A Study Of Wooden Structures

A Study of wooden structures: a contribution to the architectural history of the Bayon Style Monuments CUNIN Olivier Ph.D. in Architecture – Architect DPLG Centre for Khmer Studies. PO BOX 9380, Wat Damnak, Siem Reap (Cambodia) Email: [email protected] Web site: http://www.khmerstudies.org

Abstract: At the present time, the studies of Khmer architecture are based essentially on the stone skeleton of the temples. Concerning temples dated from the end of the 12th century, different researches have found out that the current state of the stone skeleton is the result of a sequential addition plan. Meanwhile, the interpretation of this evolution process and the use of the monuments based on the chronological sequence of the stone skeleton can lead to an erroneous understanding. Indeed, these temples have a lot of traces of ancient wooden structures inside and outside of the stone skeleton. Based on archaeological analyses of the constructions of the main temples of Jayavarman VII in Angkor: Ta Prohm, Preah Khan, Banteay Kdei and Bayon, this paper will attempt to show that the wooden structures that originally complemented these temples during each state of theirs evolution bring new evidences to complete and interpret theirs architectural history as much as to assist in the understanding of the use of these monuments.

1. Introduction Traces of structures, made of ephemeral material in Khmer monuments of the angkorian period, especially those dated at the end of the 12th century and at the beginning of the 13th century, are a long time well known fact. Nevertheless, only a few studies to this date have focused on this type of structures. This is easily explained, as no evidence of wooden structure on these monuments have been collected and only traces left on their minerals skeletons prove their past existence. Therefore it was logical to prioritise the stone remains of the religious complexes. The present paper will not depart from the rules. Indeed, the main purpose of the research presented here is to complete the relative chronology of the main complexes of the architectural programme of Jayavarman VII, keeping this new studies in line with the continuity of the research programme: “from Ta Prohm to Bayon”. The first part of this programme was carried out between March 2002 and October 2004, and was the main subject of our Phd in architecture at the National Institute of Polytechnic of Lorraine (INPL). The aim of this research was to set up a relative chronology of the main monuments of the Bayon style as much in Cambodia as in Thailand, by carrying out comparative studies. Existing relative chronologies were limited to the mineral skeleton of the monuments of our large corpus of study, leaving aside for a long time, a part as important of the works which originally constituted these monuments: Structures made of ephemeral material. Although they were not included in the relative final chronologies of each monument, a certain amount of archaeological traces of wooden structures were substantially recorded and taken into account in our approach to establish this chronology1. The methodology followed to establish this chronology was based on the synthesis of three type of analysis2 (Fig.1). The archaeological study of the constructions constitutes the basis of our approach. During this study the archaeological traces of the wooden structures became useful. The second approach based on the stylistic study of the Bayon of Ph. Stern 2 concerned the ornamented and architectural décor of these temples. Finally, an archeometric analyses based on the magnetic susceptibility of the sandstone used on these monuments defined our third approach 3. This was carried out in collaboration with the petrology unit of the Japanese governmental team, JSA, and Waseda University of Tokyo.

1

[Cunin 2004: Volume 1, 284-287] [Stern 1965] 3 [Cunin-Uchida 2002] and [Uchida-Cunin et al. 2003] 2

Fig.1: Methodology used to elaborate the chronology sequences of the Jayavarman VII monuments

In order to complete the relative chronologies of Jayavarman VII monuments, we have carried on the recording work of several archeological traces of wooden structures in these monuments. This research carried out between mid-March 2005 and mid-January 2006 was made possible thanks to the INALCO/CKS grant programme financed by the Florence Gould Foundation. As there is a large corpus of study, only the results of the works on four of the most representative monuments of the Bayon style located in the archaeological park of Angkor will be presented here: the complexes of Ta Prohm, Preah Khan, Banteay Kdei and the Bayon itself. Nevertheless, before going further in presenting the data collected, some of the archaeological type of traces that are kept in our inventory have to be defined, their meanings explained, as well as their integration into existing chronologies.

