Search Results

One of the most striking features of the excavations in rural settlement sites, conducted as part of the SARCP program, has been the very large amount of pottery retrieved. Not only has this created problems of storage, but also of data process­ ing. As a means of handling these very large amounts of pot­ tery in a convenient way and yet getting the most out of them information-wise, a method of pottery analysis founded on the distribution of wares was applied. The method was tested at a special pottery workshop or­ ganized within the SARCP framework, held at the National Museum in Colombo, March 1-10, 1992, and continued less intensively at another location until April 30th. We are most obliged to the management of the museum for putting a room at our disposal during the first intensive days. At this workshop, the total pottery assemblage from the two large scale settlement excavations at Tammannagala 1990 and Ibbankatuva-Polvatta 1991 was analysed, in addition to the less bulky ceramic assemblage from the Mapagala excava­ tion of 1990 and the excavations at Dehigaha-ala-kanda in 1990 and 1991. Examples of diagrams are published with the accounts of Ibbankatuva and Mapagala. Here it will be suffi­ cient to give a description of the methodology applied. A pottery analysis of the kind applied here is carried out under one absolute prerequisite: the excavation must be car­ ried out using a strict stratigraphical method, where the pottery is collected and registered according to stratigraphical units. If this is not done, a matrix chart according to the Hanis method cannot be put together, and without a reliable stratigraphic sequence analysis is impossible. Hence, the pottery from ear­ lier excavations at Ibbankatuva-Polvatta cannot be processed in this manner, since those excavators used the less accurate planum method for excavation. This means that the separate excavations at this same site can never be compared. The objectives of the workshop were: 1) To contribute towards working out an operative descrip­ tive code for wares. We were working from the awareness that it is important to avoid dividing wares into too many varieties; the sub-division has to be operational for dating purposes. Attention was paid to earlier attempts at classifying wares in Sri Lanka. 2) To use the stratified pottery finds for establishing rough relative chronologies based on wares (which later can be elaborated by working with rim-forms). Those relative chronologies are subsequently being correlated with 14C datings. 3) To compare analyses of total pottery assemblages with analyses of separate sub-square assemblages (this was done on the Ibbankatuva-Polvatta pottery only) and to compare analyses based on weight with analyses based on the number of sherds. This was done mainly for the inherent pedagogical value. 4) To experiment with different ways of displaying results, using tables, graphs etc. The advantage of the four chosen sites, besides their all having been excavated and recorded stratigraphically, is that work could be done with pottery assemblages that had a rela­ tively high ware variability, i.e. there was EHP1/EHP2 or even protohistoric pottery, preceeding the pottery of the Kasyapan period and later Anuradhapura period. In those later periods the ware variability appears to have been slight over a long period of time. The pottery bags were sorted context-wise and in the order of stratigraphical sequence. Then each single context was processed from the bottom up. Sherds were sorted according to wares. Special forms were filled in with weight in grams and the number of sherds for each ware type. The completed forms constitute the data base that has been used in various ways: 1) The frequency of undifferentiated wares in each context was calculated: the total number/weight of potsherds from an excavation = 100%. The number/weight of each context = X% of the total. By doing this, we acquired knowledge of the nature of deposits in different layers, which might indicate differences in the intensity of occupation over time, as well as an indication of which layers could be suspected to consist of redeposited fill. In the latter case, a correlation to the ware analysis must be made. By this method, something about the stratigraphical sequence that would otherwise be invisible, can be learnt. 2) The proportion of different wares in the different con­ texts was calculated: the total number/weight from one context = 100%. The number/weight of each variety of wares = X % of the total. By doing this, an idea of how the uses of different wares have changed over time can be gleaned. We may find the whole set of wares in most of the contexts, but the impor­ tant thing is the change in emphasis for the different wares. The data was then processed in a computer, using a pro­ gram for graphic representation. Two different types of graphs were produced: the calculation of the frequency of undifferen­ tiated pottery per context, was displayed in a simple bar-chart. Only contexts containing pottery were included. On the horizontal axis we show weight, or number of sherds, and on the vertical, the contexts starting with the stratigraphically ear­ liest at the bottom, according to the matrix chart. The proportion of different wares in each context was displayed in another type of graph: also, in this only the con­ texts containing pottery are included. The display gives a graphic representation of the stratigraphical record along the vertical axis. Thus the earliest contexts are represented by a lying bar at the bottom of the graph, etc. The horizontal axis displayes the percentages of different wares. One bar here = 100%. The various wares are hatched or screened differently. On completion of all this, yet another advantage of the method becomes evident. It is now possible to check on the details in the matrix and its division into phases. Finally, it is possible to work out a second set of graphs, displaying phases instead of contexts. The reason for not being satisfied with a display of the contexts is, of course, the results we can obtain from the first calculation: the frequency of undifferentiated potsherds in each context. Redeposited fill can perhaps be sorted out if required, alluvial layers with very few potsherds may be merged with adjacent cultural deposits, to avoid too fragmented a picture, or if they represent a hiatus in settlement, sorted out or enhanced according to what is required, etc. The method implies a lot of tedious work, but it is only hard data of this type that will move research forward, and computer technology is now making it possible to concentrate weeks of sorting, counting and weighing into one or two graphs which, within a second, will give the reader of a report a visual picture of the often very complicated stratigraphy of an archaeological site. That is what makes all the work worthwhile.

Gamini Adikari When compared with prehistoric research undertaken so far, there is some special significance in this study, in that it is the first systematic and extensive research of its kind in an area in Sri Lanka where inland ‘Dry Zone’ environmental and geographical conditions prevail. Almost all prehistoric re­ search earned out in Sri Lanka previously has been confined to the low country Wet Zone, the central highlands and the southern coastal belt (Pole 1907; Seligmann 1908; Parker 1909; Hartley 1913; Wayland 1914; Sarasin 1926; Hocart 1928; Deraniyagala, P.E.P. 1937; Allchin 1958; Deraniyagala, S.U. 1980,1988 and Deraniyagala, S.U. and Kennedy 1972). Therefore investigations into a new and important aspect of Sri Lanka’s prehistory have been initiated by this study of the prehistoric culture of the Dry Zone heartland of Sri Lanka. This paper presents a brief review of the results of research conducted into the prehistory of the Sigiriya area. In this re­ search program, attention has been focused on factors such as the location of prehistoric sites within the area, the micro-en­ vironmental conditions and the nature of cultural and natural finds discovered at those sites etc. It is a basic assumption that the emergence and character of prehistoric settlements in any given area reflect a definite con­ nection with the physical environment of that area. Hence the necessity to pay attention to the environment and geography of the area under study. Previous research conducted on these subjects relevant to this area (Vitanage 1959; Domros 1974; Epitawatte 1989) were made use of in interpreting the finds for the purpose of this study. A recent floristic survey of the area (Epitawatte and Premathilake unpubl.) and of ancient faunal remains of the area (Karunaratne unpubl.) were of consider­ able help in analysing some of the data for this study. Geological parameters Prehistoric sites situated in the area belong to two major types i.e. cave dwellings and open air sites. Our excavations indi­ cate that natural rock shelters, situated on elevated rock-knob erosional remnants, were regularly used as living places by prehistoric man in this, as often in other regions of Sri Lanka. In the geological formation of the country, the Highland series has been responsible for the creation of a large number of caves in the area (Cooray 1990). In contrast, all examples of the open air sites indicate they have been used as hunting grounds and/or as places for the production of implements. Figure 3:1 Millagala (SO. 35) in 1990. View from north-west towards Sigiriya rock. Photo: Mats Mogren. Cave sizes Rock shelters within the study area can be preliminarily sub­ divided into three main groups, according to their size- and disposition. Since there is clear evidence to conclude that natural rock shelters in the area were frequently used as prehistoric dwellings, it was considered that in a number of such caves their floor areas provided some indicators of the numerical strength of the groups of people who occupied these caves at different times. Using this hypothesis as a basic factor, the entire sample of caves in the area was classified as cluster caves, single line caves or solitary caves. Caves concentrated at certain elevated locations, like those at Sigiriya and Potana, were described as ‘cluster caves’. The best example of ‘single-line caves’ is found on the upper terrace at Dambulla, but Pidurangala also belongs to this group. Such caves are situated in a single line, or row. At both Dambulla and Pidurangala, it is feasible to speak of cluster caves as well. It can be surmised that these two types, i.e. cluster caves and single-line caves, may have accommodated, or had the potential of accommodating, a lar­ ger number of occupants during the time they were in use, than the solitary caves, situated in places like Aligala, which ex­ emplify the third category: the solitary caves. Open air sites Three open air sites can be identified in the area: Millagala, Tammannagala and Mapagala. Open air sites of this nature are located in close proximity to natural depressions at ave­ rage groundlevel, where small pools or lakes are formed or near-natural water reserves and rock pools on the summits of rocks which do not exceed 20m above ground level. It can be thought that these water resources may have been one of the fundamental environmental factors that facilitated the use of such open spaces as prehistoric hunting grounds. Excavations for prehistoric studies were not carried out at any of the sites identified as open air sites in the area. Excavations Natural rock shelters at three selected sites - Aligala, Dambul­la and Pidurangala - were subjected to excavation. The Dam­ bulla and Pidurangala excavations were carried out as parts of the KAVA collaboration project, whereas Aligala was ex­ cavated under the SARCP program. At the time the Potana cave was identified as a prehistoric site, it had been dug up for the construction of a building. This excavation was not sub­ ject to archaeological control and documentation and only the nature of the natural and cultural finds unearthed were reported. It was by the chance of seeing the debris thrown up by labourers, digging the cave floor for laying the foundation of a building, that the Potana cave was identified as a prehis­ toric site. Within the SARCP framework it was possible to investigate the remaining deposits. Figure 3:2 Map showing prehistoric study locations of research area. Aligala This natural rock shelter is situated on the slope of the eastern side of the Sigiriya rock. The inner extent of the cave is 41m2. The maximum height is 6m. The mouth of the cave faces east. It is situated at an estimated elevation of about 35-40m above the average ground level of the area. No dripledge is carved above the mouth of the cave. An area of 64m2 within the cave was subjected to excava­ tion. The total number of activity phases (deposited layers) is four, with a total depth of 1.70m. The contexts were numbered from the surface downwards (see matrix chart, fig. 4:6). The fourth layer, situated at a depth of 1.50m from the surface, is an undisturbed prehistoric deposit. A considerable amount of stone implements was found in this layer. I The third layer is 10cm thick. The compact nature of the layer shows that it is an undisturbed deposit. This layer can be described as a short term abandonment, or a residual layer created as a result of sudden and heavy rain. No cultural debris whatsoever was found in this layer. The second layer contained cultural debris. The peculiar fact about this layer is that prehistoric artefacts are found at its lower levels, while historic remains are found at the surface levels. There is reason to believe that the second layer repre­ sents a transitional stage. The first layer contains only historic artefacts. The few stone implements appearing in layer 1 and 2 point to their being disturbed layers. This disturbance could have occurred layer level were identifiable in the profile. Such pits are indi­ cated as context nos. 2,3, 8 in the matrix chart. The following are the cultural debris found from the layers: figure 3:3 Pidurangala rock shelter. Site plan Pidurangala Excavations were carried out in the natural rock shelters on the north and west surfaces of the rock outcrop called Pid­ urangala, situated about 500m north of the Sigiriya rock. The total number of caves excavated was three. The prehistoric artefacts found in all three caves are similar in quality. Res­ ults of research in the upper cave of Pidurangala, where historic artefacts are found in a higher density than the two lower caves, are described below. The cave at the highest level is larger than the others. The length of this cave is 40m, while the roof is 7m high, measured from the bedrock. The mouth of the cave faces east. There is a dripledge carved above the opening of the cave. There is an inscription belong­ ing to the 4th century AD carved on the rock opposite the cave. The central portion of the cave had been converted intoan image house in the historic period. Three layers were found during the excavation of this cave. All three layers contained cultural debris. It is the third layer which can be described more specifical­ ly as the prehistoric layer. It is situated at a depth of about 50cm from the surface. Beside stone implements, no other cultural debris was found in this layer. The second layer is entirely a layer of ash. The first layer contains a large quantity of potsherds and other objects belo­ nging to the historic period. But several stone implements too, were found in that layer itself. This is due to disturbances in the layers. Because the depth from surface level to the prehis­ toric layer is very slight, there is a strong possibility of layers mixing with one another. There is evidence to believe that the earth of the third layer may have been disturbed, owing to loose soil caused by the lighting of fires on that layer. The following are the cultural debris found in each layer: Figure 3:4 Pidurangala rock shelter. Test pit profile from north with a section of the rock Dambulla Figure 3:5 Dumbulla