THE POTTERY FROM THE EXCAVATION SEASONS 1990 - 1997 AT TRA KIEU IN THE THU BON RIVER VALLEY, CENTRAL VIET NAM*
RUTH PRIOR - INSTITUTE Of ARCHAEOLOGY, LONDON
INTRODUCTION
This paper will focus on the Early Historical ceramic assemblage from the site of Tra Kieu (TK), central Viet Nam. Tra Kieu is located approximately 20 km inland from Hoi An in the fertile Thu Bon River Valley (fig. 1).
There were four seasons of excavation between 1990 and 19971. As a result of the consistant occurrence of ceramics throughout the stratigraphy of the site, the pottery has been grouped into 3 phases. This paper will summarise the archaeology and the pottery of each phase, then move onto a brief discussion of the petrographic analysis carried out by the author on the ceramic fabrics.
* There were subsequent excavation seasons in 1998, 1999 and 2000 after this paper was written.
ARCHAEOLOGICAL DEPOSITS
There was a close relationship between the strati¬graphy found in the trenches of TK 90 and TK 93 due to the proximity of the two areas (fig. 2).
Comparing TK 90's north and west sections to TK 93's northeast and northwest sections (fig. 3 and 4) can illustrate this.
The TK 90 west section and TK 93 northwest section depict nearly identical stratigraphy, with only a slight variation in layer depths. The top soil was the same pale grey, followed by layers of rubble. In TK 93 the top soil was labeled layer 1 and the layers of rubble were grouped under layer 2, In TK 90 this same material formed the top of what Nguyen Chieu et al. (1991) recorded as Cultural Level 1 (ie: from 0.5-1.20 m). It has been renamed here as Phase 3. In the TK 93 trench section, layer 3 (a dark grey clay) followed, and corresponded with the layer in TK90 from 1.20 -1.60 m in depth. In the latter, brick, tile and potte¬ry were more evident. In the TK 90 trench, imme¬diately below this layer of charcoal rich clay, was a deposit of yellow clay, which appeared only as a thin strip within layer 5 of TK 93. The top of TK 90 Cultural Level II (as defined by Nguyen Chieu et al. 1991) appears to correspond with the top of layer 5 in the TK 93 trench. At the base of both trenches was a row of large stones, which were abutted by a layer of sand and fine gravel. From the section drawings it was clear that Cultural Level II (referred to by this author as Phase 1) began at a depth of approximately 1.90 m in the TK 90 trench.
The TK 90 north section and the TK 93 northeast section again showed similar deposits. More notable in these sections were the brick floors/ foundations. In the TK 93 trench the floor was labelled as Feature 7. In the TK 90 trench, just as 'brick'. These features all fell within Phase 2. It was concluded that TK 90, like TK 93, showed phases of collapsed buildings, inter¬spersed with layers of charcoal. It seemed certain that on this side of Buu Chau Hill there were brick and timber built structures which were burnt, collapsed and then possibly leveled, all re¬latively in situ, prior to being built over yet again (Glover et at. 1996: 169).
The majority of the spits excavated in the BCE 96 trench, about 80 m southeast of TK 90 and TK 93, related to those belonging to Phase 2 (Cultural Level I) at TK 90. The radiocarbon date from BCE 96 also supports this. However, there are in dications that the start of Phase 1 was reached because a sherd from an ovoid jar, typical of the Phase 1 ceramic assemblage, was discovered in the lower section of BCE 96, where the water ta¬ble inhibited further excavation.
Trench GDD 96 was located in a rice field about 200 m southeast of Bun Chau Hill - there all the archaeological activity falls within Phase 2, and represents a single event.
Six depositional layers were revealed, with only scattered and disturbed sherds and bricks in the top five. At the base of the excavation was a feature which consisted of two more-or-less complete spouted vessels (kendi), four bowls, some sherds and some burnt bricks - possibly pot stands at a hearth (fig. 5).
These vessels had collapsed together in a heap and were surrounded by a burnt area. One interpretation for this feature is that the burnt area represents a domestic hearth, perhaps below a building raised on wooden piles - the post holes of which lay outside the excavation area - into which the vessels have fallen (Nguyen Kim Dung and Glover pers. comm. 1996). However, the author has since re-evaluated the evidence and sug¬gests an alternative interpretation. The broken vessels sat in a shallow depression that contained charcoal. The kendi lay with their broken surfaces facing upwards - as though a weight had crashed them when collapsing on top of them. Furthermore, most of the vessels within the depression were poorly fired with friable fabrics. It is suggested that this is evidence of a failed firing. A small pit had been dug in which to fire a limited number of vessels, but which had failed. It is possible that the potters retrieved some useable vessels, hence the relatively small numbers re¬maining. The evidence from trench GDD96 is not inconsistent with what might be found from a failed open firing, as excavation of experimental firings have determined that the remains from such firings could be mistaken for the remnants of domestic hearths (Gibson 1986).