2. Example of archaeological traces of wooden structures Wood has always been used in Angkorian foundations, whether directly in masonries, as it is the case at Phnom Chisor (Ph.1) an 11th century monument where remains of supporting beams for stone constructions can be found, or on stone buildings stand alone or not constructions. Traces of such buildings of wooden structures are also found in the central sanctuary of Phnom Bakeng (Ph.2), whereas in Angkor Wat several porch pillars present traces of wooden lean-tos that where adding on to it (Ph.3). Bayon style monuments have also traces of several wooden elements directly in their masonries. Nevertheless, these are always the results of an opening whether of a door or of a new window in their masonries. The lack of sandstone lintels was then compensated by the insertion of wooden beams (Ph.4). Apart from these always subsequent beams insertions, Jayavarman VII’s temples always had from their conception several buildings made of wood. Sometimes separated from the rest of the stone buildings as in the courtyard of the first-floor and the “parvis-causeway” of Bayon (Fig.2), they were mostly directly added to their mineral skeleton. In this case on top of the usual traces on the ground we also find sockets for the purlins of their framework, built on the pediments of the building that they were completed (Fig.3). Others archaeological traces of wooden structures are specific to Jayavarman VII‘s temples. These are usually located inside the galleries and resemble arrangment of sockets aimed at the armature of a partition. Thus, we can distinguish in these monuments two types of wooden structures: on the one hand the outer wooden structures and the inner wooden structures on the other hand.

Ph.1: South inner gate of the eastern central pavilion of the first enclosure of Phnom Chisor.

Ph.2: Paving facing the central sanctuary of Phnom Bakeng.

Ph.3: Eastern pillars of the North porch of the south library of the fourth enclosure of Angkor Wat.

Fig.2: Arrangement of the potholes of the parvis-causeway of Bayon (recorded by JSA).

Ph.4: Eastern window of the south avant-corps of of the western central pavilion of the first enclosure of Ta Prohm.

Fig.3: plan of the eastern pediment of the Western entrance pavilion TP.65 of the southern cloister of Ta Prohm [Dumarcay 1973: figure 20, plate XX].

Let’s not forget, that in the monuments of the Angkorian period the false ceiling4 and the doors were made of wood. The beams, where the pins or the gudgeons were built, in order to fasten these doors, were also sometimes made of wood. Finally it was several times noticed in the monuments of the Bayon style, sockets for the beams probably used as tie-beams located in the intrados of the vault of the galleries (Ph.5), as well as in the chimneys of several towers (Ph.6). This grouping of wooden items has not been selected in our present study, as these elements, or at least their archaeological traces, rarely bring data helping to complete the relative chronologies of Jayavarman VII’s monuments5. Only outer wooden structures as well as inner ones seem likely to substantially complete the chronologies of these monuments.

Ph.5: Intrados of the vault of the second gallery of Ta Prohm.

4

Ph.6: Intrados of the chimney of the satellite temple VII of Banteay Chmar.

Remains of such ceilings are found in the first enclosure-galleries in Ankor Wat and Banteay Trop. Except supporting beams added after the opening of a door or a window, their archeological traces were taken into account during our archeological reading of the constructions of the Bayon style monuments. 5

2.1. Structures with a wooden framework As mentioned earlier the wooden inner structures were likely to be used as partitions. The outer wooden framework must have had different functions depending on their locations in the overall setting of the temples. Although we can only speculate on the shape that these stand alone buildings had, we can distinguish two types of buildings with a wooden framework, which were added on to the mineral skeleton of Bayon style temples. Figure 4, as well as pictures 7 6 and 8, present archaeological traces matching these two types of lean-tos.

Fig.4: Archaeological traces of the first and second type of outer wooden structures.

Ph.7: Example of the first type of archaeological traces of outer structure on the western facade of TP.49 (Ta Prohm).

6

Ph.8: Example of the second type of archaeological traces of outer structure on the western entrance of TP.83 (Ta Prohm).

Holes located in the upper part of the decorative lintel and of the small columns of the western entrance of the TP.49 tower of Ta Phrom are consequences of the looting of the metallic staples that were fixing these elements to the rest of the masonry of the tower.

Archaeological traces of the first type of wooden structure building that have been identified, are exclusively located on the pediments of the stone building that they were completing. As for the second type, traces are found to the lateral side of the pilasters, adding on to traces found on the pediments that they support. According to these archaeological traces and going by the work of J. Dumarçay 7, we have tried in the figure five and six to sketch8 the wooden framework of those two types of lean-tos. The first one is characterised by a simple double sided roof supported by a bending structure, whereas the second presents the same structure to which a lean-to was added on both sides.

Fig.5: Skeleton’s restitution of the outer wooden structure based on the work of Dumarçay (1973).

Fig.6: (Type 1) Skeleton’s restitution of the outer wooden structure without side lean-to. (Type 2) Skeleton’s restitution of the outer wooden structure with side lean-to.

7 8

[Dumarçay 1973: figure 21 of plate XXI] Too few elements are for the moment available in order to be able to present a more accurate sketch for the second type of lean-to.