site plan. Inset map of main rock shelter A test pit 2x2m was started on the terrace opposite the second cave of the Buddhist monastery called Rangiri Dambulu Vih- araya (see fig. 3:5). There was no possibility of any large scale excavation work in the cave itself, since it is currently used as an image house. A total of three layers was found in the excavation. The third layer contained only prehistoric stone implements. Apart from that, no other cultural remains were found there. Beneath this layer is bedrock. The second layer had been subjected to subsequent disturbances. Among the cultural debris found were potsherds belonging to the 20th century and flat tiles of an earlier period. No cultural debris was found in the first layer, which is a filled layer, probably a part of the architectural plan of the complex. Exploration Sites showing geographical features conducive to occupation and resources exploitation were systematically surveyed, to obtain evidence of prehistoric, temporary, or seasonal hunt­ ing grounds, open air sites and production sites in the area. Rocks and villus and natural water holes, situated at 3-30m level above average ground level, were paid attention to in the research. It is difficult to understand the prehistoric fac­ tors on the surface of open air sites. These could be caused by various geophysical activities. Stone implements which do not decay quickly were taken as evidence to identify these sites. The three open air sites of Millagala, Mapagala and Tam- mannagala were surveyed in this manner. Compared with the natural cave dwellings, these are very few in number. Marked differences were not observed among the open air sites, but according to surface collections and the nature of their loca­ tion, minor differences were visible. The important factor, revealed on a careful observation of these sites and their sur­roundings, is that to a large extent they are protected locations, connected to natural water reserves and waterways. The na­ ture of each place is given below: Mapagala Figure 3:6 Millagala site plan. Mapagala, an isolated inselberg towering about 30m above the plain, stands 500m to the south of Sigiriya rock. The old Yan Oya, today known as the Sigiri Oya, flows south of Mapagala. The summit of Mapagala, an almost flat area of 1000m . which slopes gently to the south, is dotted with natural depressions, where water is obtainable even during severe droughts. This location has all the classical features that sets it apart as an ideal choice for a hunting ground and open air site by prehistoric man (see fig. 9:3). Surface collection At Mapagala too, as with other tropical prehistoric occupa­ tion sites, fewer bone remains were found at open air sites than in rock shelters. Stone implements and fragments of stone debris from tool production were the only major finds. Tammannagala Situated about l/2km west of Sigiriya, Tammannagala rock, which is 60m long and about 30m wide, rises about 20m above the plain. Explorations of its summit and surroundings yielded a large amount of quartz chips and flakes, among which a few stone implements were identifiable. Quartz is not common in this area, so the presence of such quartz flakes undeniably proves prehistoric human activity in this area. We can conjecture that the Sigiri Oya (old Yan Oya) which flows about 300m east of the rock, must have been the source of raw material for the production of stone imple­ ments. Talkote Ihalavava is situated in close proximity, about 100m north of Tammannagala. Considering the presence of quartz flakes and the proximity to Talkote Ihalavava, interest­ ing problems can be addressed regarding the possibility of a continuous development from camp-sites to sedentary, agr­rian habitation sites, and also the origin of the gamgoda. Millagala This rock slab is about 8-10m high. The Pidurangala cave complex, which is situated about 200m east of the rock, is the most prominent feature of the surrounding landscape. The Sigiriya Oya flows about 100m west of Millagala. The Pidu- rangalavava, which is in close proximity to Millagala, is a construction of the modem period (see fig. 3:6). During ex­ plorations, a large number of quartz flakes were discovered strewn around the base of the rock. The rock pools at the summit attract jungle animals. The evidence indicates that this site may have been used for the production of imple­ ments and hunting by those prehistoric communities that set­ tled in the Pidurangala caves. Quartz flakes, found in large quantities upon the rock slab, indicate they are debris from the production of implements. Among them are a few fragments of broken implements. Prehistoric implements Prehistoric implements were found at all sites explored and excavated. This sample of implements is taken as a whole and its special features are discussed here. Three types of raw material have been used in the produc­ tion of implements: two basic minerals, quartz and chert; and animal bones. The bone implements found are preserved well enough to give an idea of which parts of animal skeletons were used for tool making. Decay had changed the basic shape of most of the bone tools. Implements from each field varied according to the raw material used. The density of the implements of various mat­ erials found at each site, is given in the table below. Although these stone and bone implements were used for different purposes, they can be discussed under five major categories, depending on utility and shape. They are: scrapers, pointed tools for boring and piercing, blades or cutters used for cutting, grinding stones and pitted stones. Evidence has been found that the implements were used as sub-types of these major categories. Therefore it seems they were used for special functions. Figure 3:7Artefact distribution ofAligala excavation in 1990. On reviewing the physical and cultural data discussed ab­ ove, several definite conclusions and hypotheses can be ar­ rived at, the most important being the concept of ancient settlements in the research aretf'. Considering the evidence from this’research, together with the prehistoric factors dis­ covered in other parts of the island, it is acceptable that this area under study was settled by a people belonging to the mesolithic culture, at least 3000 years ago. On an analysis of certain factors found at some sites, there is indirect evidence to conclude that the environmental ph­ enomena of those ancient settlements was somewhat different from that found today. The non-availability of sufficient data to reconstruct either the physical aspect or the social behavioural patterns of the groups of people in those settlements, was the basic obstacle in bringing the story of the prehistory of the area to a complete and logical conclusion. Up to now, there has been no evidence whatsoever from which to describe the physical appearance of Lankan mesolithic man. At Potana, however, two skeletons have been found, which may prove to be very important in this sense. They appear to have been buried in layers proceed­ ing a 14C-dated layer from ca 3,800 BC (see 14C-datings" this volume). Faunal remains of species of snails from certain sites can be identified as water-related creatures (Karunaratne, P.B. un- publ.). The implication of this is that the hunter gatherers may have collected food for the group from wet areas, and brought it back to their homes. Or it could be that the contemporary environment was more humid and conducive to water snails than it is today. The recent theories on the possible distances that prehistoric man might have travelled in a day for his sustenance (Vita-Finzi 1978), cannot be applied with certainty to ascertain how far these particular human groups must have migrated to roach wet climatic areas, in search of food. This varies, depending on the quantity of foodstuff available. The fact that no thick layers of ash have been discovered in any cave, gives rise to the hypothesis that those places were not used continuously as dwelling places. Instances of one dwelling being re-settled again are very rare, but at Aligala an abandonment phase is clearly visible between two mesolithic deposits (see Karunaratne and Adikari this volume). The evidence so far gathered is not sufficient to form any definite idea of the prehistoric population pattern and its expansion. Nevertheless the site pattern can be understood very roughly. Factors such as scarcity of samples and suitable scientific dating, the impact of present settlements brought about by urbanization of many of the prehistoric sites, of the sites being covered by thick jungle, of constant disturbances caused to the layers by wild beasts, and the inability to do continuous ex­ cavations (at sites) are responsible for the unconnected ep­ isodes in the story of the prehistory of the area - a saga which has neither a clear beginning nor a clear end. But the indis­ putable fact remains that the earliest settler in the Sigiriya- Dambulla area used microlithic stone implements, and pursued hunting and gathering for his sustenance. REFERENCES Allchin, B. 1958. The late stone age of Ceylon. Journal of the Royal Anthropological Institute. 88(2): 179-201. Correy P.G (1984) 1990 THe Geology of Sri Lanka Colombo: National Museum. Deraniyagala, P.E.P. 1937. Some Fossil Animals from Cey­ lon. Journal of the Royal Asiatic Society, Ceylon Branch 33(88) for 1935:165-168. Deraniyagala, S.U. 1980. Prehistoric Research in Sri Lanka, 1885-1980. P.E.P. Deraniyagala Commemoration Vol­ ume. Colombo: Lake House Book Shop. Deraniyagala, S.U. 1988. The Prehistory of Sri Lanka. Ph.D. dissertation. Department of Anthropology, Harvard Uni­ versity. Cambridge, Mass. Deraniyagala, S.U, and K.A.R. Kennedy. 1972. Bellanban- di-Palassa 1970, a Mesolithic Burial Site in Cevlnn Ancient Ceylon. 2:18-47. Domros, M. 1974. The Agro-climate of Ceylon. Wiesbaden: Franz Steiner verlag Gmbh. Epitawatte, S. 1990. The Physical Environment of the Study Area. The Settlement Archaeology of the Sigiriya-Dam­ bulla Region. Bandaranayake S., M. Mogren and S. Epi­ tawatte (eds.). Colombo: PGIAR. Hartley, C. 1913. The Stone Implements of Ceylon. Spolia Zeylanica. 9(34): 117-123. Hocart, A.M. 1928. Archaeological Summary. Ceylon Jour­ nal of Science. (G) 1(4): 143-164. Karunaratne, P.B. (1989). Report on an Ecological Survey of Forests and Identification of Faunal Remains from Ar­ chaeological Sites around Sigiriya. Unpublished Prelimi­ nary Survey Report. PGIAR Archive. Cat No 89/26. Parker, H. 1909. Ancient Ceylon. An Account of the Ab­ origines as Part of Early Civilization. London: Luzac and Company. Pole, J. 1907. A Few Remarks on Prehistoric Stones in Cey­ lon. Journal of the Royal Asiatic Society, Ceylon Branch. 79(58): 272-81. Premathilake, T.R. and S. Epitawatte. Vegitation Study of the Sigiriya-Ramakale Ruin Sites. Unpublished Report. PG­ IAR Archive. Cat. No. 93/11 Sarasin, F. 1926. The Stone Age of Ceylon. Nature. 117 (2946) 567. Seligmann, C.G. 1908. Quartz Implements from Ceylon 8: 113-116. Vitanage, W.P. 1959. Geology of the Country Around Polon- naruva. Geological Survey of Ceylon, Memoirs No. 1. Colombo: Department of Mineralogy. Vita-Finzi, Claudio. 1978. Archaeological Sites in their Set­ ting. London: Thames and Hudson. Wayland, E.J. 1914. Paleolithic Remains of Kosgolla Estate, Ratnapura District. Journal of the Royal Asiatic Society, Ceylon Branch. 23: 117-119. Figure 3.8 Pidurangala rock seen flrom the east upper cave of the complex seen aboue the three Photo IS Madanayake

Senarath Dissanayaka Location ‘ Sigiriya rock is reached traversing 8km to the left from the Inamaluwa Junction on the Dambulla - Habarana main highway. This world heritage site is located in the Dambulla Divisional Secretary’s Division of the Matale District. History King Dhatusena of Sri Lanka who ascended the throne in 459 C.E. and reigned from the capital at Anuradhapura had two sons named Mugalan from a peer queen and the other son Kashyapa from a commoner queen. Prince Kashyapa who was avaricious for power and wealth, not withstanding that Prince Mugalan was the heir apparent, usurped power and surreptitiously put his father to death. It is said that the monks of Mahavihara abhorred King Kashyapa having heard of the fratricidal act of the king, which according to the Buddhist Canon is an unredeemable sin. It is mentioned in the great chronicle Mahavamsa that King Kashyapa who in a failed attempt to assassinate his brother, fled to Sigiriya to clean the Sigiriya Rock of vegetation and built a flight of rock hewn steps through a lion’s head to the top of the rock and a palace, where he lived like the Kuvera in the mythical city of Alaka. Owing to the flight of steps through the Lion fascia carved into the rock, it was named Sinhagiri or the Lion rock, which with time, evolved to the likable form of Sigiriya. The reign of King Kashyapa from this city was limited to a period of 18 years from 477 to 495 C.E. Prince Mugalan, who fled to India to raise an army, returned to the country to lay siege to the rock fortress Sigiriya and challenged King Kashyapa to war. King Kashyapa realizing defeat at the hands of Prince Mugalan, committed suicide by severing his head with a sword. Thereafter Prince Mugalan donated Sigiriya and the two monasteries in the vicinity to the Bhikkhus of Abhayagiriya and Jetavana monasteries and having cremated King Kashyapa’s remains with royal honours, returned to Anuradhapura to rule the country righteously. Layout ofSigiriya City City of Sigiriya To consider Sigiriya with its inner walls and moats, gardens, inner and outer cities, the palace atop the rock and the Sigiriya Lake to be the epitome of a Sri Lankan example of design combining, simultaneously, the attribute of a city is unequivocal. While the ramparts are fortified by the moats, the rock rises to a height in the midst of the city like a stout obelisk. The area on top of the rock in its entirety is covered with structures belonging to the palace. The area to the west of the rock, when taken as a whole, could be considered to have been used as the royal park. The area enclosed by the outer wall on the eastern side of the rock is believed to be the outer city. Sigiriya Lake is in the south west of the rock. The honour of designing and construction of the above mentioned features of the Sigiriya City can be wholly attributed to King Kashyapa. Before the occupation of Sigiriya by King Kashyapa. the caves located on western slopes of the hill adjoining the Sigiriya rock were used as a monastery by Buddhist monks. The monastic complex consists of 38 caves, which were in use from the 3rd c B.C. to 2nd c. C.E. as evidenced by the epigraphs found inscribed in these caves. According to the great chronicle Mahavamsa, after the overthrow of King Kashyapa, the caves together with the Sigiriya itself reverted to be used as a monastic complex. Moreover, it is stated in the Mahavamsa that Sigiriya remained as a centre for activities of governance from the 5- 7th c. C.E. Ramparts and Moats The sightseer approaching the rock in the easterly direction from the car park of the new Sigiriya village will set his sight on the moat encircling the Sigiriya rock and the rampart leading to the western entrance to the city. Before approaching the western entrance, the visitor has to traverse the middle rampart, constructed in brick. Beyond the middle rampart it could be surmised from the ruins found in its vicinity that a moat had been in existence. When approaching the western entrance, an inner moat and an inner rampart are visible. Roughly cut stone blocks had been used to pave the banks of the inner moat and the foundation of the earthen inner rampart. The length of the ramparts all round the Sigiriya rock add up to about 10km while the total length of the moats is in the region of 8km. Apart from the western entrance to the Sigiriya city, there are entrances across the ramparts and moats on the northern and southern sides as well. During King Kashyapa’s time, there appears to have been a drawbridge to cross the moat leading to the western