All of the pottery from the GDD 96 trench related primarily to the vessel forms usually found in Phase 2; and the radiocarbon dates from the GDD 96 trench were also of this later period.
The stratigraphy in the TK 97 trench depicted a similar picture to that in the TK 90 and TK 93 trenches, with much building debris and charcoal in the upper layers that corresponded well with Phase 2. The ceramics were characteristic of this phase also. The lower spits (13-17) related to Phase 1, and contained earlier vessel forms. This lower stratigraphy was composed of alternative layers of yellow clay and coarse sandy soils. No boulders were found at the base of this trench, but as these are being interpreted as evidence for terracing, and as the TK 97 trench was not situated on the hill itself, this absence could be anticipated. More remarkable, was the discovery of a compacted 'column' of bricks in the TK 97 trench, which could have been the foundation for a later structure (More of these columns were discovered in the 1998 -2000 seasons and will be discussed in later publications).
DISCUSSION OF ARCHAEOLOGICAL DEPOSITS
In considering the evidence from all of the trenches at Tra Kieu it was possible to suggest the following interpretation. The levels in Phase 1 in the TK 90, TK 93 and TK 97 trenches, produced large quantities of ovoid jar sherds. This phase also included the line of boulders in the first two trenches. A radiocarbon date from the 1993 exca¬vation provided a calibrated date of BC 381 - 6 AD (table I). Thus, the earliest activity appeared to be agricultural in nature - probably as terracing. By averaging the radiocarbon dates from Phase 1 it can be dated to the 2nd century BC to 1st century AD.
In Phase 2 there appeared to be an increase in activity at Tra Kieu. There was a wide range of vessel forms present and within the stratigraphy there was evidence for some building, followed by destruction. The average of the radiocarbon dates from Phase 2 dates it to 2nd century AD to 4th century AD.
The excavations provided little evidence for later occupation. In the surface layers of each trench (Phase 3) glazed wares were discovered, dating mainly to the 14th and 15th centuries AD, and a few later pieces. From the TK90 surface layer a ceramic sherd was recovered which had characters engraved on it (Nguyen Chieu and Lam My Dung ND:18). These appeared to be a Coptic cross and the numerals 370. They were be¬lieved, by the excavators, to be Arabic and could date to the 8th century AD.
Overall, the archaeological evidence showed continued and substantial occupation at the site of Tra Kieu from the late centuries BC to about the 4th century AD. Phase 2 could relate to the events surrounding the expansion of the Lin-yi kingdom during the centuries following the decline of Han power, but further archaeological evidence would be required to corroborate this theory.
Surprisingly the trenches excavated in the 1990s have failed to produce much material that could be related to the period of the main temple complex at Tra Kieu excavated by Claeys in 1927-28, which is commonly dated to between the 9th and 12th centuries AD. The trenches excavated in the 1990s have consistently produced much earlier material. As Claeys did not keep records of the pottery he found, it was not possible to know for certain whether the stratigraphy present in his excavation trenches spanned the early and later periods. However, a photograph from Claeys' excavation depicts a small group of pottery whose forms appear to be similar to those found in Phase 2 of the 1990's excavations. This suggests that Claeys did discover material with a possible date range from 1st AD to 9th century AD or later.
THE CERAMIC SEQUENCE PHASE 1 ASSEMBLAGE
The earliest occupation at Tra Kieu is represented as Phase 1. Radiocarbon dates suggest a date range of 2nd century BC to early 1st century AD for the pottery occuring in Phase 1. This assem¬blage is characterised by the following vessel forms: ovoid based jars, cooking pots, flat lids, shallow bowls and dishes and pedestal cups (fig. 6), Nguyen Chieu et al. (1991) believed that that forms and decorative motifs of the Phase 1 assem blage could derive from the late Iron Age Sa Huynh vessel forms. However, Sa Huynh sites are predominantly burial sites, whereas Tra Kieu is a settlement site. Thus, the vessels found at each site would have served different functions making any continuity from Sa Huynh to Early Historic (Cham) difficult to find.
