹ But the monuments of Western Java, even if there were any, have vanished altogether, without leaving any trace behind, while nearly all the existing temples of Eastern Java belong to a much later period. The disappearance of these old monuments may be partly accounted for by supposing that they were built of wood or other perishable materials. As regards Eastern Java, which was in continued occupation for nearly five hundred years after the fall of Central Java, the older structures were possibly replaced in many cases by later buildings. In any case, the fact remains that so far as the actual remains go, the earliest monuments in Java belong to the Central region and do not carry us further back than perhaps eighth century A.D.

v2.c21.s1.t1.2

¹ For example, the construction of a temple of Agastya is referred to in the Dinaya inscription, dated A.D. 760. Cf. Part I, p. 248, fn. 3.

v2.c21.s1.t1.3

This fact is extremely unfortunate, as it hides from our view the early and formative stage of that Indo-Javanese art of which we find the fullest development in the classical monuments of Central Java. At the time when these were erected, more than four centuries had elapsed since the first Indian colonists had settled in Java. During this long period, the principles of Indian art which these colonists had introduced in their new home must have taken deep root in the soil, and undergone process of evolution. How far these modifications are inevitable consequences of normal growth amid strange surroundings, and how far they are to be attributed to a deliberate working of local factors, it is at present impossible to say. But certain it is that when this developed form of art first comes into our view, it can hardly be regarded as purely Indian. Although its essential characteristics are unmistakably Indian, it has already developed into a type which neither actually exists nor probably ever existed in India. In other words, we come across a full-fledged Indo-Javanese art, and that at its very best.

v2.c21.s1.t1.4

The history of this Indo-Javanese art is thus written in decay,—to quote a familiar phrase of Fergusson. It may be conveniently divided into two broad periods. First the Central-Javanese period ranging from 700 to 900 A.D., and secondly, the East-Javanese period from 1200 to 1500 A.D. The art of each of these periods may again be sub-divided, on sectarian, chronological or regional basis, though none of these principles is entirely satisfactory. For the sake of convenience of treatment we shall study it under the following scheme, which is more or less a combination of all the three.

v2.c21.s1.t1.5

I. Central Java.

v2.c21.s1.t1.6

(a) Hindu monuments in Dieng Plateau and its neighbourhood.

v2.c21.s1.t1.7

(b) Buddhist monuments in the Prambanan Valley and Kedu plain.

v2.c21.s1.t1.8

(c) Hindu monuments in the Prambanan valley.

v2.c21.s1.t1.9

(d) Monuments in other localities.

v2.c21.s1.t1.10

II. Eastern Java.
(a) Early Monuments.
(b) Monuments of the Singhasāri period.
(c) Monuments of the Majapahit period.
(d) Post-Majapahit period.

v2.c21.s1.t1.11

III. Sculpture of Central Java.

v2.c21.s1.t1.12

IV. Sculpture of Eastern Java.

§ 2. General Plan

v2.c21.s1.t2.1

The Indo-Javanese art, like its parent art in India, may be described as the hand-maid of religion. All the monuments of this art, so far discovered, are religious structures, and religion has provided its sole aim and inspiration from beginning to end.

v2.c21.s1.t2.2

The religious structures in Java are known by the general appellation Caṇḍi. The word originally denoted any monument over the ashes of the dead, ranging from a single stone column to a complete structure. Krom suggests that the word is derived from Caṇḍikā, a name of the goddess Durgā, who is specially associated with the dead and whose temple is invariably present in a Balinese cremation ground even to-day. Howsoever that may be, it is evident that the word Caṇḍi gradually came to denote a temple in cremation ground, and ultimately any temple.¹

v2.c21.s1.t2.3

¹ Krom—Kunst, Vol. I, pp. 142-3. Stutterheim (Djawa, 1927, p. 188) thinks that ‘Caṇḍi’ in the sense of a general temple is a wrong use of the term by modern archaeologists.

v2.c21.s1.t2.4

Before describing the individual temples in Central Java, it will be convenient to begin with a general description of the essential features of its temple-architecture. Any deviations from this general plan will be noted in the description of the particular temple where they occur.

v2.c21.s1.t2.5

The Javanese temple of the standard type may be divided into three parts as follows

v2.c21.s1.t2.6

(1) Basement;

v2.c21.s1.t2.7

(2) Main Body (referred to sometimes as Body)

v2.c21.s1.t2.8

(3) Roof.

v2.c21.s1.t2.9

The basement, as well as the main body of the temple, consists, like the classic entablature, of three distinct parts which we may designate as the base, wall (the central vertical part), and cornice (corresponding to Greek architrave, frieze and cornice). The base and the cornice have also the same essential features in the basement and in the main body of the temple.

v2.c21.s1.t2.10

I. Basement. The base consists of a flat plinth supporting a bell-shaped ogee- and a roll-moulding, though occasionally one or more string-courses (fillets or mouldings) are added to give the whole thing a complex character.

v2.c21.s1.t2.11

The cornice, consisting of one or more projecting string-courses or mouldings, is enriched with ante-fix ornaments which serve to distinguish the different parts and emphasise the corners. Just below the cornice is a frieze-band usually enriched with ornaments.

v2.c21.s1.t2.12

The walls of the basement are usually divided into panels (sometimes sunken) by pilasters, and often richly ornamented with figure-sculpture, scroll-work etc.

v2.c21.s1.t2.13

II. Main Body. The main body of the temple is usually square (sometimes rectangular) on plan with a porch or vestibule in front and projections on all other sides. The vestibule leads to the only entrance of the temple which has no other opening.