entrance. Western Middle Rampart Western Inner Moat Western Entrance between the Inner Rampart and the Moat Southern Entrance of Sigiriya City Northern Entrance of Sigiriya City Royal Pleasure Garden When entering the City from the Western Entrance, one sets foot on the Royal Pleasure Garden, built for the recreation of the king, his concubines and the royals. The Royal Pleasure Garden consists of three major sections, namely, the Water Garden, the Rock Garden and the Terrace Garden. No sooner entering the City from the Western Entrance, one steps into the Water Garden. The ponds, the associated pavilions and the parapets surrounding them are the main features of the Water Garden. It is divided into four subunits, each enclosed by a boundary wall. The striking feature of the Water Garden is that its components are designed to be located in perfect geometrical symmetry. A sightseer proceeding in a straight course towards the rock, after entering the City through the Western Entrance, will observe that the ponds and the pavilions seen on one side are symmetrically located on the other side as well. The first subunit that strikes the visitor’s eye, when entering the City through the Western Entrance of the Water Water Garden of the Royal Pleasure Garden Layout of the Water Garden Garden is the Miniature Water Garden. This could be unequivocally identified as the most attractive constituent of the Water Garden. Three building complexes with pavilions and shallow ponds with water up to ankle depth comprise this subunit. In addition, at the far right end of the Miniature Water Garden are two deep water ponds. The boundary walls around the Miniature Water Garden and the other subunit of the Water Garden show evidence to have had eaves in the past. Though it is certain that to the left of the visitor, a structure identical to the one on to his right would have been located symmetrically opposite the Miniature Water Garden, evidence to confirm the same has not been revealed from the excavation conducted in this archaeological site so far. The subunit of the Water Garden, the visitor confronts, when entering the door way is referred to by the same name. This subunit of the Water Garden is divided into three sections by cross walls. The path that lay ahead through the centre of the Water Garden runs through its central section. The central section has four ponds. As these four ponds are in the shape of the letter L, a square shaped island remains at the centre. Evidence indicates that a large pavilion had existed in this island. Generally as all the pavilions in Sigiriya are open and free of walls with their roofs supported on columns, the pavilion on this island may have had a similar construction. Steps have been constructed from the pavilion leading to the water in the ponds to facilitate the royals to reach the water in the pond for bathing. Adjoining the North East and the South West ponds adjacent to the boundary wall are two seats hewn in limestone. One could surmise that the king seated on these seats would have enjoyed sensual pleasure, observing the females of his royal entourage engaged in water sports. On to the north and south of the subunit with four ponds are the two remaining subunits, each having sixteen ponds, totaling up to thirty two ponds. Apart from these, there are four open pavilions, two on each side surrounded by water. Most of the ponds are lined with limestone implying that the bottom of the pond may have been shined with polished marble. Four of the ponds on the side are Layout of the Miniature Water Garder excessively deep, quite inappropriate for bathing or to be used for water sport. However, the other ponds are shallow with water up to little over ankle depth. Although it is difficult to speculate as to what purpose these ponds were put to use, the presence of these ponds in the Water Garden undoubtedly enhances its beauty. From each of the four ponds in the central area are subsurface paths, which provided access to the areas on either side of each pond. The visitor next emerges from the Water Garden subunit into the Fountain Garden subunit, consisting of two long ponds located opposite to each other. These could be the Dirghika or Diguvil, a variety of long ponds, referred to in the ancient times. The flights of steps constructed to facilitate access into them indicate that these ponds had been used for bathing. Proceeding further, one comes across broad and shallow water courses on either side flowing over the surface. Within these water courses are four fountains, two on each side formed by allowing water to squirt through holes constructed in circular sandstone slabs. Pond at the North East Corner of the Central Square of Water Garden Subunit Sub Surface Land Access to the North East Pond The fountains derive their water from two moats, each located on either side of the Fountain Garden. The water from the moat passes through a subsurface drain to the sandstone slab and the hydrostatic pressure of the water squirts vertically up from the opening on the slab in the form of a fountain. Proceeding further, one comes across two sets of twin ponds placed symmetrically on either side of the walkway and two pavilions beyond them. From there onwards one climbs by way of a flight of steps up to the upper precinct of the Fountain Garden to behold a magnificent site of a shallow natural broad expanse of water meandering right across the earth surface. To the north and south of the Fountain Garden are two islands, each surrounded by a moat. Of the two islands, the south island is in good preservation and is known as the Sitala Maligaya. As this island encircled with a moat is at a higher elevation than the surrounding land, it takes the appearance of a small hillock. The island is square in shape and its banks bordering the moat are hewn of stone blocks. It could be surmised by the presence of column foundations made of brick work, found all over the island that a huge pavilion had existed on this island. The structure of the pavilion is designed as a rectangle and there is no doubt that the roof of the pavilion had been supported on wooden columns. As usual this pavilion is open and being on top of the island above the surrounding land of the Fountain Garden would have given ventilation in abundance. As at present the description of it as a Sitala Maligaya or a summer palace could be most appropriate for the reason that the royals engaged in sport would have used to cool off the fatigue in the breeze after their bodily exertions. Though the island to the north of the Fountain Garden is similar to the one in the south, the building has a slightly different design. At the extreme end of the Fountain Garden is a flight of steps giving access to the Water Garden sub unit, which has an octagonal shaped pond. The geometrical symmetry, described regarding the previously mentioned three garden subunits Jacks in this garden subunit. Though the visitor would see to his left an octagonal shaped pond, he would observe on his right, a square shaped pond with earthen banks, while the banks of the octagonal shaped pond are made of stone blocks. The Inner City Rampart and the Rock Garden The wall on the eastern limit of the Water Garden is identified to be the Inner City Rampart. This rampart may have provided added security to the palace atop the rock escarpment. Prof. Senake Bandaranayake is of the impression that on entering through the Inner City Rampart, one sets foot on the land belonging to the boulder Garden. Huge boulders with and without caves are commonly found here. Almost all the boulders have furrows cut on top at their edges to provide foundation construction joints giving the impression that, each boulder had a pavilion built on it. Brick walls have been built using these construction joints and the pavilions were erected on the leveled and enclosed surface. As found in many places in the Sigiriya Garden, there had been open pavilions, erected on timber columns. There may have been sixty such open pavilions in the Sigiriya Gardens, as indicated by the chiseled marks found on the boulders. The structures in the boulder Garden are the Monastery, Deraniyagala Cave, the Audience Hall, Cistern Rock, the Asana Cave or the Cave with a Seat and the Cobra Hood Cave. Fountain of the Water Garden Subunit in the Rainv Season The South Summer Palace Octagonal Pond Furrows Engraved on Rock for the Erection of a Pavilion - The Boulder Garden Water Garden, Boulder Garden, Terrace Garden & the Main Rock The Monastery As one enters the Inner City from the Water Garden, one sees the ruins of a monastery of the later Kashyapa period. It is said that after the demise of King Kashyapa, King Mugalan constructed a temple in Sigiriya and donated it to Thera Mahanama of the Diksanda Seneviya Pirivena. It could be assumed that this monastery had been constructed in the Sigiriya Inner City as it is said that the monastery was constructed on the Sigiriya Rock. Since this is the only structure in Sigiriya City that could be identified as a monastery, it is obvious that it could be the aforementioned monastery. The land occupied by this Pabbata Vihara type Monastery is hemmed between the rampart of the Inner City and a cluster of large boulders. On the eastern side of the monastery premises is a cave presently converted to an image house with a drip-ledged inscription of the pre Christian era under its brow. The cave covered by a wall made of brick had a headless image hewn in lime stone, which is presently lying The Cave Image House of the Monastery Bodhighara in the Archaeological Museum as an exhibit. The Dagoba of the monastery is located on a rocky outcrop near the ramparts of the Inner City. The Dagoba is small in size and in its proximity to the south of the rocky outcrop, a Bodhighara of circular design, paved with stone blocks on an elevated terrace exists. Though Bell surmised it to be a vatadage, the recent excavations conducted by the Central Cultural Fund confirms it to be a Bodhighara. From the Monastery the path leads gradually to a higher level reaching the foot of the rock. Travelling beyond the monastery and climbing a flight of steps one comes across a cave situated on a plane on the left. This cave was named after the scholar named Deraniyagala who recorded its presence originally. I . Deraniyagala Cave. Although the caves in Sigiriya formerly had been prepared with drip-ledges for the use of the Buddhist monks, the conversion of the caves together with the land to take the form of a pleasure garden and the intension to adapt them as a pleasure garden can be observed in the fragments of paintings, which do not appear to be of Buddhist origin. The Sigiriya Asana cave and the Deraniyagala Cave together with the rock face of one of the rocks forming a natural archway of the path leading from the Deraniyagala Cave had figures of damsek painted on them. The aforesaid painting was discovered in the year 1983 by Professor Anura Manatunga. Though these cave paintings are not highly artistic as the frescoes on the Sigiriya rock face, they can be considered clearly as imitations and equivalent in form and style. Hence, these cave paintings could be presumed to have had a worldly origin. A significant number of female figures remain in the Deraniyagala Cave. P.E.P Deraniyagala believed that these frescoes were of the same period as the famous Sigiriya frescoes. Having studied the paintings on the face of Sigiriya and those of the caves, he has brought to light, their similarities and dissimilarities. Though the subject of the paintings in the Deraniyagala Cave is controversial, it is possible to discover some features that demonstrate the connection between the female figures depicted in the paintings and water. In the paintings one finds a damsel having the lotus on a lotus leaf and some damsels holding the lotuses in their hands. It could be assumed that Deraniyagala may have conjectured the connection between the damsels and water being portrayed in these paintings. Deraniyagala states that a pond would have been constructed in the vicinity of the cave by obstructing the flow of water with the aid of the boulders connected with the cave and the paintings depicting the damsels being near this pond confirm his belief. If these damsels could be identified as the damsels of the king’s court engaged in aquatic frolicking, then it equals what was expected from the garden associated with the paintings. It could be surmised that the description in the poem Kavsilumina of the damsels in the king’s harem donning flowers and carrying aquatic flowers while engaged in various water associated activities, not only portrayed by the beautiful bodies of these damsels carrying flowers in the paintings, but also their postural stance. Fragments of twenty two paintings of female figures were discovered by Deraniyagala in 1948. Therefore, the cave was known as Deraniyagala Cave. The technique used in these paintings does not differ from the technique used on the paintings of the main rock. Only ten paintings are in a reasonable state of preservation. The rest are incomplete parts of paintings of female figures. The paintings differ from the paintings of the main rock, which depict only the busts whilst three of these paintings show the full figure representations from head to toe. The rest of the paintings in this cave, unlike those in the main rock, show bodies below the waist of the female figures and at times even up to the knees. The bodies of the females painted in the Deraniyagala Cave are smaller in size than the bodies of those painted on the main rock. Though these females like the ones in the main rock paintings carry flowers in their hands, one painting Deraniyagala Cave and its Paintings showing a female carrying flowers placed on a lotus leaf in her left hand is unique. This feature could be considered more closer to the assumption that they are women engaged in frolicking in water in the midst of aquatic plants rather than celestial nymphs. Paintings of female figures are distributed throughout the inner surface of the cave at various heights. Closer to the cave floor is a painting depicting a type of flora. This factor contributes more to the conclusion that these women are not celestial nymphs with connections to the sky but women of the king’s court frolicking in water sport. One of the paths that lay near the Deraniyagala Cave leads directly to the Main Rock while the one to the right leads to the Cobra Hood Cave. Cobra Hood Cave The Cobra Hood Cave is named thus due to its rock face bearing a cobra headed shape. At a high point on the cave roof, a magnificent and an extraordinary ceiling painting can be seen. The ceiling painting of the cave, which was considered by Bell to have been used as a monastery, Ceiling Painting of the Cobra Hood Cave 32 contains a remnant of a painting similar to the Viyan redder (ceiling cloth) quite a common design found presently in many temples. The painting composed of geometrical motifs and botanical motifs has the image of a rooster or a similar fowl in one location of the design. This painting, a creation of a harmonious combination of the exquisite colours, has some relation to the Anuradhapura period painting on the ceiling of the Situlpavuva cave. Professor Senake Bandaranayake believes that the painting in the Cobra Hood Cave belongs to the late 5th c. CE or early 6th c. CE. By proceeding past the Cobra Hood Cave through a stone archway along the route that radiates upwards from the cave, one comes across three striking monuments, the first of them being the Asana Cave. Asana Cave Next to the rock escarpment of the Royal Audience Hall is the Asana Cave with its picturesque look of rock-cut half walls demarcating its boundary and the raised rock-cut seat inside the cave. Ground Plan and the Cross-sectional Elevation of the Asana Cave (ASCAR 1895) On the back wall of the cave, from a point just near the floor is a tree sculpted on rock, which acts as a back drop to the rock-seat. On the roof of the cave are parts of