The ovoid based jars are generally 28 cm high with mouth diameters of 18 cm. This vessel form appears to be unique to the Tra Kieu area. It is quite unlike the burial jars found on Sa Huynh sites. Ovoid based jars are probably made using a paddle and anvil technique which also results in the cord mark decoration. These vessels do not have handles. They occur almost exclusively in the lowest layers of the excavation (table II).
Their function is uncertain. In China similar vessel forms were used as water carriers, as seen in illustrations and in clay models. In order to elucidate the problem a little, residue analysis was carried out on 3 sherds by M. Cave (Institute of Archaeology, London). Infra-red analysis resulted in spectra peaks showing that the jars had contained a highly saturated plant oil. At present the exact type of oil is unknown - it is possibly a palm oil. Further residue analysis on a range of vessel forms from Tra Kieu is now being under¬taken to examine the problem of vessel function in greater detail.
It is worth noting that some ovoid based jars appear to have been reused in antiquity. The 1990 excavation revealed a cluster of such jars in situ at the base of brick foundations. Perhaps one of the functions of the jars was ritual - being deliberately deposited within the foundations during the course of their construction - a feature observed in the building of Javanese temples (J. Miksic pers. comm. to I. Glover July 1998).
Cooking pots are the second most common vessel form in the Tra Kieu Phase 1 assemblage. Cooking pots generally have spherical bodies, round bases and unrestricted mouths with everted rims (fig. 7).
The author grouped the rims into three groups based on the rim inner surface. This could be flat, concave, or convex, thus indicating variations in forming techniques (fig. 8). The con¬cave rim group were the most common type. Cau¬tion has to be used in placing too much emphasis on rim type. As the rims were hand-made considerable variation could be found on one rim - especially in width/depth.
It is possible that the cooking pots were made using coiling and paddle and anvil as they have very thin vessel walls (usually 0.5 cm). The cooking pots found at Tra Kieu are usually decorated with cord marking Nguyen Chieu (pers comm. 1997) believes that similar round based vessel forms found on the Sa Huynh sites around Hoi An (for example Hau Xa and Dong Na) illustrate that there was cultural continuity from the Sa Huynh period into the Early Historic period. However, this vessel form is commonly found throughout Viet Nam and is known from Neolithic cultures in Viet Nam. "The possibility of continuing local ceramic traditions in pottery technology was one of the issues examined during the thin section analysis of the ceramics. The flat lids frequently found within Phase 1 possibly relate in function to the cooking pots.
Pedestal cups and shallow bowls were forms common to the later part of Phase 1. Pedestal cups were identified by the following variables: cylindrical columns with a circular base. The diameter of the base exceeded that of the column. The columns ranged from 2-5 cm in height. The pedestals appeared to belong to cups or small bowls. They were commonly made from a coarse fabric. Shallow bowls had an ellipsoid shaped bo¬dy with a flat base. This vessel form was usually identified by fragments of base or rim. The height of the vessel ranged from 3-6 cm and the diameter of the mouth from 10-12 cm. An absence of rilling and other features characteristic of wheel throwing indicated that shallow bowls were predominantly hand-made. Shallow bowls are another vessel form known in the late Sa Huynh culture which could have functioned as oil lamps (Nguyen Chieu pers comm. 1997), similar to the ones used until recently in Viet Nam.
Significantly some do have blackening around the rims. Fragments of stoves occur occasionally in Phase 1. This is a common form to Viet Nam and Southeast Asia. Generally the stoves from Tra Kieu are undecorated, unlike those found at Oc Eo in south Viet Nam. These latter are often high ty decorated with geometric and zoomorphic motifs. Stoves do not occur exclusively in Phase 1, but overlap into Phase 2. They are an indication that the early assemblage at Tra Kieu share similarities with other ceramic groups beyond central Viet Nam.
Even in this early Phase there is evidence for external contact - though the form it took cannot be ascertained. A sherd of Indian Rouletted Ware was found in the lowest layer. Thin section analysis by the author confirmed that the fabric of the Tra Kieu sherd is extremely similar to a sherd found at Arikamedu - confirming that both were likely to have been made at the same production centre. Gogte (1997) carried out XRD analysis which revealed similar results.
PHASE 2 ASSEMBLAGE
Whilst cooking pots and flat lids occur in smaller numbers in Phase 2, indicating some element of continuity, the Phase 2 assemblage is dominated by two vessel forms: flat based jars and kendi, as is illlustrated by the group of vessels found in trench GDD 96. The average of the radiocarbon dates from Phase 2 show they date to 2nd century AD to 4th century AD.