The casing of these wooden frameworks remains unknown until today. It seemed that some edifices didn't have any casing, being used as porch or canopies connected between two stone buildings. Others must have been enclosed spaces, where rituals were performed. Their roofs were covered with baked clay tiles of which several examples have been found during the excavations of the Bayon style monuments 9. The restoration of the appearance of these lean-tos has no impact on the relative chronology of Jayavarman VII’s temples and we have overlooked this problem by focusing on the archaeological traces of their wooden framework. The main important subject here is to determine when these wooden framework edifices were built. Going by on our observations two building periods were established. These are based on the location of the archaeological traces related to the ornamental décor of the pediments where the wooden framework buildings were added on. The first configuration is remarkable thanks to an ornamental décor containing sockets for the purlin of the bending structure of the affixed lean-to (Ph.9). Therefore neither the tympan nor the pitches of the pediment show this type of sockets. This configuration can come from two phases of works, which nevertheless have the same chronological consequences. In the first case the wooden structure edifice was erected as soon as the architectural programme of the stone building was achieved. The decorative programme of the latter was then executed taking into account the wooden edifice. In the second case the architectural programme and the ornamental programme succeeded, but locations in the décor of the pediment would have been desgined during its construction, which means that the wooden edifice was planned from the beginning. Whatever the sequence of events, the wooden structure must have been contemporary to the stone building it was completing. Therefore it is considered that constructions with wooden framework presenting these archaeological traces configurations were part of the initial architectural programme of the monument.

Ph.9: Northern pediment of the tower TP.49 (Ta Prohm)

The second configuration of archaeological traces is characterised by sockets for the purlins in the tympan itself or in the pitches of the pediments, as well as exceptionally on the decorative lintels10 and the pilasters capitals (Ph.10). Also noticeable is that these sockets created gaps in the ornamental and narrative décor of the pediments, which lead us to think that they were not planned before the implementation of the decorative programme. This can only result from a wooden framework structure built after the architectural and decorative programme of the mineral skeleton. Therefore these are added on lean-tos, not originally planned.

9

Angkor Conservation files mentioned several fragments of tiles and finials during excavations at the Preah Khan of Angkor as well as at Bayon and other Bayon style temples. 10 We especially refer to the central sanctuary of Ta Prohm of Bati. It is not as in picture 10 the consequences of the looting of the metallic staples of the decorative lintel.

Ph.10: Eastern pediment of the tower TP.60 (Ta Prohm)

2.2. Inner partitioning As mentioned earlier, the archaeological traces of the inner wooden structures are characterized by alignments of sockets for the armature aimed at what should have been partitionings (Fig.7, Ph.11 and Ph.12).

Fig.7: Alignement’s example of inner wooden archaeological traces.

Ph.11: Western door of TP.104 (Ta Prohm).

Fig.12: Southern corner of the gallery TP.38 (Ta Prohm).

These partitions were made to divide the galleries in order to create an equal number of rooms which functions remains to be established. However it seems possible that this succession of small spaces could have accommodated the several statues that populated Jayavarman VII monuments. Doors were built in these partitions in order to access each room. Figures eight and nine show a reconstruction proposal by H. Arahi of an armature of these partitionings11. Nevertheless other variations must have existed.

Fig.8: Reconstruction of the wooden structure of the inner partion (based on the work of Hisao Arahi).

11

[Arahi 200: 49, Volume 2, figure IV.07]. It is a reconstruction of the armature of the partitionings of the first enclosure of Bantay Kdei. Archeological traces located in this gallery have the same features as those seen in the others Jayvarman VII’s complexes.

Fig.9: Reconstruction of the wooden structure of the inner partion (based on the work of Hisao Arahi)

When were these partitionings done? Once more our observations have led us to distinguish two implementations phases. The first one is characterised by construction starting as soon as the architectural programme was finished. The decorative programme and the renovation of the inner surfaces of the walls of the galleries where this partitionings are located were executed once these were built. Indeed, the partitionings once taken off unveil a projecting surface as well as gaps in the cornices of the walls where their armatures were fixed (Ph.13). The second implementation was done after the decorative programme and the renovation of the walls of the inner galleries; as there are no gaps in the inner cornices of the galleries as well as any interruption of the renovation of the inner surfaces of their walls at the partitionings locations (Ph.14).

Ph. 13: Gallery PK.48 of the first enclosure of Preah Kan

Ph. 14:Gallery PK.46 of the first enclosure of Preah Khan.