paintings of female figures probably belonging to King Kashyapa’s reign. On top of the layer of paintings is a slaked lime plaster layer applied during a latter period, on which are two male figures drawn freehand in black. One figure has clasped hands while the other has his right hand in some sort of ritualistic gesture (mudra) and carries in his left hand some indefinable object. These two figures had been drawn long time after the original paintings were drawn probably between the 10th and the 12th centuries. Bandaranayake conjectures from the graffiti written on them that the fragments of female figures may belong to the latter part of the 5th c. CE or early 6th c. CE. He says from the paleographical analysis of the script of one of the graffiti that it may be dated to the 6th c. CE. Audience Hall and the Cistern Rock The two monuments, the Cistern Rock and the Audience Hall Rock are the structures that had been created adjoining the Asana Cave of the boulder Garden. Whether these monuments make up a section of the garden activities is questionable. The Cistern Rock is an indentation chiseled out of a rocky outcrop from the top, vertically down to form a tank with rock pieces placed bonded to each other on edges to form its banks. The banks comprising the bonded rock pieces protrude out of the rocky outcrop. The tank is 13 feet X 10 feet and is 6 feet deep and to allow the surplus water to drain out when the tank is full, overflowing drains have been engraved on top of the rock and on its slopes. The Audience Hall is formed on the second part of the same rocky out crop, on which the Cistern Rock stands and that had been split into two parts in the past. To chisel a smooth textured floor, creaseless half walls and sleek seats on top of a naturally even block of rock, split from the parent rock would have been undoubtedly an easy task. Audience Hall Terrace Garden, Audience Hall and the Cistern Rock after the Excavations done by Bell The Audience Hall comprises a well leveled stone floor with a raised stone seat 15 feet 10 inches wide, 7 feet 10 inches long and 1 foot 9 inches high on its west edge. A half wall is cut in rock on the west, north and south sides of the rock surface having a total area of 60 feet x 37 feet in length and breadth. A flight of limestone steps gives access to the Audience Hall from its south-east corner. The column orifices in the east edge of the hall floor without a half wall infer that a timber fence like structure existed along this edge. This structure was given the name Audience Hall by Bell. He may have come to this conclusion owing to the presence of the large stone seat. The existence of stone seats and halls with a stone seat where the performances of the dancers and singers were watched in the royal gardens suggests, that this may not be an Audience Hall but an appendage of the Royal Garden. When considering the issue in another angle from the evidence available in Sri Lanka that Audience Halls were located in close proximity to the Royal Palaces and the fact that King Kashyapa’s palace was at the top of the main rock, it would be safe to assume that his Audience Hall was closer to his palace. The Sigiriya Frescoes The path with the steps that goes past the Deraniyagala Cave and the flight of steps near the Audience Hall are met at a point near the Mirror Wall. From that point onwards the path leading to the top.of the rock is along the precipitous western edge of the rock. This path from its beginning extends for about 100 meters flanking the Mirror Wall. Having reached the Mirror Wall path and proceeding a short distance on it, one comes across a steel spiral stair case ending up to the rock pocket, containing the world famous Sigiriya Frescoes. The two rock faces above the Mirror Wall containing the frescoes have images of 19 well preserved female figures and 4 partially faded ones. However, from the remnants of plaster and paint visible on the western face of the rock, it could be surmised that a large area of the rock face would have been covered with frescoes. Accordingly an area of 5600 square meters of rock face originally covered with paintings may have prompted John Still to conjecture that this to be the largest painted space in the ancient world. The Spread of Paintings in Pockets A and B and their Numbers, ASCAR 1896 Paintings A 3 and A 4 Paintings B 3 and B 4 - after demaging in 1967 Paintings B7 and B8 Painting B9 Painting BIO Paintings Bit and B12 It is mentioned in three of the graffiti, that there had been 500 paintings. Considering the extent of the surface covered with paintings, this statement is surely not an exaggeration. The colours used in the paintings are limited to red, yellow and green and only the area of the female body above the waist had been the subject of the paintings. The part below the waist is shrouded in a cloud like stratagem. Their complexions are in two categories; golden and the bluish hues. In a number of graffiti, the colour of the bodies of the female figures is described as golden or bluish colour (dark) of the blue lily. While all the women wear a choli to cover their bodies below the waist, some have their bosom exposed. Some of them cover their bosom with veils and some in jackets. Their necks and hands are bedecked in jewelry. Their heads are adorned with headdresses decorated with flowers and floral garlands. All the women carry flowers in Hands of flowers in hands. Some of the identifiable flowers they carry are the lily, the lotus, the campaka and the temple flower. One elderly female carries a garland made of flowers. Generally all the female figures in the frescoes have beautiful faces, full bosoms, slim waists and broad hips as described in ancient literature. The artist has been skillful enough to show the curves of the body and make the figures light and dark. The lines used in the paintings; particularly those used for outlining the drawings are enlivening. These paintings can be placed in the category of the finest examples of paintings of the ancient South Asian art tradition. Although it is acclaimed that the paintings have been executed by those artists trained in the Ajanta art tradition or by the Ajanta artists themselves, scholars such as Benjamin Roland emphasizes the fact that the Sigiriya Frescoes stand out on their own and are no second to Ajanta paintings. Many interpretations have been brought forward about the female figures in the frescoes. Bell was the first to suggest that the frescoes depict ladies from the Audience Hall together with their attendants making a pilgrimage to the Pidurangala temple, situated north of Sigiriya. He also stated that all these damsels in the frescoes had been drawn with their heads pointing to the north. Coomaraswamy was of the opinion that these damsels were celestial maidens. Benjamin Roland endorsing Coomaraswamy stated that the damsels whether as single individuals or as duets represented nymphs or celestial maidens. Dr. Senarat Paranavitana had a completely different theory. He interpreted the damsels with a golden complexion as lightning princesses (vijju-kumari) and those with a bluish complexion as cloud damsels (rnegha-latd). Expanding on the allusion in the Mahavamsa to King Kashyapa as Kuvera living in Sigiriya referring to Alaka, the abode of Kuvera, Paranavitana states that Alaka pur a or Alakamandava being constantly engulfed in clouds and lightening and in an attempt to personalize this sentiment, King Kashyapa had the figures of the damsels in gold and dark shades. Dr. Raja de Silva taking the qui from the German Swamy Gauribala believed that the damsels represented Goddess Tara. Martin Wickramasinghe opines that these damsels portrayed ladies of the Audience Hall engaged in water sport. Considering the poses, attires of these paintings of the damsels and the fact that these paintings are on the western rock face, which is the back drop of the Sigiriya Garden, Wickramasinghe’s theory seems to be more acceptable Just after Kashyapa’s period when most of the original paintings at Sigiriya were available, the poets who incised graffiti at Sigiriya hardly knew about the story behind the Sigiriya frescoes and had not come to any conclusion. Many of the poets addressed them as those they had met in their everyday life or described them as the ladies grieving the demise of the King. Some had no hesitation to call them goddesses. However, from the graffiti on the Mirror Wall, one could glean that after the demise of King Kashyapa, as at present, hordes of people have thronged Sigiriya to see the Frescoes, the Mirror Wall and the Lion’s Head. Having seen the Frescoes, they had inscribed as graffiti on the Mirror Wall, the sensuous thoughts the damsels would have engendered in their minds. The Mirror Wall and the Graffiti on the Mirror Wall . Mirror Wall, presently remains 100m in length. On the sheer surface of the rock face where the slope was less acute, ledges were cut to provide the foundation for the brick mirror wall. The gap between the wall and the rock face had been judiciously filled to provide an alleyway to the rock. The average width of this alleyway is about 7 feet. The Mirror Wall, constructed in brick and plastered with lime mortar is about 9 feet 6 inches in height. The top surface of the alleyway is paved with smoothened Limestone slabs. The Mirror Wall got this name from the high sheen on its inner surface. The sheen on the surface of the mirror wall had been done so skillfully and to an optimum finish, to date it acts like a mirror reflecting the images of the passerby. From the graffiti inscribed on it, the Mirror Wall then was known as ketabita, katbita and kedapat pavura (mirror wall) and presently as ketapat pavyra, all indicating to the term “mirror wall”. While Paranavitana read 685 graffiti appearing on the Mirror Wall and transliterated them, Benil Priyanka read and published 800 of them comprising whole poems, parts of poems and single sentences. The graffiti read by them belonged to the period from 6-7 c. C.E. to the 13 c. C.E. However, as the bulk of the graffiti belongs to the period from 8 c. to 9 c. C.E., it could be assumed that most of the visitors to Sigiriya rock had arrived to admire frescoes during this period. According to the poems in the graffiti there appears to have been frescoes on the rock face below the Mirror Wall. Of the Frescoes, one of the damsels with a veena in her hands, another damsel carrying Asoka leaves in her hands and many carrying flower garlands in their hands as described in the graffiti have been obliterated and not available today for viewing. Those who came to admire the damsels in the Frescoes having seen the Mirror Wall, the Lion’s Head and the Palace on top of the rock may have had the extreme bliss of happiness. It is regrettable to observe the wanton damage some of those of the present generation who visit Sigiriya are causing to the Mirror Wall when they scratch on it, in huge deep letters their name or their initials when comparing the visitors of the yesteryear with sharp instruments inscribed in small letters their poems with the intension of causing least harm to the glazed surface. The other matter of importance regarding the graffiti inscribed on the Mirror Wall in Sigiriya is that the poems of the graffiti can be placed as the oldest poetic literature on record discovered up to the present time. A photograph showing the outside view of the Mirror Wall and the Rock Surface covered with Frescoes in the ancient times In order to give some insight into the language, the subject matter, the writers of the Sigiriya Graffiti and the period they were inscribed, the lines and the translations of a few poems are given below. No. 360 Svasti, TT salami labu kara sii m[ehi] -abuyu-ni [gala] (Vana niva)si Sival-mf ti yaha ko[] mam a gatmi TRANSLATION Hail! *0 damsel; (you) who stood here on the rock. I shall shake your hand which is hanging down; you (who reside in the forest) are (like unto) forest honey’. (Thus). I accepted her as good. The poem belongs to the 8 c. CE. No. 577 Svasti. Kuma a giye vida mata gosin dT gl ri [si sey] Ba[li vi] mehi hindi mun [tay] balimi ata mal gat[ten] TRANSLATION Hail! "Wherefore have you gone to the summit? Having gone (there), wherefore did you. according to your wish, give a song and look at these (ladies) who remained here?’ "I looked at (them) because (they) have taken flowers in (their) hands". Poem belongs either to the 8 c. or 9 c. CE. No. 555 Vidu-ki[da] mata [a]ti da ebimo [di]si-tak di[ i] vT Viyevun band=un beyadihi ran-[vanu]n no [me-vT] bu a TRANSLATION I am Kitala. who came from Mahana-vuta. We peeped (to see) whether there was a streak of lightening above, (and then) was seen all that is worth seeing, (namely) the golden-coloured one who stood on the mountain side, without speaking, in the manner of those separated (from their lovers). Poem belongs to the early 9 c. CE. No. 417 Svasti. E raj'hu a gnak [ballmo] a apa [a]sni Balamin her sey in ba a[nni] at[nen] a hay TRANSLATION Hail! We, who came here, looked at a damsel of that king with our (own) eyes. Looking at (them) in the manner of a roguish woman, she speaks with those who come here by means of her hands.(That is by a gesture of her hands or by hasta mudra) Poem belongs to the 9 c. CE. No. 556 Ve ava a vasi Sivala-batimi bala ba{da gT] Balanana a rama T ahasa poku a ka a biti Gehinuhu sitatama gT-rasa p ke[naka] men no TRANSLATION The sky. the pond and the mirror-wall are delightful to look at. Are not the women in the painting like those who have drunk the nectar of songs? The poem belongs to the 8 c. CE. It starts with. “I am Sivala Bati. a resident of Ve ava a. Having looked at (Sihigiri. this) song was composed (by me)”. No. 19 Tana-ran-malT ve a atani gat ho ran-van li Niridu melen ekal no me be[ ey] anna hay yavat TRANSLATION She. the golden coloured damsel, who (wears) a golden chain on her breast and has taken a lute in her hand, does not speak to anyone else whomsoever, as the king died at that time. Laya Sivala recited this song and I. Mahamet, wrote this. Poem belongs to the 8 c. CE. No. 187 Rata-ata-palu-[ma]lena abula Ia-pataka seyi li me [Bi i]vi me saba[vina] himabuyii no ba anano kT no-vadana branched out through the lion’s mouth. The gigantic size of the Lion’s Head could be gauged from the cut marks left on the rock face at that time to indicate the limit of its construction. The two paws, which remain now of the lion had been constructed in brick and plastered with a thick layer of lime mortar. The visitor commencing his vertical ascent of the rock surface from the lion’s paws, having climbed a short distance assisted by a steel flight of steps enters a section with a slight declivity. The recesses cut on the rock face indicate that an easy route had been created then to climb the rock face similar to the access constructed to reach the mirror wall. By traversing this path, one reaches the flight of steps that leads to the rock summit. Rock Summit and the Royal Palace The summit of the Sigiriya Rock is 3.5 acres in extent. Covering the whole area of the summit are the ruins of the palace and the attached residential park. As the Mahavamsa quite categorically reports of a resplendent palace built by King Kashyapa and the only magnificent structure fit for a king, built in Sigiriya is the one on top of the summit, then undeniably this structure is the royal palace. The graffiti contains a number of poems mentioning a palace in Sigiriya but one such poem was written seemingly when the palace was in the process of decay. The rock summit slopes from west to east and less steeply from north to south. The palace had been designed by taking into account this surface profile. Although most of the structures had been done during King Kashyapa’s time, H.C.P. Bell in his excavations conducted in 1895 on top of the rock has revealed that some of the structures had been done in the later periods. The massive rectangular structure at the highest elevation of the rock could be the main part of the palace. It is quite evident that the building would have had another floor, considered from the remnants of a number of flights of steps originating from the ground floor. Walls of both the ground and the floor above may have been in timber. The rectangular structure on the terrace below the main building could undoubtedly be a part of the palace. On the east of the buildings is a huge pond of 90 feet in length and 68 feet in width, excavated in rock. The north east and south sides of the pond could be identified as the Palace Graden Park. The seat carved in rock adjacent to the pond could be a part of the Palace graden. The foot paths around the palace are paved with Limestone slabs. Layout of the plan palace complex Method of supplying water to the palace on rock summit is uncertain. However, one can be certain that the 5 ponds on top of the summit may have rendered invaluable service in satisfying, to some extent, the water requirement. Palace complex : before excavation (1895-96) Largest Pond before Conservation Largest Pond after Conservation Model of the Royal Palace (Courtesy : Bandaranayake 2010 : 32) Rock Seat of the Palace Garden The Upper Palace