The main form is a flat based jar with ellipsoid shaped body, constricted mouth with vertical rim and stamped decoration (fig. 9).
The style of the jar is reminiscent of the jars first produced in China during the Warring States period and continuing into the Han period. Such jars are first known in Viet Nam when production began at the kiln site of Tarn Tho in Thanh Hoa province. During the 1920's and 30's Olav Janse excavated the kilns and associated brick built tombs at Bim Son and Lach Truong which contained vessels made at Tarn Tho. Thus, at Tra Kieu there is a vessel form with clear Chinese inspiration. A similar vessel has been found at Hau Xa, a Sa Huynh site near Hoi An, which appears to have some settlement evidence as well as the usual burial jars. The Chinese style jar was found in one of the upper levels at Hau Xa: comparable with the Early Historic Phase 2 at Tra Kieu (Lam Thi My Dung and Nguyen Chi Trung ND).
Phase 2 is not just dominated by a Chinese vessel form. The second most characteristic form is the spouted vessel, frequently referred to as a kendi (fig. 10).
The kendi forai is distinctive. Those found at Tra Kieu are recognised by the following variables: globular bodies, a high narrow neck with restricted mouth and rims that are slightly everted, recurved or upright. They also exhibit spouts and either flat bases or, more frequently, a ring foot base. Kendi do not have handles, for pouring, they are gripped by the neck. The body of the kendi is often decorated around the shoulder. Decoration is restricted to one or two parallel lines on the surface between shoulder and body, or double or single wave patterns. Yamagata (1997:180 and 1998:76) has noted that the kendi from Tra Kieu share some similarities with those from Oc Eo of 3rd century AD. The kendi form is believed to have originated in India, and is found across Southeast Asia. As Phase 2 begins as early as the 2nd century AD, the Tra Kieu kendi are probably the earliest yet known from dated contexts in Southeast Asia.
Despite the Phase 2 assemblage being very different to that of Phase 1, showing both Chi nese and Indian elements, there did not appear to be any significant change in the way the pots were produced. Those in Phase 2 were still predominantly hand made using a combination of coil building and paddle, and anvil. It is possible that some of the finishing was done by rotating the vessel on a turn table, tamed by hand or foot. They may also have been formed upturned on top of another vessel, as seen today at Phari Rang, and which would account for the bases of the Chinese style jars being flat and smooth.
PHASE 3 ASSEMBLAGE
After a hiatus of unspecified length there are the Phase 3 ceramic forms. These are all glazed wares ranging in date from the 10th to the 18th centuries AD. They include South Chinese Martaban type jars, Chinese stonewares and porcelains. They have not been studied in detail by the author.
THE CERAMIC FABRICS
The author carried out detailed petrographic ana¬lysis of the ceramic fabrics from Tra Kieu in order to characterise the fabrics mineralogically and texturally. Until then, the fabrics had been de¬scribed in reports as being fine, moderately coarse or coarse, possibly sand tempered. The textures were not defined so it would be difficult to be consistent in the recording of the ceramics. It was also hoped that the thin sections would reveal more on the technology used by the potters in the production of the vessels.
The author classified the fabrics primarily on variations within the fine grained matrix of the fa¬bric (what might essentially be considered the clay). This method was employed because of the similarities shown throughout the ceramics in the larger non-plastic inclusions. It was felt that qualitative analysis of the non-plastics alone would not distinguish the ceramics into fabric groups.
Two main types of clay appeared to have been used in the production of the Tra Kieu ceramics:
1. Mature clays. These are devoid in unstable mi nerals such as feldspars and biotite. Quartz is the only mineral readily identifiable. Such clays can occur in marine environments. Such deposits are known in the Thu Bon River Valley.
2. Immature fluvial clays. These contain a range of micas, ampfaiboles, feldspars and quartz, all of which are characteristic of the surrounding meta-morphic terrain. Numerous Quaternary alluvial deposits can be located alongside the Thu Bon River.
Within each of the two main clay types, subdivisions could be made based on the presence or absence of certain minerals. This provided the basis of the fabric groups of the Tra Kieu ceramics.
The thin section analysis further demonstrated that any given clay may have been treated in a num¬ber of ways by the potter - for example the clay may have been levigated, tempered, or used in its natural state. For example, the clay for fabric group 1 (a mature clay) occurs as a fine ware, and then the same clay occurs as a moderately coarse ware and as a coarse ware. Similarly with the fluvial clays.