v2.c21.s1.t2.14

The walls on three other sides are divided into panels as in the case of the basement, though sometimes there is a projecting niche in the centre of each. As already stated, the walls stand on a base and support a cornice both of which have the same features as those of the basement. The walls of the main body of the temple are, of course, naturally of greater height in proportion to the base and cornice than is the case with the basement.

v2.c21.s1.t2.15

III. The Roof. The roof consists of a series of gradually diminishing storeys, each repeating on a minor scale the general plan of the temple itself, *viz.*, a cubic structure with four niches on four sides. To the four corners are four diminutive turrets which again are miniature reproductions of the temple.

v2.c21.s1.t2.16

In the cases of some great structures, the upper stages of the roof are made octagonal, instead of rectangular, in order to relieve the monotony. The roof was crowned by an ornament, the exact nature of which is uncertain, as no single specimen has been discovered *in situ*. Judging from the fragments and also from the corner turrets, it appears to have been of the shape of a cushion with a high and pointed finial. Sometimes the cushion supported a bell-shaped *stūpa* ending in a pinnacle.

v2.c21.s1.t2.17

The interior of the temple is a plain square chamber. Its walls, rising vertically up to a certain height, support a series of horizontal courses of stone which, projecting one in front of the other, form an inverted pyramid of steps, and is terminated by a high and pointed hollow cone,—the whole corresponding to the storied pyramidal roof outside. Stucco or plaster work is found in some temples. In some cases it was part of the original plan, but in others it is a later addition. Sometimes even colour or paint was used. The decorative ornaments which consisted of well-known Indian motives such as rosette, garland, floral scroll, arabesque, various naturalistic designs and floral or geometrical patterns will be described in connection with individual temples. It may be noted here that all these patterns are derived from India and there is no trace of local flora or fauna. One very frequently occurring motive, however, deserves particular notice. The Dutch archaeologists describe it as Kāla-Makara. It really consists of two separate motives Kāla and Makara though sometimes they are found united to form a single combined motive. The Kāla-head (also called Banaspati) is shaped like the head of a monster, and is taken to be an effigy of the terrible god Kāla. But it is really derived from the Indian motive of lion’s head and Coomaraswamy rightly describes it as a grotesque Kīrttimukha¹. It is a conventional lion’s head with protruding eyes, broad nose, thick upper lip with two big projecting teeth on two sides. The Makara is a favourite motive of ancient Indian art. There are differences of opinion regarding its original significance. According to Vogel, it was originally a representation of crocodile. On account of a strange combination of elephantine and fish-like properties Grunwedel describes it as a 'sea-elephant'². In Java the Makara has the body of a fish and the head of an elephant, and is often developed into a floriated ornament.

v2.c21.s1.t2.18

¹ Krom—Kunst, Vol. I, p. 162; Indian Art, pp. 60-61; Coomaraswamy, pp. 202-3.

v2.c21.s1.t2.19

² Indian Art, pp. 60-61; Grunwedel—Buddhist Art, p. 57. Stutterheim regards Makara as a zodiac sign (I. A, L. Vol. III, pp. 27 ff.)

v2.c21.s1.t2.20

The Kāla-head is usually placed right over a gateway, or above a niche, in the centre of the enclosing arch, ending in a Makara-head at its two ends. Makara-heads are also placed at the foot of door-jambs. These motives are also found at the top of the staircase and various other parts of the building. In short, the Kāla-Makara motive, combined, or separated into two elements, occurs almost everywhere, and is one of the most favourite decorative devices of Indo-Javanese art. Though its elements are originally derived from India, it is, according to vogel, "the outcome of an indigenous combination and development³," and should be regarded as Indo-Javanese rather than Indian. This may be doubted, as the complete design also occurs in India⁴.

v2.c21.s1.t2.21

³ Indian Art, p. 62.

v2.c21.s1.t2.22

⁴ J. G. I. S., Vol. II, pp. 38 ff.

v2.c21.s1.t2.23

The decorative reliefs and sculptures in Central-Javanese temples were always restrained and in perfect correlation with the general architectural plan. They were designed to accentuate and emphasise the outline, rather than overshadow it by exuberance, or distort it by wrongly conceived ornamental patterns.

v2.c21.s1.t2.24

Two peculiarities of Javanese temples may be noted here. In the first place, columns and pillars are wholly lacking. Secondly, the arches are all constructed on horizontal principle, as in ancient India, and the true or radiating arches are conspicuous by their absence.

v2.c21.s1.t2.25

These two peculiarities made it impossible for the Indo-Javanese architects to cover large spaces by means of big arches and domes. Hence they could not plan any temple on a very big scale. They sought to make up for this deficiency, either by developing the projecting bays on the three sides into separate side-chapels, with separate roofs, or by surrounding one main temple by a large number of subsidiary temples, so as to give a monumental or colossal character to the whole structure.

v2.c21.s1.t2.26

Individual temples or groups of temples were surrounded by an enclosing wall. It is also likely that the gate of this enclosure was covered by a structure like south-Indian Gopuram. No trace of such a structure has, however, been found in Central Java. Its existence in Eastern Java, in the later period, is proved by actual remains and literary descriptions, and it may be presumed that this element of architecture was also taken from Central Java.

v2.c21.s1.t2.27

The stone used for the temples is usually trachyte, but a softer kind of sandstone or marlstone has been used in Lara-Jonggrang, Plaosan, and Sajivan temples. Mortar was used in the upper part of the temples, though not in the base, where the shaped and dressed stones were simply set side by side and kept in their position by various devices of connecting-stones.