Senake Bandaranayake CONTENTS Plan of Sigiriya 6 Sigiriya - World Heritage city 9 History 10 Historical periods 12 The Sigiriya complex 16 Urban form 16 Inner and outer city 18 Ramparts, moats and gateways 20 Royal gardens 23 Water gardens 23 Moated palaces 24 Miniature water garden 26 Citadel and monastery 28 Boulder gardens 30 Terrace gardens 33 The climb to the summit 34 Mirror wall 35 Lion staircase 37 Paintings and poems 38 Meaning and style 42 Boulder garden paintings 46 Graffiti poems 48 Palace 51 Archaeological finds 53 Souvenir sculptures-‘Art about art’ 55 The Sigiriya earring 58 Inscriptions 59 Cultural landscape: the Sigiriya hinterland 62 Pidurangala and Ramakale 65 Environment and bio-diversity 67 Acknowledgements 70 SIGIRIYA-WORLD HERITAGE CITY One of Asia’s major archaeological sites, the World Heritage city and palace at Sigiriya (the ‘Lion Mountain’) are a unique combination of 5th century urban planning, architecture, engineering, hydraulics, garden design, painting, sculpture and poetry. It has attracted the attention of modem antiquarians and archaeologists since the early 19th century and from the 1890s has been the subject of intense archaeological activity. How this book is arranged After a brief introduction to the site, its history (page 10), general plan and urban layout (page 16), the main features at Sigiriya are described in the order that the visitor would follow when entering from the Western Entrance or the Southern Gate. The sequence is as follows: 1. Ramparts, moats and gateways - page 20 2. Water gardens and moated palaces - page 23 3. Inner citadel wall and monastery complex - page 28 4. Boulder and terrace gardens - page 30 (The paintings in the boulder gardens are described on pages 46-47.) 5. Mirror wall gallery - page 35 (The paintings above the mirror wall and the graffiti poems are described on pages 38-45 and 48.) 6. Lion staircase terrace -page 37 7. Palace on the summit - page 51 There are also short sections on archaeological finds (page 53), souvenir sculptures (page 55), the Sigiriya earring (page 58), ins­ criptions (page 59) and the Sigiriya hinterland (page 62). Surrounded by forests and set in a cultural landscape of remarkable natural beauty and historic interest, the ancient city, palace and garden complex is centred on the Sigiriya rock, an inselberg rising aboutl 80 metres (600 feet) above the surrounding plain. The site itself goes back to pre- and proto-historic times. The area around the main Sigiriya rock was the location of a Buddhist rock-shelter monastery from the last few centuries BC. The present city, palace and garden remains date from the time of their creator, Kasyapa 1 (477-495 AC). Opposite page'. Fig. 2 Inner moat (from north-west comer) HISTORY The history of Sigiriya extends from prehistoric times to the 17th and 18th centuries. Excavations in rock-shelters and the investigation of open air sites in and around Sigiriya have yielded stone and bone tools and human and faunal remains from the prehistoric period. The earliest evidence of human habitation within the environs of the main rock is in the Aligala rock-shelter, which lies immediately to the east of the Sigiriya rock. This is a major prehistoric site with an occupational sequence starting nearly 5000 years ago and extending up to early historic times. It is also one of the sites associated with early iron production in Sri Lanka, dating from around 900 BC. The use of iron is one of the key technologies connected with the proto- historic transition from itinerant food gathering and hunting to village settlement, agriculture - especially the cultivation of millet and rice - and irrigation. The earliest monuments in the Sigiriya region are megalithic cemetery sites, one example of which is Ibbankatuva, near Dambulla. They provide evidence of the presence of early farming and iron producing communities in Sigiriya and its environs, from about 1000 BC onwards. The historical period at Sigiriya begins about the 3rd century BC with the establishment of a Buddhist monastic settlement on the rock-strewn western and northern slopes around the base of the rock. As in other similar sites of this period, partially man-made rock-shelters or ‘caves’, with deeply incised protective grooves or drip-ledges, were created in the bases of several large boulders. There are altogether 30 such shelters at Sigiriya. Several are dated to a period between the 3rd and 1st centuries BC by the donatory inscriptions carved in the rock face near their drip-ledges. The inscriptions record the granting of these caves to the Buddhist monastic order to be used as residences. Starting from its proto-historic origins, the Sigiriya region was to become one of the major centres of iron production in Sri Lanka in the early historic period, especially between the 1st and 4th centuries AC. Also dating from this time is the construction of the Mapagala fortified complex, with its ‘cyclopean’ walls made of massive stone blocks, the first expression of large-scale architectural construction at Sigiriya. Kasyapa I (477-495AC) Sigiriya comes dramatically, if tragically, into the political history of Sri Lanka in the last quarter of the 5th century during the reign of King Dhatusena I (459 - 477) who ruled from the ancient capital at Anuradhapura. A palace coup by Prince Kasyapa, the king’s son by a non-royal consort, and Migara, the king’s nephew and army commander, led to the seizure of the throne and ultimately to the execution of Dhatusena. Kasyapa, much reviled for his patricide, established a new capital at Sigiriya, while the crown prince, his half-brother Moggallana, went into exile in India. The chronicles ascribe a period of 18 years to Kasyapa’s reign, calculated by historians to extend from 477 to 495, although it is possible that he ruled for a somewhat longer period. The king and his master-builders gave the site its present name, ‘Simha-giri’ or ‘Lion-Mountain’, and were re­ sponsible for most of the structures and the complex plan that we see at Sigiriya today. His reign came to an end with the return of Moggallana, and Kasyapa’s defeat and suicide on the battlefield. Sigiriya after Kasyapa This brief Kasyapa episode was in fact the golden age of Sigiriya. In the post- Kasyapa period, Sigiriya reverted to being a monastic centre. Monasteries were located within the Inner Citadel area, and in the neighbouring sites of Pidurangala and Ramakale. The innermost royal precincts, consisting of the palace itself and the access to the summit, seem to have been little used or, more likely, abandoned. The city and its suburbs, however, continued to be occupied, and visitors to the city climbed the rock to view the paintings, the Lion Staircase and the palace. This epoch lasts till about the 13th or 14th century, after which Sigiriya disappears for a time from the history of Sri Lanka. It appears again, first in the reign of Rajasimha I of Sitavaka (1551-93) and then as a distant outpost and military centre of the Kingdom of Kandy, in the 17th and 18th centuries. Modern recovery The city’s name, location and royal connections were never forgotten, but it had reverted to being a small village settlement until the early 19th century when antiquarians, including a scholar monk from the neighbouring temple at Pidurangala, began to take an interest in the site. They were followed several decades later by archaeologists, who have now been working there in research and conservation programs, for over 100 years, since the 1890s. Successive Commissioners of the«Archaeological Department were responsible for directing research, restoration and conservation over the decades. Major contributions were by H.C.P. Bell (mainly from 1894 to 1900, who restores the access to the palace, excavates the summit and maps the entire complex) and Senarat Paranavitana (1930s and1946-54; the doyen of Sri Lankan archaeology, who deciphers the graffiti, identifies and excavates the Water Gardens and the ' western sector, and re-interprets the paintings and the site) The Central Cultural Fund’s wide spectrum Cultural Triangle heritage management program began work at Sigiriya in 1982. This involved the first large-scale, controlled, stratigraphic excavations at the site and extensive conservation and preservation action. It focused attention not only on the best- known and most striking aspects of Sigiriya - the royal complex of rock, palace, gardens and the western fortifications - but also on the entire city and its rural hinterland. Archaeological direction and multi-disciplinary team research were bv the University of Kelaniya Archaeological Team and, later, the university s Postgraduate Institute of Archaeology (PGIAR). A ‘total archaeological landscape’ survey was initiated in 1988. HISTORICAL PERIODS Sigiriya as we see it today is essentially the creation of Kasyapa I (477-495 AC) and his master-builders, but the site has a many­ layered history. Over the millennia it has undergone many transformations due to both natural and human activity. The geological origins of the Sigiriya rock go back to Middle Pre­ Cambrian times, between 3500 and 2000 million years ago. At present, the first traces of human activity are from about 5000 BC. The main rock itself was first occupied during Kasyapa’s reign. As the chronicles record, he shifted the capital from the ancient city of Anuradhapura to Sigiriya, and built the palace on the summit, laying out the fortified city and garden complex around the rock. However, the site has a rich pre-Kasyapa and post- Kasyapa history which has been reconstructed by archaeological investigation, including the most recent research over a 20-year period under the Central Cultural Fund’s Cultural Triangle program. The transformations at Sigiriya cover at least twelve distinct periods of human activity. These provide a working periodisation that can be summarised in the following way : Period 1: Prehistory. Sigiriya and the surrounding area were probably occupicdby prehistoric humans from between 10,000 and 20,000 years ago, although the earliest dates available from present investigations go back to about 5000 BC. Period 2: Proto-history. Significant developments took place in the Sigiriya region during the proto-historic transformation. This involved a fundamental change from itinerant food gathering and hunting to agriculture - especially the cultivation of millet and rice - village settlement, irrigation and the production and use of iron, during the period between about 1000 and 300 BC. Period 3: Early monastic. History begins from about the 3rd century BC. This is marked by the establishment and proliferation of early Buddhist monastic settlements dating from the 3rd to the 1 st century BC. These developed around rock-shelter residences with donatory inscriptions recording the granting of these shelters to the Buddhist monastic order. Period 4: Pre-Kasyapa. The period between the 1st and 5th centuries AC sees the development of large-scale iron production in the area around Sigiriya, and the construction of the fortified Mapagala complex, with its ‘cyclopean’ walls and terraces, located immediately to the south of the Sigiriya rock. Settlement sites from this period have been found on the slopes around the rock itself and in the immediate vicinity of Sigiriya. Period 5: Kasyapa I (477-495AC). This is the major construction phase at Sigiriya. The dates for Kasyapa’s reign are approximate ones, based on the chronicles. He may well have ruled for a longer period, although the material and managerial resources he could command had the potential to construct what we see at Sigiriya today during a reign of 18 years. Nearly two decades of extensive excavation and research since 1982, have given clear evidence of many sub-phases of construction during the overall Kasyapa phase - and also indicate that the complex.was not completed when this major construction effort came to an end. Period 6: Later Monastic A. The first phase of the later monastery period, from the 6th to the early 8th century, sees the foundation of a new Buddhist monastery in the western sector and the Boulder Garden area. It also sees new developments in the monastery at Pidurangala and probably the foundation of the Ramakale monastery to the south. Technologically, the period is marked by the beginnings of the use of dressed gneissic-granites in building construction. The urban complex and the artistic traditions of the ‘School of Sigiriya’, both painting and terracotta sculpture, continue to flourish. The period also has sub-phases of collapse, destruction and rebuilding whose exact chronology and trajectory are yet to be firmly established. The earliest graffiti by visitors to the site date from the last years of the 5th or the early 6th century. Period 7: Later Monastic B. The period from about the 8th to the 10th century is characterised by similar but less clear alternation of phases of development, collapse and reconstruc­ tion. The monasteries of Pidurangala and Ramakale are marked by a high level of building activity. Inscriptions in the form of‘immunity grants’ indicate the power and wealth of these monasteries. The Water Gardens at Sigiriya are modified. The rock, the abandoned palace and the paintings become the focus of great visitor interest. The poetry in the graffiti often displays high literary qualities. Period 8: Polonnaruva period. The period from the 11th to the 13th century is coeval with the rise of Polonnaruva as a major urban centre and the political capital of the country. During this time there is clearly a decline in the construe- i tional activity of the monasteries in and around Sigiriya, ex­ cept in centres such as Dambulla, which enjoy royal patron­ age. New building technologies, however, such as the renewed and improved use of brick masonry, plaster walls and stucco mouldings, and changes in tile and pottery forms, are indi­ cated by surface remains and excavations at Sigiriya, Pidurangala and elsewhere in the region. Period 9: ‘Abandonment’. The period from about the late 13th or early 14th century to the 16th/17th century appears in the present archaeological, epigraphical and historical record as an era of abandonment of the urban and monastic centres at Sigiriya and in the Sigiriya region. The area, no doubt, continued to have its rural settlements and local irrigation systems. The last phase of Period 8 - or the first phase of Period 9 - is marked by the appearance of a final layer of paintings (sketches) in the Boulder Gardens showing an interesting ‘weakening’ of the classic realist style of the Sigiriya tradition. Excavations have shown similar modifications in garden arrangements in the Water and Boulder Gardens. Period 10: The Kandyan kingdom. Sigiriya and its surrounding area appear once again in the historical and archaeological record as part of an outer province of the kingdom of Kandy. It is remembered in Kandyan historical