3. Contributions of the archaeological traces of the wooden structures in order to elaborate chronological sequences. After having presented archaeological traces of wooden framework edifice and of partitionings observed on the mineral skeleton of the Bayon style temples, as well as the chronological information that could been taken from it, we will demonstrate, using an example, the contribution of this data in the elaboration of the chronological sequences of a monument of this period. This example is located to the eastern side of the third enclosure of Ta Prohm (Fig.10). It includes the edifice TP.58, the sanctuary Tower TP.57, the “passage gallery” TP.80, the lateral entrance pavilion North TP.103 as well as the gallery TP.104 connecting with the central entrance pavilion of the third enclosure (Fig.11).

Fig.10: location of the edifices at Ta Prohm used as an example for the contribution of the study of the wooden structures to elaborate chronological sequences.

Fig.11: Example of the studies of wooden structures’ traces to the chronological sequence’s elaboration of the north-eastern part of the third gallery of Ta Prohm.

Outcomes from observations done on all this cluster of edifices starting by the sanctuary Tower TP.57 are presented here. The pediment of the south entrance of the latter show sockets for purlins of the skeleton of a wooden structure framework (1 of the figure 11). The location of these sockets in relation with the décor of the tympan and of the pitches let us think that the wooden lean-to, which was extending south of this entrance, was part of the original architectural programme of this tower (Ph.15).

Ph.15: Archaeological traces on the pediment of the southern entrance of the tower TP.57.

The pediment of the eastern entrance of this sanctuary tower show sockets for a bending structure identical to those from the pediment of the south entrance (2 of the figure 11). Their location with the tympans and the pitches of the pediment brings us again to think that the wooden framework edifice which was added on to it was part of the original architectural programme of this tower (Ph.16). This lean-to is nevertheless different from the one south of TP.57. Indeed, the southern pilaster of the eastern entrance shows a socket where one of the elements of the framework of this wooden edifice was fixed (3 of the figure 11 and Ph.17). And yet this type of traces is not seen on the pilasters of the south gate. Therefore it seems, as far as the south lean-to of TP.57 is concerned that it is a simple edifice with two sided simple roofing whereas the eastern lean-to had side constructions. Only the southern pilaster of the eastern entrance of TP.57 is nowadays accessible following the collapse of the eastern side of the edifice TP.80. Thus, this pilaster, as it is still the case for the North pilaster was hidden by this edifice before the collapse of the south side. Therefore TP.80 can only be subsequent to the tower TP.57. As the lean-to extending originally east of the TP.57 can only have coexisted with the present edifice TP.80, it was probably disassembled during the construction of TP.80. The addition of this sandstone building entailed a renovation of the décor of the surfaces where the new masonries were resting (Ph.17). Archaeological traces on the pediment of the western entrance of TP.103 (4 of the figure 11) help us further in our relative chronology proposal. Indeed, as for the pediment of the eastern entrance of TP.57, the location of the sockets of the purlins indicate that the wooden structure edifice, which was adding on to the west of TP.103, was part of its initial architectural programme (ph.18). But this pediment was facing the eastern entrance pediment of TP.57. Therefore it is likely that it is the same wooden structure edifice which is connecting between the tower TP.57 and the side entrance pavilion of TP.103. It is not known when this wooden edifice was substituted by the present “passage-gallery” TP.80, retaining its function and most likely its external appearance. Indeed from the outside TP.80 is made of two lean-tos, whereas inside it forms a single space. Let's point out that during our observations at Ta Prohm as well as at Preah Khan or Banteay Chmar we have noticed several time the same process of substituting the wooden structure edifices by their stone equivalent.

Ph.16: Pediment of the eastern gate of the tower TP.57

Ph.17: South pilaster of the eastern gate of the tower TP.57

Ph.18:“passage-gallery” TP.80 and pediment of the western entrance of TP.103.

Our observations made at TP.104 discovered the existence of a partition dividing into two spaces this section of the gallery (ph.19). A study of the immediate surrounding of the archaeological traces left by the armature of this partitioning (1 of the photo 19) invite us to conclude that it is subsequent to the architectural programme of this gallery. Indeed, the cornice as well as the renovation of the wall, where this partition was located, were already done. Originnally the section of the gallery had only one space, which was then divided into two with the help of a wooden partition probably including a door. Other traces on the mineral skeleton of this gallery help us to relocate the installation of this partition in a larger programme aiming at modifying the spatial observance of this part of the temple. Indeed we noticed that the western door of this gallery is not original. This opening done in a plain wall, necessited the laying of two wooden beams in order to compensate for the absence of a proper lintel (2 of the photo 19). The eastern door facing this opening was a window which had only its basement dismantled (ph.20). But these two doors are in line with the edifice TP.58 which masonries and style studies as well as archeometric data have revealed its posteriority compared to TP.57 and TP.103-104, while being prior to TP.80 12. It is therefore possible that the building of this edifice triggered in its extension, the creation of a new cross section in the gallery TP.104. This new access caused the laying of a partition on the northside in order to do delimit it clearly, whereas the south side presented originally a Dutch door. Finally, a reminder that the sanctuary Tower TP.57 was originnally extended onto its south side with a wooden lean-to. Although the size of this wooden framework structure remains unknown, the present edifice TP.58 does not leave a lot of space for such a lean-to. Thus, its building could have had for consequences the dismantling of the latter.