§ 3. Dieng Temples

v2.c21.s1.t3.1

It is generally assumed by the Dutch archæologists that the groups of temples on the Dieng plateau are the oldest temples now extant in Central Java.¹ The arguments which may be brought forward in support of this view are based, in the first place, on a number of inscribed stones which formed part of the temples, and an inscribed gold plate which belonged to the treasury of a temple; and secondly, on the architectural motives of decoration. The alphabets of the inscriptions are similar to what we find on one bearing a date 731 (=809 A.D.). The decorative motives also belong to the earliest phase of Indo-Javanese art. On these grounds the temple-groups of Dieng are regarded as contemporaneous with, if not earlier than the Caṇḍi Kalasan which was built in 778 A.D. It must be admitted, however, that these arguments cannot be regarded as conclusive.

v2.c21.s1.t3.2

¹ But vogel holds a different view; cf. Indian Art, p. 68. For a description of the Dieng temples, cf. Krom-Kunst, I, pp. 165-204. Plates 1-4.

v2.c21.s1.t3.3

The plateau of Dieng, called in old time Dihyang, has a height of about 6500 ft. and is surrounded by hills on almost all sides. About the centre of this plateau, standing in a line north to south, are four temples called Caṇḍi Arjuna, Caṇḍi Srikaṇḍi, Caṇḍi Puntadeva, and Caṇḍi Sembadra, all facing west, while there is a fifth, Caṇḍi Semar, which lies opposite to Caṇḍi Arjuna, and faces east. That these five temples belonged to the same group is indicated by the fact that they were all surrounded by one enclosing wall.

v2.c21.s1.t3.4

There are three other isolated temples on the same plateau, called Caṇḍi Dvaravati, Caṇḍi Gaṭotkaca and Caṇḍi Bima. In addition to the remains of other temples, Caṇḍi Parikesit, Caṇḍi Nga-Kula Sadevo, there is a long underground tunnel, which served as a drain-pipe of the Arjuna group of temples.

v2.c21.s1.t3.5

The names of five of these temples are borrowed from the heroes and heroines of the Mahābhārata, Puntadeva, being a Javanese name for Yudhiṣṭhira. These names have, of course, been given in later times, and although they indicate the popularity of the great Epic, they do not throw any light on the original nature of these temples, any more than do the similar modern names of the temples of Māmallapuram.¹

v2.c21.s1.t3.6

¹ Failure to recognise this obvious point has led Mr. Havell to the view that the temples were dedicated to the Pāṇḍava heroes, and he has accordingly interpreted the human heads in Caṇḍi Bima as those of the great hero Bhīma. This is, of course, quite wrong. cf. Vogel in J. R. A. S., 1917, pp. 371-376.

v2.c21.s1.t3.7

The Dieng temples, generally speaking, belong to the standard type described above. The ornament was very restricted and simple, being composed mostly of plain courses, pilasters, and ogee mouldings. The projections on three sides contained niches. The lotus-cushions in these niches with holes in the centre prove that they once supported images of gods which have now disappeared. These niches, the door of the vestibule leading to the temple, and niches on two sides of it (where they occur) on the front wall of the temple, were all crowned by Kāla-Makara ornament.

v2.c21.s1.t3.8

The Caṇḍi Puntadeva has finer decoration than the rest. The ante-fix ornaments are more profusely used both in the basement and the main body of the temple. The niche in the centre of each of the three walls covers its entire length and has an ornamental base and border-frame. At about two-thirds of its height, there is a sort of upper niche, whose decorated base with an ante-fix, ogee mouldings, and cushion cover the entire upper part. The walls are enriched with ornamented corner-posts and decorated pilasters. The first or the lowest stage of the roof, which alone is preserved, almost exactly repeats the main features of the temple and is adorned with corner towers. There are two flights of steps, each enclosed between stone curbs with volute ends. One of them leads to the top of the basement and the other continues the ascent up to the door of the vestibule. The vestibule repeats the features of the main temple on a smaller scale, but has a separate and lower roof composed of two slanting panels from a central rib, and having, of course, no stages above.

v2.c21.s1.t3.9

Among the Dieng temples Caṇḍi Bima (Pl. 1, Fig. 1) or Vṛkodara deserves more particular notice as it differs in many respects from the standard type. The ground plan is almost square, the front and back sides being 16 and the other two sides about 14 ft. in length. There are projections in the centre of the three side-walls while the vestibule is fairly large, with a projection of about 6½ ft. and width of about 11½ ft.

v2.c21.s1.t3.10

The niche, on one side, is provided with Kāla-Makara ornament, and the side-posts of all the niches are ornamented; but otherwise, the flat walls of the three sides are severely plain. The cornice rests on a double-ogee moulding, the upper of which is teethed or indented with small beam-ends, while below it, and just above the flat panels, there is a projecting course decorated with garland motive.

v2.c21.s1.t3.11

The special characteristic of the temple, however, lies in its roof. The first stage of the roof is a repetition of the base-mouldings, the upper ogee moulding being decorated with double lotus-petal motive. Above the cornice of this is a plain ogee moulding, a low flat band, and an upper ogee decorated with double lotus-petal. We have thus a repetition of the main features of the first stage though on a very small scale.

v2.c21.s1.t3.12

The second stage of the roof which springs up immediately above these, is characterised by a double row of three arched niches on each side, one such double row being on the projection, and one on each side of it. The niches on the side also occupy the corner, thus leaving very little of the plain part of the roof exposed. The four niches in each corner may thus be viewed as a whole as forming a corner turret, and it is surmounted by a very short round column ending in two round courses. From the lower of these hangs a garland, and the upper supports a lotus cushion.