writings as one of the great urban centres of earlier times, and functions as a regional military centre and staging point in the communication network of the kingdom. Period 11: Antiquarian interest. The first phase in the modem recovery ofSigiriya is in the early 19th century-following renewed royal interest in the great temple at Dambulla, and the Buddhist revival of the late 18th century. Many ancient rock-shelter monasteries in the region are re-occupied. A scholar monk at Pidurangala copies the inscriptions in the area, and modem antiquarians ‘re-discover’ the site. Period 12: Modern recovery. Archaeological investigation, restoration and conservation work by the Archaeological Department begins in 1894. This includes major periods of activity under successive Archaeological Commissioners, notably H.C.P. Bell (mainly 1894 to 1900) and Senarat Paranavitana (mainly 1946 to 1954). Work continues from the mid- 1950s during the tenure ofC.E. Godakumbure (late 1950s, 1960s), Raja de Silva (1970s) and Roland Silva (1980s, 1990s). An entirely new phase of multi­ disciplinary team research and extensive preservation and conservation action begins in 1982, under the heritage management master plan of the Central Cultural Fund’s Cultural Triangle program and the archaeological direction of the University of Kelaniya and the PGIAR (Archaeological Director: Senake Bandaranayake, 1982-1999; S. Epitawatte 1999-). THE SIGIRIYA COMPLEX Often wrongly described as a fortress or castle, Sigiriya is in fact a well-planned royal city and a multi-faceted, multi-period settle­ ment, several square kilometres in extent. The entire archaeologi­ cal complex consists of the following elements: 1. The central core around the rock and the palace 2. The western sector - essentially the water gardens 3. The eastern sector - the inner and outer cities 4. The Mapagala fortified complex 5. The Pidurangala monastery 6. The Ramakale monastery 7. The Sigiriya lake (Sigiri Mahavava) and canal system. Urban form The royal complex and the city, ramparts, moats, gateways and gar­ dens of Sigiriya together form one of the best-preserved and most magnificent examples of ancient urban planning and palace and garden architecture in South Asia. The remains of the palace are located on the summit of the Sigiriya rock. The rock and the palace form the centre of an elabo­ rate city complex, consisting of two large urban and garden zones, lying respectively to the east and west of the central core. The original plan of the city was in the form of a regular rectan­ gle with five main sectors, consisting of: (a) the palace and palace gardens on the summit of the rock, with intricate staircases and galleries leading up to it; (b) a fortified inner royal precinct or citadel in the area around the base of the rock, which also incorporated terrace and boulder gardens and, on the highest terrace, the massive Lion Staircase leading to the gallery giving access to the summit; (c) the inner city (or ceremonial precinct); (d) outer city zones to the east of the royal centre, surrounded by suburban settlements outside the ramparts; and (e)the western precinct, laid out as an extensive water gar­ den system with its own palace buildings and pavilions. The city is designed on a square module, approximately 2750 metres from east to west and 925 metres from north to south. The east-west and north-south axes of the module meet at the centre of the palace on the summit. All eastern and western entrances are directly aligned with the central east-west axis, and in the western sector an ‘echo’ plan duplicates the layout on either side of east-west and north-south axes. The city is surrounded by moats and ramparts, including a double moat and triple rampart on the west. In its total conception, Sigiriya presents a brilliant combination of symmetry and asymmetry in a deliberate interlocking of geometrical plan and natural form. It is an outstanding example of mid-first millennium planning mathematics, displaying a high degree of sensitivity to the incorporation of irregular, organic natural features in a plan based on an intricate square module. To the south of the urban zone, as demarcated by the ramparts and moats, lies the fortified Mapagala complex, now found to pre­ date the Sigiriya city. Adjoining it is the city’s main reservoir, the Sigiri-vava, formed by a seven-kilometre long earth dam, fed by local precipitation and a canal system. To the north and south of Sigiriya, respectively, are the Pidurangala and Ramakale monas­ teries, the former predating and also contemporaneous with the royal city. Sigiriya, like the ancient capital at Anuradhapura and the 12th century city of Polonnaruva, provides us with a clear model of a Sri Lankan royal city, with its well defined zonal divisions and many urban elements, such as the palace, royal precincts, ram­ parts, moats, gateways, gardens, hydraulic systems, outlying mon­ asteries and suburban settlements. Inner and Outer City Visitors to Sigiriya walk across the Water Gardens of the western sector and the Citadel and Boulder Gardens on the way to the rock and the summit. Hidden under heavy forest to the east of the rock and the Citadel extend the outlines of the Inner and Outer City. The Inner City measures about 700 metres from east to west and 500 metres from north to south, with a high earthen rampart, gate­ ways and a buried moat. On a low rock outcrop in the centre of the Inner City is evidence of a pavilion, directly aligned with the east­ ward-oriented throne on the summit, suggesting that this area was a ceremonial precinct connected visually and symbolically with the palace on the top of the rock. Beyond this Inner City is the Outer City area, a rectangle about 1500 by 1000 metres. Excava­ tions and surveys in this area have also indicated suburban set­ tlements beyond the walls. We still have no precise idea how these two urban zones func­ tioned internally, or the exact nature of their relationship with the royal citadel, the rock, the palace and pleasure gardens to the west, apart from the clear spatial zoning of the urban plan. It seems likely that this eastern area was the ‘metropolis’ of the Sigiriya complex, where the bulk of the city’s population lived-nobles, officials, traders, craftsmen, soldiers, servants and slaves. Fig. 4 Planning module RAMPARTS, MOATS AND GATEWAYS The limits of the Sigiriya complex are clearly demarcated by an intricate system of moats and ramparts, lying mainly to the east and west, but also to the north and south of the rock. The best- preserved and most elaborate are those of the western precinct, which include a substantially excavated and restored inner moat and a high earth embankment with stone retaining walls forming the inner rampart. A middle rampart extends over a substantial section of the western area. Beyond it are the outer moat and a low earth embankment or outer rampart. The total length of the ramparts around Sigiriya, excluding the stone boundary wall of the Citadel and the Mapagala fortifica­ tions, is about ten kilometres, while the moats, when fully plotted, may add up to nearly eight kilometres. The moats were fed by the man-made lake or reservoir, the Sigiri Vava, lying to the south of the main rock. The earth dam of this great reservoir extends from the Sigiriya rock to the Mapagala and then continues southward for another eight kilometres. Fig. 5 Inner moat and rampart (from south-west comer) Fig. 6 Southern entrance It is possible today to walk the more than one kilometre length of the inner rampart, the most accessible section of which is to the right just inside the western entrance. The inner and middle ramparts had tile-roofed walkways along their length. It is likely that the middle and outer ramparts and the outer moat were incomplete and that construction was interrupted with the sudden and dramatic termination of Kasyapa’s reign. Three excavated and conserved entrances are found in this west­ ern area: the great northern and southern gateways, large enough for vehicular traffic, and the small western entrance or ‘water- gate’. The original access to this water-gate was by a causeway across the outer moat, a wide entrance in the middle rampart and probably a drawbridge across the inner moat. ROYAL GARDENS Sigiriya provides us with a unique and relatively little-known ex­ ample of what is one of the world’s oldest historic gardens, and the oldest surviving large-scale garden form in Asia, whose lay­ out and internal features are still in a fair state of preservation. Three distinct but inter-linked garden types are found here: water gardens, boulder gardens, and stepped or terrace gardens encir­ cling the rock. A combination of these three garden types is also seen on a reduced scale in the palace gardens on the summit of the rock. Water Gardens The Water Gardens are the most extensive and intricate garden form at Sigiriya. They occupy the central section of the western precinct. Three principle gardens lie along the central east-west axis. Garden 1, the largest of these, consists of a central island sur­ rounded by water and linked to the main precinct by cardinally- oriented causeways. The quartered mandala or char bagh plan thus created constitutes a well-known ancient Asian garden form of which the Sigiriya version is the oldest surviving example. The entire garden is a walled enclosure with gateways placed at the head of each causeway. The largest of these gateways, to the west, has a triple entrance. The cavity left by the massive timber door­ posts indicates that it was an elaborate gatehouse of timber and brick masonry, with multiple, tiled roofs. Garden 2, the ‘Fountain Garden’, is a narrow precinct on two levels. The lower, western half has two long deep pools with stepped cross-sections. Flowing into the pools are shallow ser­ pentine ‘streams’ paved with marble slabs and kerbs. Fountains, consisting of circular limestone plates with symmetrical perfora­ tions, punctuate these serpentines. They are fed by underground water conduits and operate on a simple principle of gravity and pressure. With the cleaning and repair of the underground con­ duits, in rainy weather the fountains operate even today. Two relatively shallow limestone cisterns are placed on oppo­ site sides of the garden. Square in plan and carefully constructed, they may well have originally functioned as storage or pressure chambers for the serpentines and the fountains. The eastern half of the garden, which is raided above the western section, has few distinctive features: a serpentine stream and a pavilion with a lime­ stone throne being almost all that is visible today Garden 3 is on a higher level and consists of an extensive area of terraces, halls and pools. In its north-eastern comer is a large oc­ tagonal pool and terrace at the base of a towering boulder, a theat­ rical juxtaposition of rock and water at the very point at which the Water Gardens and Boulder Gardens meet. A raised podium and a drip-ledge for a lean-to roof are the remains of a ‘bathing pavil­ ion’ on the far side of the pool. The central feature of Garden 3 is two ‘L’-shaped pools which define the area in front of the en­ trance to the Citadel. These ‘L’-shaped pools are at the eastern limit of Garden 3. Be­ yond them are the steps of the entrance and massive brick and stone wall of the Citadel. This towering citadel wall, now reduced to less than half its original height, was once a dramatic backdrop to the Water Gardens, echoing an even more dramatic view of the great rock and the palace on its summit further to the east. When seen from the Water Gardens, the wall extends from the tall boulder above the Octagonal Pool to a matching bastion on the south-east formed by wide brick walls, and a series of boul­ ders which surround a cave containing a rock-cut throne. The pools and other water-retainir.g features of the gardens were interlinked by a network of underground conduits, initially fed by the Sigiriya lake and probably connected at various points with the surrounding moats. Moated Palaces The three water gardens constitute a dominant series of rectangu­ lar enclosures of varying size and character, joined together along a central east-west axis. Moving away from this to the wider con­ ception of the western precinct as a whole, we see that its other significant feature is a sequence of four moated islands, at right angles to the central axis of the Water Gardens. These follow the principle of symmetrical repetition or ‘echo planning’ - the two inner islands, on the one hand, and the two outer islands, on the other, forming pairs. Fig. 8 Inner moat and rampart Fig 9 Cistern and vertical drain in Boulder Gardens Fig. 10 Moated palace Fig. 11 Fountains in Water Garden 2 The two inner islands, closely abutting the Fountain Garden on either side, are partially built up on surfacing bedrock. They are surrounded by high rubble walls and wide moats. The flat­ tened surface of the islands were occupied by ‘water palaces’ or ‘cool palaces’. Bridges, built or cut into the surface rock, pro­ vided access. Further to the north and south, almost abutting the ramparts, are the two other moated islands, still unexcavated but clearly displaying the quartered or char bagh plan. Intricately connected with the water gardens of the western precinct are the double moat surrounding it and the great artifi cial lake extending southward from the Sigiriya rock. Miniature Water Garden Just inside the western entrance, to the west of Water Garden 1, is a garden very different in character from those described above. This was discovered during excavations in the 1980s. It contains at least five distinct units, each consisting of pavilions of brick and limestone surrounded by paved water pools and winding wa­ ter-courses. The two units at the northern and southern extremi­ ties are badly eroded, but the general layout of the major portion of the garden and of the three central units is clear. A striking feature of this ‘miniature’ garden (it is in fact about 90 by 30 metres) is the use of water-surrounds with pebble or polished marble floors, covered by shallow, slowly moving water. These probably served as a cooling device, but at the same time had great aesthetic appeal, creating interesting visual effects, en­ hanced by the sound of moving water. In the south-eastern comer of the garden is a deep cistem for water storage. Cavities in its thick brick walls indicate that it once had a structure, probably of wood, which supported a roof or, more likely, a water-lifting device Fig. 12 ‘Miniature’Water Garden Another distinctive aspect of this complex is the geometrical intricacy of the garden layout. While displaying the symmetry and ‘echo-planning’ characteristic of the Sigiriya water gardens as a whole, this miniature garden has a far more complex inter­ play of tile-roofed buildings, water-retaining structures and wa­ ter-courses than is seen elsewhere in Sigiriya. More than one phase of construction is evident here. The garden was originally laid out as an extension and ‘miniaturised’ refine­ ment of the main water garden macro-plan and belongs essen­ tially to the last quarter of the 5th century. Added to in later years and remodelled, it was subsequently abandoned and again par­ tially built over in the last phases of the post-Kasyapa period be­ tween the 10th and 13 th centuries It is almost certain that a similar garden lies buried beneath the lawns of the unexcavated parallel sector in the northern half of the water-gardens, an ‘echo’ or ‘twin’ of this miniature garden. Citadel and Monastery At the eastern extremity of the Water Gardens, are the main en­ trance steps to the Citadel, a walled, inner, royal precinct, cover­ ing an area of about 15 hectares. It has a broadly elliptical plan, more or less defining the outer limits of the hill slopes around the base of the rock. This boulder-strewn hillside has been fashioned into a series of terraces, forming not only a clearly defined and protected area but also boulder and terrace gardens around the