Ph.19: western gate of the TP.104 gallery

Ph.20: Eastern gate of the TP.104 gallery

Based on our observations and their conclusions supported by the already known relative chronology of the mineral skeleton, we have tried to suggest with figure 12 a synthesis of the chronological sequence of this area of Ta Prohm.

12

[Cunin 2004: 304-305, Tome I]

Originally tower TP.57 was completed by two outer wooden structure: one to the south side without lean-to and one to the east side with lean-to connected to the west entrance of the building TP.103.

Then, the southern outer wooden structure of the tower TP.57 was disassembled to build edifice TP.58. At the same time a door to the axis of the new building was opened to the gallery TP.104 and a wooden partition was installed.

Finally, the outer wooden structure between TP.57 and TP.103 was disassembled and replaced by the stone building TP.80.

Fig.12: Relative chronology of the group of edifices (TP.57, 58, 80, 103 and 104) based on the syhtnesis of the archaeological traces study observed in situ.

4. Inventory of the archaeological traces of the wooden structures in the main temples of the Bayon style The inventories of the archaeological traces of the wooden structure edifice and of the partitionings presented here are based on basic research of the mineral skeleton of the four monuments of our study. As these latter are in ruin, some parts of their mineral skeleton have completely disappeared, whereas some others are still inaccessible because of collapses. Therefore these inventories are not exhaustive. Nevertheless, as will be presented later, they still represent a sufficient base to complete the chronological sequences of these temples. The other temples of this period are also useful to set up the relative chronology, as they are showing archaeological traces of ancient wooden structures. Nevertheless the results of our research have been limited for practical reason to only four of Jayavarman VII’s religious complexes. Apart from Banteay Chmar, their involvement as monuments of smaller size is of lesser importance than the four temples studied for our paper. Colour coding associated with a pictogram was chosen in order to differentiate the type of archaeological traces that we have recorded during our field trips. We have distinguished the traces of the wooden structures on the one hand, the traces of the partitioning and of the potholes on the other hand. All our observations are summarised in the following figures. 4.1. Ta ProhmTa Prohm

Fig.13: Archaeological traces distribution in the third gallery of Ta Prohm.

Ta Prohm is indisputably Jayavarman VII’s temple presenting most of the archaeological traces of wooden structures edifices. Indeed, no other monuments of the Bayon style present such a density of archaeological traces of this type. If several sanctuary towers of the temple were at a certain period extended with one or more wooden lean-tos, the presence of traces of partitioning in this galleries is relatively small compared to the one that are found in the Preah Khan of Angkor. Also noticeable in “the hostel”13 TP.216 located between the fourth and fifth enclosure, are archaeological traces of wooden lean-tos at its extremities.

13

This type of edifice is better known as “darmaçâlâ” and corresponds to the 121 “fire houses” mentioned on the stele inscription of the Preah Khan of Angkor, found on the main roads of Jayavarman VII kingdom.

Fig.14: Archaeological traces distribution of TP.216 of Ta Prohm

4.2. Preah Kahn

Fig.15: Archaeological traces distribution in the second gallery of Preah Khan.

Fig.16: Archaeological traces distribution in the northern complex of Preah Khan.

Fig.17: Archaeological traces distribution in the southern complex of Preah Khan.

Fig.18: Archaeological traces distribution in the western complex of Preah Khan

Fig.19: Archaeological traces distribution of PK.144 and PK.170 of Preah Khan.