v2.c21.s1.t3.13

The third stage of the roof is similar to the first stage, though the upper ogee moulding has got a single, not double lotus-petal, with a garland hanging beneath. But the most striking difference consists in a double row of niches on the projections which, together with the two niches of the second stage and those of the upper one, form a successive series of niches vertically placed one above another.

v2.c21.s1.t3.14

The fourth stage was a repetition of the second stage with slight modifications. The upper part of the roof has completely disappeared.

v2.c21.s1.t3.15

The arched niches are of horse-shoe pattern and resemble the caitya-windows of India. They have plain borders and contain human heads or rather busts in high relief ( Pl. XII, Fig. 2 ) or some ornamental motives.

v2.c21.s1.t3.16

The general appearance of the roof is very different from the normal type, and may be regarded as unique.

v2.c21.s1.t3.17

Dr. Coomaraswamy takes the corner turrets to be three-quarter ribbed *āmalakas* and suggests that a complete *āmalaka* crowned the summit. In his opinion 'the roof structure corresponds exactly with that of a typical Indo-Aryan Śikhara, such as that of the Paraśurāmeśvara at Bhuvaneśvar.'¹ It is difficult to subscribe to this view, at least without a great deal of reserve. As he himself admits, the roof "consists of diminishing horizontal stages" and thus preserves the essential features of Dieng Temples and Dravidian style.

v2.c21.s1.t3.18

¹ Coomaraswamy, p, 202.

v2.c21.s1.t3.19

The vestibule repeats the features of the main building on a small scale. It had a separate roof with two stages corresponding generally to the first two stages of the roof of the main building. The top of this roof covered from view only the first stage of the main roof. The doorway in front of the vestibule cuts through the mouldings of its basement and reaches down to the plinth.

v2.c21.s1.t3.20

It is not necessary to give a detailed description of the other temples of Dieng. Although they are comparatively small in dimensions, their simple and clear outline, and restrained but well-conceived decorations endow them with a special importance. The sculptures are also characterised by a simplicity and vigour worthy of the temples which they adorned.¹ On the whole the art of Dieng is characterised by a sobriety and dignity which reminds us of the Indian temples of the Gupta period.

v2.c21.s1.t3.21

¹ Detailed discussion of sculpture will be found in Chapter II.

v2.c21.s1.t3.22

The images found in the plateau of Dieng belong exclusively to the Brāhmanical pantheon. We have images of Śiva, Durgā, Gaṇeśa, Brahmā and Viṣṇu. The temples were thus Brāhmanical, and to judge from the extant remains, mainly of Śaivite character. The representation of the three chief gods with their riders is very important from iconographical point of view. For the riders have all human bodies, and only the head of a bull or a swan, or the beak of a bird on the human body indicates the riders of Śiva, Brahmā and Viṣṇu.

v2.c21.s1.t3.23

A general view of the architectural remains of Dieng indicates that it was a holy place of pilgrimage rather than a town with a considerable number of settled population. Beyond the normal requirements for priests and servants of the temples and the occasional rush of pilgrims, there are no vestiges of a large number of dwelling houses, and there is hardly any space for the accommodation of a large population. The Dieng or Dihyang, to call it by its old name, thus resembled the Jaina City of Temples in India, such as Śatruñjaya or Palitānā. We learn from the Chinese accounts that the king of Java visited it once a year, and presumably the nobles and other classes of people made occasional pilgrimages to this holy place.

v2.c21.s1.t3.24

To the south and east of the Dieng Plateau we come across the ruins of a large number of ancient temples. To judge from a number of inscriptions scattered about in the neighbourhood, they belong probably to the middle of the ninth century A.D. The isolated remains also indicate that they were mainly Śaivite in character. These temples, built on the standard type, offer sometimes examples of exquisite decoration.

v2.c21.s1.t3.25

Among these, Caṇḍi Pringapus¹ has been rendered famous by the enthusiastic description of Brandes. It is a small beautiful temple measuring about 16½ ft. by 15 ft. It is of the standard type save that the basement consists merely of a plinth and plain walls without any moulding, and that there are no projections on the three sides of the temple. The walls are, however, divided into three panels by pilasters, and the sculptures on these panels as well as those on the front constitute the excellence of the temple. Brandes has compared the decoration of this temple with that of Mendut and emphasised the fact that artistic principles of construction did not differ according to the sectarian character of the structures. In conclusion attention may be drawn to the repetition of cornices, an architectural peculiarity which played such a prominent part in later East-Javanese art.

v2.c21.s1.t3.26

¹ Krom—Kunst, Vol, I, pp. 210-215, Pl. V. Brandes—Oudheid-kundig Rapport, 1903, p. 10.

v2.c21.s1.t3.27

Another interesting group of temples is that on Mount Ungaran known as Gedong-Sanga² which lies to the east of Dieug and south of Semarang. The name Gedong-Sanga means nine houses and no doubt refers to the nine groups of temples which are now found on two ranges of hills, separated by a deep ravine which to-day forms the boundary between Semarang and Kedu. It is a difficult question to decide whether all the temples belong to one series. The difficulty of communication between them, caused by the depth of the ravine, would, no doubt, incline us to regard them as two different groups. But the similarity of architecture is in favour of regarding them all as belonging to one and the same series. It is conceivable that in old times the communication was rendered easier by means of a sort of staircase which is now lost.

v2.c21.s1.t3.28

² Krom—Kunst, Vol I. pp. 224-239, Pll. 6-7. Detailed descriptions of the recent explorations are to be found in O. V. 1916, 1917 etc.

v2.c21.s1.t3.29

The temples of Gedong-Sanga are not arranged in a definite plan like those of Dieng, but lie scattered about on different plateaus without any symmetery of plan. They belong to the standard type and need not be described in detail. They were mostly Śaivite in character, but some were devoted to worship of Viṣṇu. Among the verious images found in these temples particular mention may be made of a new form of Durgā. Instead of standing she sits on the bull, and holds the Asura, not by the hair, but by the neck.