rock. The Citadel is entered today from the Water Gardens, over the broad steps of its western entrance. (An alternative modem path­ way also connects it directly with the inner, southern car park.) On either side of these steps are the vestigial remains of the citadel wall. Built, like the steps themselves, of rough-hewn blocks of stone, and originally faced with brickwork and plaster, it also incorporated large natural boulders that lay in its path. The result of many phases of construction and decay, the original 5th century base of this once-towering wall (see also page 24) lay several metres below the present ground level. Immediately inside the entrance to the Citadel is the centre of the post-Kasyapa monastery. The most prominent features here are a bodhighara, a circular temple for a sacred Bodhi tree, a rock-shelter which functioned as an image-house, and the remains of a stupa on top of a boulder. The chapter-house or uposathaghara, which at this period also functioned as the residence of the head of the monastery, is indicated today only by the rectangular out­ line of the building located on a high boulder, presiding over the monastic complex. The basic 5th/6th century plan of this monastery makes it one of the earliest dated examples of a distinct monastery type, referred to in Sri Lankan archaeological terminology as a pabbata vihara. The present remains of the monastery belong to several phases of construction, dating from the late 5th or 6th century to the 12th or 13th century. The rock-shelter image-house was also originally a monastic residence of the early monastic phase in the 2nd or 3rd century BC. It is dated by a donatory inscription of that period. Fig. 13 Monastery and western wall of Citadel Boulder Gardens The entrance to the Citadel and the Monastery is also the entrance to the Boulder Gardens, at their western extremity. These gardens present a design in marked contrast to the symmetry and geom­ etry of the Water Gardens. It is an entirely organic or asymmetri­ cal conception, consisting of a number of pathways which link several clusters of large natural boulders, extending from the south­ ern slopes of the Sigiriya hill to the northern slopes below the Lion Staircase terrace. One of the most striking features of these gardens is their rock- associated architecture, a distinctive aspect of the ancient Sri Lankan architectural tradition. Every boulder here once had a build­ ing or pavilion erected on it and a painted rock-shelter below. What seem to us today like steps and drains or honeycombs of holes on the sides or tops of boulders are in fact the foundations or footings of ancient brick walls and timber columns and beams. The natural surface of the boulder was retained in the middle, between the building on top and the painted rock-shelter below. Fig. 14 ‘Preaching Rock’ Fig. 15 ‘Audience Hall Rock’ and ‘Cistern Rock’ At several places in the Boulder Gardens are the remains of rock­ cut and terracotta water courses and water-retaining structures, in­ dicating that controlled water movement formed part of the garden architecture in this area as well. One of the most visible of these is in the complex of boulders and rock-shelters now referred to as the ‘Audience Hall Rock’ and the ‘Cistem Rock’'. These are two neigh­ bouring boulders, the one fashioned into an audience hall or council chamber and the other supporting a large rock-cut stone cistem or water tank, formed by massive slabs of granite. The Audience Hall boulder has a flattened summit and a large five- metre long throne carved out of the living rock. The adjoining Cis­ tern Rock was originally watered by an elevated aqueduct carrying water from a large reservoir built on the south-western slopes of the hill around the great rock. The reservoir itself was designed to store water collected from the palace on the summit. After the water from the cistem was used, it was drained carefully down the side of the Fig. 16‘Cistem Rock’ and ‘Audience Hall Rock’ Fig. 17 Boulder archway boulder into the paved marble pathway between the two boulders. Above this pathway, a stone lintel and walls connected and prob­ ably provided an elevated access between the two rocks. Two other small stone seats or thrones are also found within this complex. It would appear that what we have here are two buildings used in a court ceremony involving a council chamber and a bathing ritual. Beneath these two boulders are rock-shelters, containing sev­eral layers of fragmentary paintings. The rock-shelter below the cistern also has a large throne or altar and graffiti similar to those on the Mirror Wall. Other remarkable features in the Boulder Gardens are the two boulder archways, with their original limestone staircases, as well as various flights of steps and passageways constructed of pol­ ished marble blocks and slabs. Terrace Gardens Although there is no clear boundary between the Boulder and the Terrace Gardens, the one flowing Into the other at nearly every point, to some extent the two boulder archways mark the transi­ tion from the Boulder to the Terrace Gardens. The latter consti­ tute the third garden form at Sigiriya. The terraces have been fashioned out of the natural hill at the base of the Sigiriya rock by the construction of a series of rubble retaining walls, each terrace rising above the other and running in roughly concentric rings around the rock. Two great brick-built staircases with limestone steps traverse the terraces on the west, connecting the pathways of the Boulder Gardens to the precipi­ tous sides of the main Sigiriya rock itself. In an area not as yet accessible to visitors, are the remains of a third staircase, built partly of timber, which gave access io the rock from the north. The staircases offer clear views of the terraces. It is very likely that the terraces formed the base of elaborate gardens planted with trees and flowering shrubs, but we have little idea today of their origi­ nal character. THE CLIMB TO THE SUMMIT It was with Kasyapa that the great rock - ‘difficult of ascent to humans’ as the chronicles describe it - was made accessi­ ble by elaborate sets of stairways and galleries from the west, the north and along the steep, precipitous sides of the rock. The two staircases, now much restored, which traverse the Terrace Gardens, meet on a landing or lobby, placed almost at the middle of the western face of the rock. This landing was once covered by a roof, evidence of which can still be seen in the beam holes cut into the rock. From this lobby, a covered ambulatory or gallery - the Mirror Wall and the marble-paved pathway it encloses - provided access across the belly of the rock to the ‘Lion Staircase Terrace’ Mirror Wall The Mirror Wall is constructed of brick and plaster. It dates from the 5th century and has been substantially preserved in its original form. Built up from the side of the rock itself with brick masonry, the wall has a highly polished plaster finish from which it gets its ancient name, the Mirror Wall. Visible from a great distance, it is a rare and dra­ matic survival of the construction tech­nique employed at Sigiriya to combine masonry and natural rock. The famous Sigiriya paintings are found in a depression high above the Mirror Wall gallery. The pol­ ished inner surface of the wall contains hun­ dreds of graffiti, de­ scribed on page 48 below. Frag-mentary paintings have also been found on the outer surface. Lion Staircase One of the most dramatic features at Sigiriya is the great Lion Staircase, now preserved only in two colossal paws and a mass of brick masonry surrounding ancient limestone steps. On the terrace in front of the lion are antechambers and courtyards forming the final reception area before the ascent to the summit. Built against the rock face, the Hon itself was a massive architectural sculpture, consisting of the head, chest and paws of a colossal sphinx-like figure. It was originally made of timber, brick and plaster and had an internal staircase. This ‘Lion-Staircase-House’, as it is called in the chronicles, was in effect the gatehouse to the solitary stairway leading to the palace on the summit. The first few flights of this stairway are the limestone steps found inside the body of the lion. The actual structure of the Lion Staircase House itself can be at least partially reconstructed from the evidence still remaining at the site. The lion, so impressive even in its ruined state today, must have afforded a vision of grandeur and majesty when it was intact. Remarkably, we have poems recording its impact on ancient visitors to the site. We saw at Sihigiri the King of Lions whose fame and splendour remain spread in the whole world. Having ascended Sigiriya to see what is (there) I fulfilled my mind s desire and saw His Lordship the Lion From another poem on the Mirror Wall which refers to ‘the face of the great lion,’ it is possible to surmise that the lion was relatively well preserved, at least until the 9th century We know from the chronicle account of Kasyapa’s construction of Sigiriya that the Lion Staircase House was one of the principle features of his plan of the Sigiriya complex. At the same time, it made a major symbolic statement operating on several levels of meaning, enhancing the power and majesty of royal authority and invoking ritual notions of dynastic origins, the lion being the mythical ancestor and the royal symbol of the Sri Lankan kings. Opposite page: Fig. 19 Lion Staircase PAINTINGS AND POEMS The most famous features of the Sigiriya complex are the 5th cen­ tury paintings found in a depression on the rock face more than 100 metres above ground level. Reached today by a modem spi­ ral staircase, they are but fragmentary survivals of an immense backdrop of paintings that once extended in a wide band across the western face of the rock and the Mirror Wall. The painted band seems to have gone as far as the north-eastern comer of the rock, covering an area nearly 140 metres long and, at its widest part, about 40 metres high. As an early writer observed: ‘The whole face of the hill appears to have been a gigantic picture gallery ... the largest picture in the world perhaps’ (John Still, Ancient Capi­ tals of Ceylon, Colombo, 1907:15). Fig. 20 Western face of rock. The original painted area can be seen as a distinct ‘band’ above the Mirror Wall. All that survive of this great painted backdrop are the female figures preserved in two adjacent depressions in the rock face Fig. 21 Apsara Fig. 22 ‘Fresco pockets A and B* Twenty-two numbered figures are shown here in a drawing of 1896 by D.L.G.. Perera of the Archaeological Survey. Oil on canvas facsimiles of the paintings are now preserved in the National Museum in Colombo. There are five figures in Pocket A’ and seventeen in Pocket B\ Of these, eighteen are clearly discernible - twelve are relatively complete and well- preserved (Al, A3, Bl, B5-15),four partly preserved (A4, A5, B2, Bl4); and two partly damaged in an act of vandalism in the 1960s (B3, B4). The rest are extremely shadowy or fragmentary (A2, Bl 5-17), one of which is only rep­ resented by a part of a hand. In recent times, careful scrutiny has revealed another shadowy and fragmentary figure between B2 and B3.Although the term fresco' is still commonly used in referring to th. paint­ ings, it is no longer believed that they really utilised the true fresco technique. known as ‘Fresco Pocket A’ and ‘Fresco Pocket B’. (Three other depressions: Fresco Pockets C, D and E higher up the rock also contain patches of plaster and pigment and, in at least one instance, fragments of a painted figure.) Traces of plaster and pigment else­ where on the rock face and on the outer surface of the Mirror Wall provide further evidence of the extent of the painted band. The figures represent apsaras or celestial nymphs, a common motif in the religious and royal art of Asia. The Sigiriya paintings have been the focus of considerable in­ terest and attention in both ancient and modem times. The poems in the graffiti on the Mirror Wall (discussed on page 48 below), dating from the 6th to the 13th or 14th century, are mostly ad­ dressed to the ladies in the paintings. The paintings also seem to have been studied and reproduced in the 18th century by the Kandyan artists who painted the great mural cycle at the World Heritage site of Dambulla nearby. Antiquarian references to the figures in the ‘fresco pocket’ date back to the 1830s. The first proper descriptions in the 19th century are based on an examina­ tion of the paintings by telescope from the gardens below. The first person to find his way into the fresco pocket and come face to face with the paintings in modem times was a Public Works De­ partment engineer C. A. Murray. He made tracings, copied them in pastel and published a paper about them in 1891 The first real study of the paintings, however, begins with the commencement of archaeological operations at Sigiriya in 1894 by H.C.P. Bell. Facsimile copies in oils were made by Muhandiram D.A.L. Perera of the Archaeological Department in 1896-7. Meaning and style An important but unanswerable question is: how did the present figures relate to the entire composition of the painted band ex­ tending across the rock face? Their fragmentary nature and unu­ sual location have led to the Sigiriya paintings being interpreted in a number of ways, sometimes quite fancifully. Of the proposals that deserve scholarly consideration, the three most important are those of Bell (1897), and the two well-known Sri Lankan schol­ ars, Ananda CoomAraswamy, writing in 1908, and Senarat Paranavitana, half a century later. Fig. 23 Apsaras Bell’s idea that they portray the ladies of Kasyapa’s court in a. devotional procession to the shrine at Pidurangala is a purely im­ aginative construction and has no precedent in the artistic and social traditions of the region or the period. It seems quite likely, however, that the court ladies and their costumes and ornaments provided models for the Sigiriya artists. As such, the paintings would reflect the life and atmosphere, the ideals of beauty and the attitude to women, of the elite society of the time. Paranavitana’s suggestion that they represent Lightning Prin­ cesses (yijjukumart) and Cloud Damsels (meghalata) is an inter­ pretation at once more literary and sociological. It forms part of his elaborate hypothesis which attempts to explain Sigiriya as an expression of the cult of divine royalty, the entire palace complex being a symbolic reconstruction of the abode of Kuvera, the God of Wealth. While these identifications may seem to us today an over-inter­ pretation, too specific to accept in its totality, and deriving from Paranavitana’s attempt to see the Sigiriya palace and royal com­ plex primarily as an expression of divine kingship, they do draw our attention to important sociological dimensions in the under­ standing of ancient works of art. There is no doubt that the spatial organization and symbolism of the Sigiriya complex is profoundly determined by the cult of the king and the ideology of kingship. The great tapestry of paintings at Sigiriya, the palace on the sum­ mit and the lion staircase are all part of a complex ‘sign-language’ expressing royal power and ritual status Coomaraswamy’s identification of the Sigiriya women as apsaras is in keeping with well-established South Asian traditions and is not only the simplest but also the most logical and acceptable in­ terpretation. Recent studies have reinforced this idea showing that apsaras are often represented in art and literature as celestial be­ ings who carried flowers and scattered them over kings and he­ roes as a celebration of victory and heroism. We can say almost with certainty that the Sigiriya ladies are celestial nymphs, very similar in essence to their successors thirteen hundred years later in the ‘Daughters of Mara’ panel in the Dambulla murals. Iiis also likely that they had more than one meaning and function, as ex­ pressions of royal grandeur and status, and as artistic evocations of courtly life, with aesthetic and erotic dimensions. 