Contrary to Ta Prohm, Preah Khan is characterised by a huge amount of archaeological traces of partitionings in its galleries. They are mainly located in the first and second galleries (figure 15). Their density seems to decrease in the galleries of the secondary complexes (figure 16 to 18), but this is only due to the fact that they are obstructed by collapses of their supperstructures which sometimes hinder any type of observations. Although there are some wooden framework edifices in the main complex and its annexes, they are less frequent than the one seen at Ta Prohm. Nevertheless, they are still a major data for understanding the expansion of the monument. Among the edifices showing archaeological traces of wooden lean-tos is “the hostel” of Preah Kahn. There are archaeological traces of wooden framework edifices, extending on both extremities of the edifice PK.170 (figure 19). It seems that these extensions, done with ephemeral material, of the mineral part of this type of foundations, typical of Jayavarman VII’s period, are systematic. As we find the same type of archaeological traces on the extremities of the “hostel” of Banteay Chmar and of the Preah Kahn of Kompong Svay as well at those along the ancient road of the kingdom that we have visited. 4.3. Banteay Kdei

Fig.20: Archaeological traces distribution in the second gallery of Banteay Kdei.

In Banteay Kdei, traces of wooden framework edifices as well as of partitionings are located in the area of its first enclosure (figure 20). Although the edifice PK.55 located outside the third enclosure had traces of wooden lean-tos on both extremities (figure 21). In this edifice an important quantity of Buddhist images were discovered in 2001 during archaeological excavations carried out by the international mission on the scientific studies of Angkor vestiges (Sophia mission, Japan).

Fig.21: Archaeological traces distribution of BK.55 of Banteay Kdei.

4.4. Bayon

Fig.22: Archaeological traces distribution in the Bayon complex.

Apart from individual edifices originnally located to the eastern causeway-parvis as well as in the courtyard of the first level, traces of wooden framework edifices and of partitionings are mainly located in the Bayon in its gallery of the second level (figure 22 and 23).

Fig.23: Archaeological traces distribution in the second level of the Bayon complex.

5. Insertion of the wooden structures in the relative chronologies of the mineral skeleton of the temples Similar to the examples that we have just seen, we have tried based on the inventory of the archaeological traces of the wooden structures and on the study of the immediate surrounding of their location, to complete a relative chronology of the mineral skeleton of the four temples of our study. Due to the importance of these monuments, it is impossible here to go into details for each of them and to justify the insertion of the wooden structures as proposed. We will therefore, for the moment, only give the results of our research using graphs along with the first observations resulting from the comparison of these completed chronologies. As an introduction for each temple, we will show graphs of the relative chronology of their mineral skeleton resulting from our prior works in order to better apprehend the insertion of the discovered wooden structures.

5.1. Ta Prohm

Fig.24: Synthesis of the relative chronology of the fourth enclosure of Ta Prohm’s stone skeleton.

Figures 25 and 26 show that the mineral skeleton of Ta Prohm was from its initial architectural programme completed with wooden structures. These will then be replaced during subsequence construction phases by their equivalent in stone, as was the case for the eastern cloister gallery during the fourth phase of building (figures 29 and 30), or the edifice TP.80 as on our previous example. Note that the central sanctuary was originally connected to the eastern and western pavilions of the first enclosure-gallery. We find the same type of connection between the central sanctuary and its first enclosure in Banteay Kdei as well as in Banteay Chmar and in several other monuments of the same period. Other parts of the mineral skeleton of Ta Prohm originnally without wooden extensions will then be afterwards completed with such structures. The most remarkable extension of this type is the second enclosuregallery which was more than likely built in wood during the fourth phase of the temple construction (figure 29 and 30), then replaced by the stone gallery as it stands nowadays during the sixth and last phase of building (figure 31 and 32). We find the same process of substitutions for the second enclosure of Preah Kahn and of Banteay Kdei.

Fig.25: Distribution of the wooden structures in the second phase of the Ta Prohm’s fourth enclosure.

Fig.26: Distribution of the wooden structures in the second phase of the Ta Prohm’s third enclosure.

Fig.27: Distribution of the wooden structures in the third phase of the Ta Prohm’s fourth enclosure.

Fig.28: Distribution of the wooden structures in the third phase of the Ta Prohm’s third enclosure.

Fig.29: Distribution of the wooden structures in the fourth phase of the Ta Prohm’s fourth enclosure.

Fig.30: Distribution of the wooden structures in the fourth phase of the Ta Prohm’s third enclosure.

Fig.31: Distribution of the wooden structures in the sixth phase of the Ta Prohm’s fourth enclosure.

Fig.32: Distribution of the wooden structures in the sixth phase of the Ta Prohm’s third enclosure.

5.2. Preah Kahn

Fig.33: Synthesis of the chronological squences of third enclosure of Preah Khan’s stone skeleton

Fig.34: Distribution of the wooden structures in the second phase of the Preah Khan’s third enclosure.