§ 4. Buddhist monuments in the Prambanan Valley

v2.c21.s1.t4.1

From the Dieng Plateau and its neighbourhood we come to the Prambanan valley¹ which forms the rich treasure-house of the products of the art of Central Java. It is dominated by a hill-range in the east and watered by the Opak river which runs across it from north to south, while a smaller river, Sarageḍug, runs south from the hill. The Prambanan valley proper lies to the north and west of the hill-range, along the Opak, while the Sarageḍug constitutes a separate valley named after it. These two valleys and the hill range itself contain a large number of Hindu and Buddhist temples belonging to different types and different periods.

v2.c21.s1.t4.2

¹ Krom-Kunst, Vol. I, pp. 240 ff, Pl. 9.

v2.c21.s1.t4.3

The region stands on the border of the modern districts of Jogyakerta and Surakerta. It was a seat of one or more cities or capitals, and not merely a city of temples like Dieng.

v2.c21.s1.t4.4

Of the Buddhist temples in the Prambanan valley the famous Caṇḍi Kalasan² (Pl. II, Fig. 2) deserves to be mentioned first. An inscription dated 778 A.D.³ informs us that the temple was built by a Śailendra king and dedicated to the Buddhist goddess Tārā. The image of Tārā which must have once occupied the shrine has now disappeared.

v2.c21.s1.t4.5

² Krom-Kunst, I, pp. 257-264, Pll. 11-12.

v2.c21.s1.t4.6

³ See above, Bk. II, Chap. I.

v2.c21.s1.t4.7

The inscription also refers to the gift of the village Kalasa, a name preserved in modern Kalasan, a neighbouring village. Hence the temple is referred to as Caṇḍi Kalasan. As we can assign a fixed date to this structure, it forms one of the few definite landmarks in the history of Indo-Javanese architecture.

v2.c21.s1.t4.8

The basement of the temple, which is in a very damaged condition, offers the normal type *i.e.* square in shape with projections on four sides. The temple stands on this basement leaving a margin on all sides, about 15 ft. wide, which obviously served as a processional path. The little that is left of the base of the temple shows that it conformed to the normal type and its central vertical part was decorated.

v2.c21.s1.t4.9

The main body of the temple is also square on plan with projections on four sides. These are formed into four chapels accessible by a double flight of stairways, one from the ground to the top of the basement, and another from the basement-terrace to the door of the chapel itself. The chapel on the eastern side is really an ante-chamber leading to the main cella in the centre, but the other three chapels do not communicate with it, and are really subsidiary and independent shrines.

v2.c21.s1.t4.10

The door of the chapel forms the most striking feature of each of the side-walls. A fantastic Kāla-head of colossal proportions forms the crowning piece of the decoration above it, and extends over the entire upper part of the wall up to the point where the roof begins. The hair of the monster is conventionally arranged in various curls in the shape of a pyramid. From the terrible mouth of the monster hang down five lotus-flowers, and out of its two corners appear two lions from which issue two flowering branches ending in the usual Makara on the two sides of the doorway.

v2.c21.s1.t4.11

The actual door-opening is set in a sort of niche. The two pilasters of the niche are crowned by two Gaṇa figures who support with their hands a beautiful framework above the double-curved arch of the niche. This entire framework is nicely carved with leaf-scroll and birds. It has a central piece with pointed ends, both above and below. The lower end divides the arch of the niche into two segments and the upper one almost reaches the lotus-flowers hanging from the mouth of the monster. This central piece may also be regarded as a conventionalised pattern of Kāla-Makara ornament. Immediately above the doorway, set within this niche, is the figure of a goddess holding a lotus-flower in each hand.

v2.c21.s1.t4.12

The decorated door-frame, described above, covers only a narrow vertical portion in the centre of the wall. At the two ends of the wall, and extending over its entire length, are again two vertical decorations of recalcitrant spirals within two plain flat posts. The portion of the wall, on each side, between this decorated corner-piece and the central doorway, is plain above, with a richly ornamented human figure in a niche below, just above the Makara figures on the two sides of the doorway.

v2.c21.s1.t4.13

The two side-walls of the projection as well as the remaining parts of the main wall on either side of the projection are decorated in the same style. Two long but narrow recalcitrant spirals within two flat posts form the two vertical borders of the broad central part. This central part has an arched niche, whose opening covers its lower half, and the decorations above the opening, the upper half. The two pilasters on two sides of the niche support a beautiful framework. Its lower part is formed by the Kāla-head whose two ends, immediately resting on the pilasters, are formed by two pairs of Makaras, each pair looking in opposite directions and projecting a little beyond the top-piece of the pilaster. Above the Kāla-head is the representation of a richly decorated temple, emerging out of a mass of cloud below, in which are found floating two divine beings on each side with lotus-flowers in their hands.

v2.c21.s1.t4.14

The niches must have contained figures which have, however, all disappeared, only the footstools remaining in situ. The holes visible above and beneath the niches indicate the existence of some sort of screens which were necessary to protect the images from rain and sun.