In order to define the textures, grain size analysis was carried out (by which the mineral inclusions were counted and measured). The resulting data was processed using two clustering programs - K-means cluster analysis and Principle Components Analysis (PCA). The PCA plot clearly shows 3 groups (fig. 11). The same samples were found to fall into the same groups with the K-means analysis. Group 1 represents the fine wares, group 2 the moderately coarse wares and group 3 the coarse wares. There is a bimodal distribution in group 2.
With the fabrics and the texture characterised it was then possible to look at how they related to the vessel forms.
Whilst there was a strong correlation between vessel form and grain size distribution (the texture), there was seemingly little correspondence between vessel form and fabric group. A range of forms were found in the same fabric group. So a flat based stamped jar could be found in the same fabric group as a cooking pot - but perhaps with a different texture (each fabric group was also subdivided by texture, but because the fine grained matrix was the characterising feature fine and coarse can co-exist in the same group).
FIRING CONDITIONS
Early Historic pottery kilns associated with the pro¬duction of any of the vessel forms described above, have not yet been located in the Thu Bon River Valley. It is likely that the wares in Phase 1 were produced mainly using a bonfire firing technique, as most of the vessel sherds show evidence of 'fire clouding'. The vessels of Phase 2 show more overall even and controlled firing, which may have involved kilns of some type. It is possible to suggest that the ceramics were generally fired below 1000°C.
A series of experiments were conducted by the author on local clay samples collected in the Thu Bon River valley, in order to evaluate firing conditions for the Tra Kieu ceramics. The results of the analysis are summarised in table III.
Table III summarises the main alterations witnessed during the firing experiments when tempe¬ratures were increased from 700 °C to 1000°C. The alteration in amphibole was noted in all of the clay samples where this mineral was present. The colour of the amphiboles in PPL, when examined under a polarising microscope, became more in¬tense (usually changing from green to brown) with an increase in firing temperature.
Previous authors (Echallier and Mery 1992; Whitbread 1985 and Porat 1989) have noticed similar alterations to amphibole. Whitbread (1985: 32) states that when hornblende is fired at a tempera¬ture of 925 °C for 25 minutes in an oxidising atmo¬sphere, the alteration in colour intensity affects pleochroism and masks birefringence. Porat (1989:29) similarly noticed that in an oxidising kiln at a temperature of 8OO-850°C there was a change in green hornblende to red oxyhomblende. Deer etal. (1992:252) confirm that the alteration from hornblende to oxyhomblende occurs at 800°C.
Biotite underwent alteration when in samples fired at higher temperatures. The alteration manifested as a reduced level of birefringence.
An unexpected indicator of the temperature at which a ceramic was fired was the presence or ab¬sence of rice trichomes in the sample's matrix. Initial examination of the ceramic thin sections re¬vealed that many of the fabric matrices contained frequencies of rice trichomes. When the clay samples were examined it was noted that clay sample 14 similarly contained frequent rice trichomes. Rice trichomes are biogenic silica and previous study (Fujiwara 1993) has established that at temperatures of 1000°C and above this material will melt - thus the rice trichomes will no longer be visible. Comparison between the thin sections of clay samples at 700°C and 1000°C showed that whilst the trichomes were visible in the sample fired at 700 °C they were not visible in the sample fired at 1000°C. This suggests that the trichomes had been destroyed by the firing at a higher temperature. This information acted as a guide for estimating the firing temperature of the Tra Kieu ceramics to be below 1000°C.
DISCUSSION
The results of the thin section analysis were interpreted by the author as indicative that there were a series of small production centres throughout the Thu Bon River Valley which made the full range of vessel forms then in vogue, using locally available materials (clay and temper). Because ovoid based jars and the Chinese style jars occurred in the same fabric groups (and sometimes with the same texture), it is possible that some of the centres continued in production over a long period of time. The technological ceramic traditions did not seem to rapidly evolve, even when distinctly new vessel fonns were introduced. Over the Early Historic period represented at Tra Kieu ceramic production did not appear to become centralised, but seemed to remain on a domestic smaller scale, as indicated by the range of fabric groups present in both Phases 1 and 2. In total 12 clay variations were identified, and whilst some could be the result of variation within a single clay deposit, it is just as likely that the variation is due to the collection of clay from different clay deposits. It is suggested here that had centralisation of production occurred there would perhaps have been fewer fabric groups and greater standardisation in vessel forms.
A similar situation of small scale production scattered along the Thu Bon River Valley can be seen in the present day brick production. The kilns, located near the banks of the river, utilise clays dug directly from the paddy fields - and if the same practise was used in the Early Historic period, this may account for the presence of rice trichomes found in the ceramic fabrics - and when sand temper is required the banks of the river are used as a source.