Fig.35: Distribution of the wooden structures in the second phase of the Preah Khan’s third enclosure

Fig.36: Distribution of the wooden structures in the fifth phase of the Preah Khan’s third enclosure.

Fig.37: Distribution of the wooden structures in the fifth phase of the Preah Khan’s second enclosure.

As emphasised during the inventory of the archaeological traces of the wooden structures, compared to Ta Prohm, Preah Kahn shows only a few wooden lean-tos. Nevertheless there is a second enclosure with wooden framework dating back from its first construction phase (figure 34 and 35). It was dismantled and substituted by its stone equivalent. This new avatar showed several partitionings forming successions of alcoves where part of the statutory of the numerous divine population of the temple was stored (figure 36 and 37). Many occurences of several partitionings in the first enclosure-gallery of the monument of the first construction phase (figure 34 and 35) are found, which were then completed (figure 36 and 37). Those difference rooms once shaped must have stored several statues as is the case for the second enclosure. Therefore, although from the outside the first two enclosures of Preah Kahn are shaped like galleries, this appearance does not imply that the main functions of this latter were ambulatory.

5.3. Banteay Kdei

Fig.38: Synthesis of the relative chronology of the third enclosure of Banteay Kdei’s stone skeleton.

Fig.39: Distribution of the wooden structures in the first phase of the Banteay Kdei’s third enclosure.

Fig.40: Distribution of the wooden structures in the first phase of the Banteay Kdei’s second enclosure.

Fig.41: Distribution of the wooden structures in the fourth phase of the Banteay Kdei’s third enclosure.

Fig.42: Distribution of the wooden structures in the fourth phase of the Banteay Kdei’s second enclosure.

As early as during the first phase of construction of Banteay Kdei, a second enclosure with wooden framework esixted as well as a serie of partitionings in the north and south parts of its first enclosure. Their functioning must have been similar to those of the first two enclosures of Preah Kahn. The wooden gallery was substituted by its stone equivalent during the fourth phase of construction of the temple. Thus, three of the main temples of the Bayon style present the same succession of events regarding their second gallery. Considering other edifices with wooden framework from Ta Prohm, Preah Kahn and Banteay Kdei, it seems that a large part of the architectural history of these monuments is the consequence of the gradual substitution of originnally built wooden structures by their stone equivalent. These replacements could be due to two major constraints which the project managers had to face at these periods. On the one hand the importance of the architectural programme of Jayavarman VII and on the other hand limited deadlines regarding its implementation. Therefore, in order to gain time during the implementation of the initial architectural programme of these huge complexes, the architects would have used the stone for the most important parts. Secondary buildings would have had to do with wooden framework while waiting for the work to be achieved on all the main temples’ architectural programmes. Whether at Ta Prohm, Preah Kahn or Banteay Kdei, the original architectural programmes remained close to their present state. Only the material in some of the edifices was substituted in order to perpetuate its composition. Therefore, to only study the mineral skeleton of these monuments could lead to false interpretations. This is confirmed with the case of the inner galleries of the Bayon that we will now study.

5.4. Bayon

Fig.43: Synthesis of the relative chronology of the Bayon’s stone skeleton.

Contrary to Ta Prohm, Preah Kahn and Banteay Kdei, Bayon was the subject of a complete transformation of its initial architectural option during the third phase of construction corresponding to the date of its external gallery. But as mentioned previously, the wooden lean-tos and the partitionings in this temple are located in its inner gallery. The following figure 44 relates the inventory of the latter with the Buddha’s niche friezes and inscriptions location.

Fig.44: Inventory of the wooden archaeological traces in the Bayon’s inner gallery.

The Buddha’s niche friezes are systematically interrupted where the areas for the wooden partitioning armature are located, leaving sections of the wall without any décor. Therefore, these partitions would have been laid out before the works of the decorative programme and can be considered as integrated to the initial architectural programme of this part of the temple. The presence of this partitionings explained the location of two inscriptions K293-9 and K293-10, located on a pillar of the first span of the galleries facing the little face towers BY.38 and BY.39, and not on the abutment of this latter. Thus, spaces gained by the presence of partitionings facing the small faces towers of the inner gallery of the Bayon were probably acting as antechamber. Figure 45 shows a reconstruction of part of the partitionings as well as of the wooden lean-tos of the inner gallery of the Bayon. Nevertheless the distribution of these partitions is not synchronic. Based on the relative chronology of the mineral skeleton of the Bayon and on the data coming from the study of the archaeological traces of the wooden structures, figure 46 is an attempt to restitute the architectural history of the South-East angle of its inner gallery.