v2.c21.s1.t4.15

The upper end of the decorated walls contains an architrave of three flat bands supporting a richly decorated frieze, and lastly the cornice, with ante-fix ornaments. The decoration of the frieze consists of a row of dwarfish Gaṇas below, and a row of garlands and rosettes above. The ante-fix ornaments of the cornice are Gargoyle heads in the angles, and figures of goddesses in the centre.

v2.c21.s1.t4.16

The roof begins immediately above the cornice and at first follows all the windings of the walls below. Its base, similar to that of the temple itself, supports a row of decorated niches, alternately larger and smaller, most of which contained Buddha images. These belong to the class of the four Dhyāni-Buddhas, Akṣobhya, Ratnasambhava, Amitābha and Amoghasiddha, facing the appropriate directions, as in Barabuḍur.

v2.c21.s1.t4.17

The second and third stages of the roof are octagonal and carry niches with images of Dhyāni-Buddhas.

v2.c21.s1.t4.18

From the middle of the third stage arises the crowning ornament of the roof, viz., a large bell-shaped stūpa with a high pointed pinnacle.

v2.c21.s1.t4.19

We may now turn to the interior of the temple. The chapels on the three sides are about 11½ ft. square. The walls rising straight up to a projecting cornice, approach one another by means of projecting steps, and then, rising up for some distance like a perpendicular boiler, ultimately end in a point. Each of the side-walls has a niche, about 3½ ft. wide, decorated with Kāla head and scroll-work, and above it two divine figures on two sides. The back-wall contains a footstool which has place for three images.

v2.c21.s1.t4.20

The chapel on the east side is naturally somewhat different. It measures about 11 ft. by 8 ft. and has three small niches on each of the side walls. The back-wall has an opening which gives access by four steps to a door-opening above, from which descending again by three steps one comes to the main cella.

v2.c21.s1.t4.21

This chamber is about 24½ ft. square. The plain walls rise straight up to the cornice, where in each corner a big projecting angular stone gives the room an octagonal appearance. Then the walls approach one another by projecting steps, rise perpendicularly high for some distance, and then again inclining towards one another end in a top-piece just under the bell-shaped stūpa of the roof. In the north and south walls, at the height of about 13½ ft. from the floor are two small apertures, about 4 ft. wide, which opened into the side-chapels and were no doubt meant for purposes of ventilation.

v2.c21.s1.t4.22

The back-wall of the cella is entirely covered by the throne which once contained an image of Tārā,¹ as we learn from the Kalasan Inscription. On two sides of the throne are the usual motifs, a lion standing on a half-lying elephant, while the horizontal top-frame of the back of the throne has two Makaras on two sides.

v2.c21.s1.t4.23

¹ Mr. Schnitger says that the throne must have been meant for a person with hanging legs, and so it was for a male god, and not for the goddess Tārā; as such the Kalasan inscription cannot also belong to the temple. (Oudheidkundige Vondsten in Palembang, Bijlage B, 2nd Ed, p. 3 notes).

v2.c21.s1.t4.24

Finally, it may be noted here that the building was covered by stucco paint. That this was original part of the plan, and not a later addition, clearly appears from the fact that the decorations and sculptures were not fully worked out as the finishing touch was to be given by the stucco-work.

v2.c21.s1.t4.25

About half a mile to the north of Caṇḍi Kalasan lies the Caṇḍi Sari or Caṇḍi Bendah² (Pl. II, Fig. 1). It is a rectangular building about 19 yds. by 11 yds. with the longer sides running north-south, and the main entrance at the middle of the eastern side. The high basement and the lower portion of the structure is very much damaged, and the vestibule which stood in front of the main entrance has entirely disappeared.

v2.c21.s1.t4.26

² Krom—Kunst, I, pp. 264-269, Pl. 13.

v2.c21.s1.t4.27

It is a two-storeyed building, the upper storey being set back a little behind the lower. The division is marked by a projecting cornice. A similar cornice divides the second storey from an attic or parapet wall on the top, which, being adorned with niches and gateways, looks like a third storey from below. This is crowned by a number of Dāgobas with high pinnacles.

v2.c21.s1.t4.28

The centre of the eastern face of the first storey is covered by an ornamented doorway. Leaving some blank space on either side—where evidently the vestibule was attached to the main building—the rest of the wall on each side is decorated by an ornamented window-opening flanked by two decorated panels. The other three faces are similarly decorated, there being two window-openings on the northern and southern faces and three on the western.

v2.c21.s1.t4.29

The gateway has the usual Kāla-Makara ornament, the Makaras resting on kneeling elephants, and having birds in their open jaws. The window-opening is flanked by Kinnaras and supports a richly decorated framework ending in two Makaras. The panels contain human figures, a male on the one and a female on the other side of the window-opening.

v2.c21.s1.t4.30

The decoration of the second storey resembles that of the first, only a window-opening takes the place of the central door on the eastern face, and there are two small niches on two sides of it with a divine figure in each.

v2.c21.s1.t4.31

In the parapet walls above the second storey, there are three niches corresponding to the three window-openings below, and two additional smaller niches, corresponding to those of the floor beneath. These niches are empty, but perhaps the bigger ones with heavy decorations on the top contained the images of Dhyāni-Buddhas.

v2.c21.s1.t4.32

The richness and variety of the decorations of the niches and windows constitute the most striking feature of the temple.

v2.c21.s1.t4.33

The doorway leads to a rectangular chamber divided into three compartments, 11½ ft. by 19 ft. each, by walls over 3½ ft. thick. There were communicating doors between these rooms and in the left hand chamber was the staircase for going to the upper storey. All the rooms had niches, and the two side-rooms also window-openings. There was a broad altar-like bench against the back-wall of the three rooms on the ground-floor, and probably they contained images of Buddhas and Bodhisatvas.