REFERENCES
Aurousseau, L., 1923, "Le premier conquete chinoise des pays annamites", BEFEO 23:137-264.
Claeys, J.-Y., 1927, Fouilles a Tra Kieu. BEFEO 27: 468-82.
Claeys, J.M., 1928, "Fouilles a Tra Kieu", BEFEO 28: 578-96.
Deer, W.A., Howie, R.A. and Zussman, I, 1992, An introduction to the rock forming minerals. Harlow: Longman UK Group Ltd.
Echallier, J.-C. and Mery, S., 1992, "Devolution mine-ralogique et phsico-chimique des pates clacaires au cours de la cuisson: experimentation en laboratoire et application archeologique", in S. Mery (ed.) Sciences de la terre et ceraniiqiies archwlogiques: experimen¬tation, applications 16. Saint-Louis a Cergy: Centre Poiytechnique.
Fujiwara, H., 1993, "Research into the history of rice cultivation using plant opal analysis", in Pearsall, D.M. and Pipenio, D.R. (eds), Current research in phytolith analysis: applications in archaeology and paleoecology. MASCA Research Papers in Science and Archaeology Vol. 10, 147-158. Philadelphia: MASCA, University Museum of Archaeology and An¬thropology.
Gibson, A., 1986, The excavations of an experimental firing area at Stamford Hall, Leicester: 1985. Bulletin of the Experimental Firing Group 4: 4-14. Glover, I.C. and Yamagata, M., 1995, "The origins of Cham civilization: indigenous, Chinese and Indian in¬fluences in central Vietnam as revealed by excavations at Tra Kieu, Vietnam 1990 and 1993", in Yeung, C.T. and Li, B. (eds), Archaeology of Southeast Asia, 145-69. Hong Kong: University Museum. Glover, I.C., Yamagata, M. and Southward!, W., 1996, "The Cham, Sa-Huynh and Han in early Vietnam: excavations at Buu Chau Hill, Tra Kieu, 1993", in Bellwood, P. {edl), Indo-Pacific prehistory: the Chiang Mai papers Bulletin of the Indo-Pacific Prehistory Association 14, 166-176. Canberra: Indo-Pacific Pre¬history Association.
Gogte, V. D., 1997, 'The Chandraketugarh-Tamluk re¬gion of Bengal: source for the Early Historic Rouletted ware from India and Southeast Asia", Man and En¬vironment XXII (1): 69-85.
Lam Thi My Dung and Nguyen Chi Trung (ND), "The site of Hau Xa I and the continuity of cultures in the early Ciiristian era in Hoi An (in Vietnamese). Paper to appear in Khao Co Hoc.
Nguyen Chieu, Lam My Dung (ND), Unpublished excavation report about the 1990 excavation at Tra Kieu (in Vietnamese).
Nguyen Chieu, Lam My Dung, Vu Thi Ninh, 1991, "Ceramics from the excavations of the ancient Cham site at Tra Kieu, 1990" (in Vietnamese), Khao Co Hoc 4: 19-30.
Porat, N., 1989, Composition of pottery - application to the study of the interrelations between Canaan and Egypt during the 3rd millennium BC. PhD thesis submitted to Hebrew University. Stein, R.A., 1947, Le Linyi. Bulletin du centre d' etudes sinologiques de Pekin Vol. I!.Whitbread, I.K., 1985 "Firing and reftting tests in the context of ceramic thin section studies", Bulletin of the Experimental Firing Group 4: 30-48. Yamagata, M., 1997, "Formation of Lin-yi: derived from the archaeological materials of Chinese origin found at Tra Kieu, central Viet Nam", Journal of South¬east Asian Archaeology 17: 167-84. Yamagata, M., 1998, "Formation of Lin Yi: Internal and external factors", Journal of Southeast Asian Ar¬chaeology 18: 51-89
Nguồn: R. PRIOR, THE POTTERY FROM THE EXCAVATION SEASONS 1990 -1997 AT TRA KIEU IN THE THU BON RIVER VALLEY, CENTRAL VIET NAM
SOUTHEAST ASIAN ARCHAEOLOGY 1998: 149-169
Proceedings of tfce 7th International Conference of the European Association of Southeast Asian Archaeologists Berlin, 31 August - 4 September 1998, Wibke Lob© and Stefanie Reimaim Editors
© Centre for South-East Asian Studies, University of Hull Hull, Great Britain, 2000