Fig.45: Inner gallery’s cross-section with the restitution of the wooden structure.

First phase. Since the first construction phase of the Bayon the small face-towers BY.38, 40, 41, 43, 44, 46, 47 and 49, and of the inner gallery were originally preceded by antechamber(s) occupying one span of the gallery (1).

Second phase During the second construction phase, the new galleries on each corner of the current inner gallery.

Third phase After the construction of the current outer gallery, several wooden partitions (2) and stone extensions were added on.

Fourth phase During the iconoclast period the wooden partitions of the antechamber of the small face-tower might have been disassembled. At the same time the North, South and West doors of the small face-tower were closed with a wooden structure (3).

Fig.46: Relative chronology of the South-East angle of the inner gallery of the Bayon based on the synthesis of the archaeological traces study observed in situ.

The chronology of this part of the Bayon completed by the wooden structures brings us a new interpretation of the evolution of the spatial observance of the temple. Indeed, in its initial architectural programme (phase 1) and in its immediate retake (phase 2), the outer gallery of the temple was its present inner gallery (1 of the figure 47). It was then the first buffer between the profane space of the city (Angkor Thom) and the main sacred space of the Bayon, its present-day third level. During the building of the existing outer gallery (3 of the figure 47) the former

outer gallery became inner (2 of the figure 47) therefore losing its initial function as first buffer of the monument. It is probably at this time that the project managers of the Bayon allocated a new function to this gallery by including it in the sacred area of the temple. Therefore several partitionings were built adding on to those existing originally in order to multiply the areas dedicated to worship.

Fig.47: Restitution of the second and the third phase of the Bayon complex.

6. Conclusions As previously mentionned, we have only shown here the first results of the study of the wooden structures in the four most representative temples of Jayavarman VII’s architectural programme. Although inventories of the archaeological traces of the wooden structures edifices and of the partitionings have been realised for other monuments of this style, they are still to be interpreted. Finally the new relative chronologies that are being put forward for Ta Prohm, Preah Kahn and Banteay Kdei remain mostly to be argumentated. This could be done while setting up their respective monograph. Nevertheless we can already assert, based on our first results, that these monuments presented in their first architectural programme several edifices with wooden structures as well as partitionings which were fully parts of the ritual observance. These wooden lean-tos could have resulted from the constraint imposed by the architectural programme of Jayavarman VII: on the one hand its density and on the other hand its implementation. The partitionings invite us to reinterpret the function of the enclosures-galleries in the big complexes of this period. Indeed these were not only ambulatory areas, but could also have had different functions depending on the configuration of the partitions. Therefore all research, aimed at studying the ritual observance of these monuments, based only on their existing state and epigraphy sources could only lead to misinterpretations. According to us only the relative chronologies of their mineral skeletons completed with structures made of ephemeral materials constitue the only solid base for such studies.

7. Bibliography Arahi, H. 2003. Étude de l’évolution architecturale et des techniques traditionnelles utilisées au temple de Banteay Kdei, à Angkor, Cambodge: Leuven, Heverlee: “R. Lemaire” CCHTB, K.U. Boisselier, J. 1966. Manuel d’Archéologie d’Extrême-Orient, première partie, ASIE DU SUD-EST tome 1, LE CAMBODGE, Paris: A. et J. Picard et Cie. Cunin, O. 2004. De Ta Prohm au Bayon. Volume 1: Analyse comparative de l’histoire architecturale des principaux monuments du style du Bayon. Volume 2: Contribution à l’histoire architecturale du temple du Bayon, Annexe 1: Documents graphiques. Annexe 2: Documents photographiques. Nancy: INPL (Online thesis: http://tel.ccsd.cnrs.fr/tel-00007699) Cunin, O & Uchida, E. 2002. Contribution of the magnetic susceptibility of the sandstones to the analysis of architectural history of Bayon style monuments. Tokyo: JSA. pp. 205-254. Dumarçay, J. 1973. Charpentes et tuiles Khmères. Paris: MAEFEO n°8, Adrien-Maisonneuve. Dumarçay, J. 2000. L’évolution de la couverture de tuiles à Angkor, du IXe au XVIe siècle, Udaya, APSARA, Phnom Penh, pp. 125-129. Stern, P. 1965. Les Monuments khmers du style du Bayon et Jayavarman VII. Paris: PUF. Uchida, E - Cunin, O - Shimoda, I - Suda, C - Nakagawa, T. 2003. The construction process of the Angkor monuments elucidated by yhe magnetic susceptibility of sandstone, University of Oxford: Archaeometry vol. 45 Part 2, pp. 221-232.