v2.c21.s1.t4.34

The exact nature or purpose of the building is difficult to determine. The lower floor was evidently used as a temple. The upper floor is generally believed to have been used for residential purposes. The fact that there are arrangements for opening and shutting the windows lends support to this view, though it generally offends religious sentiments that human beings should live above the heads of the gods.

v2.c21.s1.t4.35

The Caṇḍi Sari may thus be regarded as a combined temple and monastery. Some scholars hold that it is the monastery referred to in the Kalasan Inscription. Krom, however, refuses to accept this view, partly on account of its distance from Caṇḍi Kalasan, and partly on the ground that Caṇḍi Sari can by no means be regarded as exclusively a monastery.

v2.c21.s1.t4.36

Next to Caṇḍi Sari the two groups of temples, Caṇḍi Lumbung and Caṇḍi Sevu deserve our attention. The Kelurak inscription,¹ dated 782 A.D., was found in the neighbourhood, and must have originally belonged to one of them, probably the former.

v2.c21.s1.t4.37

¹ See above Bk. II, Chap. I.

v2.c21.s1.t4.38

The group called Caṇḍi Lumbung² consists of the main temple and sixteen others around it, arranged in the form of a square with five on each side. The main temple belongs to the standard type, but the others have no projection except on the front. Neither the size nor the decoration of these temples mark them out as prominent ones.

v2.c21.s1.t4.39

² Krom—Kunst, I, pp. 272-4, Pl. 14.

v2.c21.s1.t4.40

Far different, however, is the case of Caṇḍi Sevu the biggest Buddhist sanctuary except Barabuḍur³ (Pl. II, Fig. 2).

v2.c21.s1.t4.41

³ Krom—Kunst, I, pp. 274-294. Pll. 15-18. O. V. 1915 (141), 1917 (42-45, 76-78), 1919 (pp. 90 ff.)

v2.c21.s1.t4.42

A rectangular paved courtyard, measuring about 200 yds. by 180 yds., is surrounded on each side by two rows of temples, altogether 168 in number. The main temple which occupies the centre of the courtyard is similarly surrounded by two rows of temples in the form of squares, with 12 and 8 respectively on each side, thus making a total of 72. The main temple is thus surrounded by 240 temples, and there are traces of five more between the first two and the last two rows. The group thus formed had, again, one anterior temple on each side, at a distance of about 330 yds. There was thus a total of 250 temples including the main temple.

v2.c21.s1.t4.43

The main temple, situated on an elevated plane, formed the worthy centre of this vast complex of sanctuaries, each successive row of which was on a lower plane than the others. The sloping roofs of the vestibules of the first and last rows of temples accentuated the gradual slope and gave a pyramidal appearance to the whole.

v2.c21.s1.t4.44

Very little remains of the subsidiary temples, and there is no need to describe them in detail. A short inscription, engraved on one of them reads as follows: "*Mahāprattaya sang Ranggunting*" (*var*. Ragunting, Ragarunting, Ranggunting) *i.e.* "gift of Ranggunting". There are several other short dedicatory records of this kind. These make it evident that the subsidiary temples were the gifts of private persons. These short records may be referred on palaeographical grounds to the ninth century A.D.¹

v2.c21.s1.t4.45

¹ Krom refers them to the beginning of the ninth century A.D. (op. cit, p. 283) but Stutterheim refers them to the second half of the ninth century A. D. ( B. K. I., Vol. 85, pp, 494 ff. )

v2.c21.s1.t4.46

The main temple stands on a raised pavement, about 40 yds. square, enclosed by a balustrade with four gateways in the middle of the four sides. The general plan of the temple resembles that of Caṇḍi Kalasan. It is square on plan with projections on four sides, three of which form three separate chapels, while the fourth,—here, again, that on the east--forms an antechamber to the main shrine. There are, however, two important differences. In the first place the path of circumambulation on the basement does not go round the entire temple, but is interrupted by the wall of each side-chapel. Each of the side-chapels has, however, a rectangular vestibule at the back which communicates with the circumambulation path by a door, and thus one can walk round the central shrine through these vestibules. Secondly, in Caṇḍi Kalasan one and the same roof covers the main and the subsidiary chapels, while in Sevu, the four side-chapels have all independent roofs, with that of the central shrine towering above them all.

v2.c21.s1.t4.47

The vertical panels in the basement are decorated with vase and lotus motive, while the balustrade round the temple is adorned with figures engaged in dance and music. The walls of the temple are severely plain but the monotony is broken by two niches at the two ends. The walls of the side-chapels were, however, richly ornamented with complex variegated patterns including rosette-circles with flowers and animals, combination of trident and wheel etc. The frieze of the chapels just below the cornice was decorated with human figures under a band of birds and garlands.

v2.c21.s1.t4.48

A short staircase leads to the interior of the side-chapel which measures 11½ by 7¾ ft. At the back a door leads to the vestibule referred to above. On the right and left hand walls are three niches with pointed arch, one larger between two smaller ones. The pilasters of these niches rise out of a vase with lotus and end in a lion's head. In the four corners of the room, as well as in the central niches of the side-walls, are footstools for images.

v2.c21.s1.t4.49

The back-vestibule also has three niches on its back-wall (which is also the outer side of the wall of the cella), with a footstool for images in the central niche. The images have, however, all disappeared.

v2.c21.s1.t4.50

The back-vestibule on the eastern side had of course a doorway through which by means of a staircase one could reach the cella measuring about 18 ft. square. The walls of the cella are plain and the construction of the arch above was of the usual type. Its back-wall was wholly covered by the throne which once contained a colossal seated image, probably of bronze.¹

v2.c21.s1.t4.51

¹ Stutterheim describes a small bronze fragment which he believes to be part of the head of the image (B. K. I., Vol. 85, pp. 487-491 and plate).

§ 5. Buddhist monuments on the Kedu Plain

v2.c21.s1.t5.1

About the time when the older temples were erected in the Prambanan valley, or perhaps a little later, the foundation of the famous Barabuḍur marked the Kedu plain as another important centre of religious sanctuaries in Java. Numerous temples, both Brāhmaṇical and Buddhist, must have at one time clustered round this central figure. Two of them deserve special mention, as they were more closely associated than others with the Barabuḍur. These are Caṇḍi Mendut and Caṇḍi Pavon.

v2.c21.s1.t5.2

The recent "restoration" of Caṇḍi Mendut (Pl. III.) gives a somewhat different idea of the structure as a whole from what it originally was. But it is not difficult to follow the essential features of this beautiful temple.²

v2.c21.s1.t5.3

² Krom—Kunst, I, pp. 295-320, Pll. 19-21.

v2.c21.s1.t5.4

The basement measuring about 30 yds. by 26 yds. has a height of 12 ft. The projections are shallow except on the front, the north-west, where room had to be made for the porch leading to the doorway of the temple. The basement is composed in the usual way, with a nice tooth-band, between the lower ogee and roll mouldings, and several square fillets both above and below the vertical part. The vertical part contains a number of decorated panels separated by pilasters, the whole being nicely arranged with a view to produce a very good effect by the play of light and shade.

v2.c21.s1.t5.5

A flight of steps enclosed between stone curbs leads to the top of the basement. In addition to the Kāla and Makara ornaments at the top and the bottom, the curb walls of the steps were decorated with panels representing Jātaka scenes.

v2.c21.s1.t5.6

The temple occupies the centre of the basement leaving round it a passage, about 8 ft. broad. It is surrounded by a railing which has decorated panels and also twelve spouts with Makara-heads for draining out water.

v2.c21.s1.t5.7

The temple is about 15 yds. square with small projections on three sides and a large porch in front. The base of the temple has some additional courses in the lowest part, viz., a vertical plinth, divided into rectangular panels, a frieze with garland motive etc. The panels are all decorated, sometimes with recalcitrant spiral of floral design combined with birds, and sometimes with illustrations of familiar fables about animals etc.¹

v2.c21.s1.t5.8

¹ Cf. Notes by Crucq ; O. V. 1930, p. 323.

v2.c21.s1.t5.9

The walls of the temple are divided into three parts by the projections, a broad vertical compartment in the centre with two smaller ones on two sides. Each of these has a broad plain flat post with a smaller one decorated with recalcitrant spiral at each end. The intervening space is decorated by a vigorous and beautiful sculpture. It represents a divine figure with attendant worshippers, flanked by two pilasters, rising out of a vase and supporting two Gaṇa figures. The addition of trees and flowers adds infinite grace to the well conceived scenes.

v2.c21.s1.t5.10

The central panel on the north-east has an eight-armed goddess richly ornamented and seated on a lotus-cushion. Above the halo behind her head is a conventional tree richly bedecked with foliage, garlands, and umbrellas. On each side of the goddess is a human figure, also richly ornamented and with a halo, holding a lotus stalk in one hand and a fly-whisk in the other. The goddess holds in her right hands, conch-shell, thunderbolt, discus and rosary, and in the left a round object, axe, elephant-goad and probably a book. In the background are divine worshippers floating in clouds, etc.

v2.c21.s1.t5.11

The corresponding panel on the opposite side has a lotus-pond from which arise three lotus-cushions, one bigger in the centre with two smaller on two sides. On the central one supported by two Nāgas sits the chief figure, perhaps a female, richly ornamented and with a halo. Of the four hands two are placed on the lap, in dhyāna-mudrā, and the other two hold a rosary and a book. On the two side-cushions sit two worshippers, one with a flower and the other with a jewel on a lotus-cushion. On each side is a tree, with pots of money or treasure near the foot, while in the background is another big tree with umbrellas, divine figures, etc.

v2.c21.s1.t5.12

The central panel of the back-wall has the figures of Bodhisatva Avalokiteśvara in the centre, in a field as richly decorated as the other two.

v2.c21.s1.t5.13

The central panel is of course lacking on the front side, its place being taken by the vestibule.

v2.c21.s1.t5.14

The identification of the two goddesses in the central panel on the two sides is a matter of difficulty. Foucher has identified the one on the north-eastern side as that of Cundā, and the corresponding one on the opposite side has been taken to be the same goddess in her benevolent form. But the presence of Avalokiteśvara on the back-wall has induced others to regard the goddess as Tārā. The lotus-pond scene supports this view, as it agrees with the mythical legend about the birth of the goddess.

v2.c21.s1.t5.15

The eight side-panels have each the figure of a Bodhisatva in the centre. By small variations in the drapery and attributes the artist has left no doubt of his intention to represent the eight Bodhisatvas who were held in special honour in the Mahāyāna pantheon. Beginning from the panel to the proper right of the doorway these are, (1) Sarvanivaraṇa-viṣkambhī, (2) Maitreya, (3) Samantabhadra, (4) Kṣitigarbha, (5) Vajrapāṇi, (6) Mañjuśrī, (7) Khagarbha, and (8) Padmapāṇi.

First 140 of 1,006 passages. Read on

Linked texts (14)