45 Such an interpretation with its varying levels of ambiguity al­ lows us to accommodate both Bell’s and Paranavitana’s sugges­ tions at either end of a semiological spectrum. It also makes it possible to view the painted band at Sigiriya as a rare and early survival of a royal citrasala or picture gallery, well known in Sri Lankan and Indian literature and described in the chronicle ac­ counts of 12th century palaces and audience halls at Polonnaruva. The style and authorship of the paintings have been as controver­ sial a question as that of their identity. Early writers, such as Bell and even Coomaraswamy, saw them as extensions of the Central Indian School of Ajanta, or several related traditions such as those of Bagh or of Sittanvasal in South India. Bell even suggested that ‘artists trained in the same school - possibly the same hands - executed both the Indian and the Ceylon frescoes’. These were views expressed at a time when very little was known of the ex­ tent and character of early Sri Lankan painting. The American art historian Benjamin Rowland was among the first to observe carefully the actual painterly technique at Sigiriya and to note in what specific way it differed from Ajanta and other subcontinental traditions. ‘The Sigiriya paintings outside of their exciting and intrinsic beauty are perhaps most notable for the very freedom they show at a period when the arts were tending to be­ come more and more frozen in the mould of rigid canons of beauty... The apsaras have a rich, healthy flavour that, in contrast, almost makes the masterpieces of Indian art seem sallow and effete in over-refinement... Just as the drawing is more vigorous than that of the more sophisticated artists of India, so colours are bolder and more intense than the tonalities employed in the temples of the Deccan’ (Benjamin Rowland, The Wall-Paintings of India, Central Asia and Ceylon, Boston, 1938:84, 85). These insights have been pursued and reinforced by contempo­ rary Sri Lankan scholars who rightly argue that while the Sri Lankan paintings belong to the same broad traditions of South Asian art as the various subcontinental schools of the time, the specific char­ acter and historical continuity of the Sri Lankan tradition give it its own distinctive place in the art of the region. Thus the Sigiriya paintings represent the earliest surviving examples of a Sri Lankan school of classical realism, already fully evolved when we first encounter it in the 5th century at Sigiriya Boulder Garden paintings The art of Sigiriya is not confined to the great rock itself. Of equal archaeological and even aesthetic interest, though less well pre­ served and visible, are a number of paintings found in the shelters Fig. 24 Fragmentary paintings from Rock-shelter 7 at the foot of the rock in the area that formed the Boulder Gar­ dens in the time of Kasyapa. As we have seen, this was also the centre of both the ancient and the post-Kasyapa monasteries. Significant fragments of paint-ings can be seen in at least five of these shelters. Many oth­ ers also contain traces of plas­ ter and pigment, indicating an extensive complex of painted caves and pavilions in the whole of the Boulder Garden area. The most ambitious compo­ sition can be found on a large area of plaster in Rock-shelter 7 where there are faint traces of several female figures carrying flowers and moving amidst clouds, very much like the apsaras on the main rock above. Even in ornamentation and gen­eral figural treatment these women are broadly similar to those in the famous paintings, except for the fact that at least three of them are not cut off at the waist by clouds but are full­ figure representations with legs bent in a conventional flying posture. Altogether there are about half a dozen identifiable forms here barely discernible in traces of body colour and linework. Fig. 25 Details from ceiling paintings in Rock-shelter 9 The most extraordinary, and certainly the most dramatic mani­ festation of the painter’s art at Sigiriya, are the remains of ceil­ ing paintings in the rock-shelter popularly known as the ‘Cobra­ hood cave’ (Rock-shelter 9) on account of its equally dramatic rock formation. The shelter itself dates from the earliest historical phase of occupation at Sigiriya and bears a donatory inscription belonging to the last few centuries BC. The painting combines geometrical shapes and motifs with a free and complex rendering of characteristic volute or whorl motifs. It is nothing less than a masterpiece of expressionist painting, displaying considerable imaginative range and artistic virtuosity in a way not seen else­ where in the surviving paintings of the Sri Lankan tradition. The characteristic brushwork style and tonal qualities of the Sigiriya school are immediately noticeable here. There is little doubt that this awning is contemporary with the paintings on the main rock Considered in their totality the paintings in the Boulder Gar­ dens at Sigiriya, though vestigial, provide important evidence of the continuation of the Sigiriya school over a fair period of time. Excavations have shown several post-5th century phases of oc­ cupation in the rock-shelters in this area, continuing until per­ haps as late as the 12th or 13th century. This situation is paral­ leled by the layers of plaster and painting which provide evi­ dence of several successive phases of painterly activity at Sigiriya. Graffiti poems The Sigiriya paintings have preoccupied visitors to the site over many centuries. After the abandonment of the palace in the 5th or 6th century and the establishment of a monastery in the Boulder and Water Gardens to the west of the rock, Sigiriya became a place of secular pilgrimage for visitors from all over the country who came to see the paintings, the palace and the Lion Staircase. Greatly inspired by the paintings, they composed poems, mostly addressed to the ladies depicted in them, and inscribed their verses on the highly polished surface of the Mirror Wall just below the painting gallery. Known as the ‘Sigiri graffiti’ and dating from about the 6th to the early 14th century, hun­ dreds of these cover the surface of the wall and some of the plastered sur­ faces in the rock-shelters below. Nearly 700 of these writings have been deciphered by Paranavitana (1930s to 1950s) and another 150 recently by Benille Priyanka (1990s). The poems ex­ press the thoughts and emotions of ancient visi­ tors to Sigiriya. They contain not only reveal­ ing comments on the paintings themselves and other remains at the site, but also provide insights into the culti­ vated sensibilities of the time and the contem­ porary appreciation of art and beauty. Fig. 26 Rubbings of selected graffiti poems. For readings see S. Paranavitana The Sigiri Graffiti London, Oxford University Press, 1956: Nos. 133, 197, 208, 42, 75 (reading from top to bottom) Opposite page and page 50: see pages 70-71 below for sources of readings and translations of poems Their bodies' radiance like the moon wanders in the cool wind... The song of Lord Kital: Sweet girl standing on the mountain your teeth are like jewels lighting the lotus of your eyes talk to me gently of your heart. I am Lord Sangapala I wrote this song: We spoke but they did not answer those ladies of the mountain they did not give us the twitch of an eye-lid. Are they frightened the ladies with the golden skins that they stand so silent? This long-eyed girl says nothing but a flower flaunts in her hand. Lovely this woman excellent the painter! And when I look at hand and eye I do believe she lives. We being women sing on behalf of this lady. You fools! You come td Sihagiri and inscribe these verses (yet) not one of you brings wine and molasses remembering we are women. This divine maiden is like the centre of a flower cleansed by moonbeams... Your hair floats not on gentle breezes their welcome blowing. Why spend you the cold night thus grieving? O beautiful doe-eyed lady is it because Sigiriyas lord left not for you his gift so dear Ladies like you make men pour out their hearts and you have also thrilled the body making its hair stiffen with desire. The long-eyed women are parted from their lover, are grieving for the king. Their eyes are blue lotuses. The song sung to the painting The wind blew Thousands and hundreds of thousands of trees which had put forth buds fell down. The curlew uttered shrieks. Torrents came forth on the Malaya mountain. The night was made to be of the glow of tender copper-coloured leaves by fireflies beyond count. O long-eyed lady the message offered by you what sustenance does it afford? Is it not our lord ...who has returned to the palace he once lived in? ...is he going away again seeing that we remain behind? Having seen them (the ladies in the paintings) one is not content... see also, indeed, the mansion in which they lived so happily. Having heard... that the mansion of the king is said to please the mind I looked at it (but) even if I had as many eyes as there are stars I would have found no pleasure in it whatsoever! (With) tears in my eyes I saw how a king had lived as he pleased on the mountain inaccessible lofty as the sky. PALACE The summit of the Sigiriya rock is a stepped plateau with a total extent of around 1.5 hectares. A brick-walled staircase originally gave access to the summit. It began within the lion gateway and probably had limestone stairs and a tiled roof. Footings for the brickwork, cut into the rock face above the lion, indicate the path of the upper sec­ tions of this stair­ case. The palace was the centre of the royal city At its highest point it lies 180 me­ tres above the sur­ rounding plain and 360 metres above mean sea level. Not only the loftiest and innermost precinct of the Sigiriya complex, it is also the geo­ metrical centre of the modular grid on which the plan of Sigiriya was based. The central north­ south and east-west axes of the entire for­ tified complex inter­ sect near the mid­ point of the palace area. Located close to this centre is a massive rock-cut throne, which faces the inner city and ceremonial precinct to the east of the rock. The earliest surviving example of a royal palace in Sri Lanka, with its layout and basic ground plan clearly visible, it provides important comparative data for the study of Asian palace forms. Fig. 28 Plan of palace The palace complex divides into three distinct parts: the outer or lower palace, occupying the lower eastern part of the summit; the inner or upper palace, occupying the high western sector; and the palace gardens to the south. All three sectors converge on a large and beautiful rock-cut pool bordered on two sides by a stone- flagged pavement. This garden area is a combination of water, terrace and boulder gardens on a smaller scale than the great royal gardens below. A marble-paved walk runs like a spine down the centre of the complex. It separates the upper palace from the lower palace, forming an axial north-south corridor leading directly to the throne. The summit has a number of water-retaining structures for rain­ water harvesting. Extensive tile and nail remains found in recent excavations are evidence that the buildings here were roofed. It is likely that rainwater run-off from these roofs provided the main water source on the summit. A wide rock-cut drain, running north­ south along the western edge of the rock and then vertically down the rock-face, carried surplus water from the summit to a large reservoir built against the south-western comer of the rock. As described on page 31 above, an aqueduct from this reservoir car­ ried water down to the ritual bathing tank and royal audience hall in the Boulder Gardens. Footings cut into the rock-face indicate that a wall, similar in principle to the Mirror Wall, encircled the summit along its edges. The roofs, pavilions and upper storeys of the palace would have risen above this wall and been clearly visible from the gardens and the plain below. The palace on the summit and the great lion presided over the surrounding countryside, a powerful expression of both actual and symbolic'royal authority and control over ‘a landscape of power’, radiating across the territory of the Sigiriya kingdom. ARCHAEOLOGICAL FINDS With more than a century of investigation at the site and intensive research over the last two decades, a great deal of archaeological information and material has been gathered at Sigiriya. Most of it relates, of course, to the major architectural remains and the layout of the city, palace and gardens. Smaller finds, such as pottery, beads, coins, sculpture, roof tiles, drain pipes and a variety of iron objects, give a vivid picture of the way the site was used and also help to determine the chronology of various levels and phases of activity. As in most archaeological sites, a huge collection of potsherds constitutes the bulk of finds. This enables us to construct a de­ tailed occupational sequence in various sectors of Sigiriya. Among the most interesting finds are the cups, bowls, tray-bowls, serving dishes, lids, cooking and storage jars and flower pots from the period of royal occupation. Import ceramics, although relatively rare, and large numbers of Roman coins, are evidence of the Sigiriya kingdom’s trade and exchange contact with both East and West Asia. Fig. 29 Some pottery forms (derived from excavated vessels and potsherds) Recent excavations in the Sigiriya area have shown it to be a major centre of iron production. This may partly explain the ex­ tensive finds of iron nails, some more than 30 centimetres long, and iron bands and other artefacts used in architectural construc­ tion, as well as agricultural and military implements. Sigiriya’s largely timber architecture has long since disappeared. Its shadow, however, is found everywhere, not only in the post-holes and beam­ holes cut in the living rock or surviving as hollow spaces in brick masonry, but also in the iron elements referred to above and in elegant terracotta architectural detailing. Rooftiles have also been found in large quantities in most parts of the site. Two interesting chance finds many years ago, are the Sigiriya ear ornament, an extremely rare and unique example of Sri Lankan jewellery, and a blue-glazed jar of West Asian origin. Souvenir sculptures: ‘Art about art’ It is clear that Sigiriya was for many centuries a centre of artistic and literary activity, as evidenced by the multiple layers of paint­ ing and the graffiti poems. Closely connected with the paintings and the poetry is a series of miniature terracotta figurines found in the de­ bris of collapsed structures in the Boulder Garden area on the western slopes at the base of the rock. These are among the most interesting ar­ chaeological finds from the Cultural Triangle excavations at Sigiriya. Most of the figurines are female tor­ sos rendered in the familiar ‘classic­ realist’ style of the Middle Historical period (circa 6th to 13th century). The modelling shows a characteristic con­ cern with three-dimensional form and a sensitivity to both anatomical and decorative detail. From their archaeo­ logical context and style we may ten­ tatively date them to a period between the 7th and the 10th centuries. As far Fig. 31 Souvenir terracotta sculptures Fig. 32 Terracotta architectural ornamentation as we know, terracotta figures of this specific type have only been found at Sigiriya but they are clearly related to a contemporary tradition of fine terracotta architectural ornamentation and sculp­ ture associated with Sigiriya and other sites in the region. It is particularly interesting that these figures are representations or models of the famous apsaras of the Sigiriya paintings. The concept of the unity of sculpture and painting (i.e. the equivalence of the three-dimensional and the two-dimensional image) is a ba­ sic principle of South and Southeast Asian art. What is rare, per­ haps even unique, at this early period is to find works of art which are deliberate representations, or in this case actual models or mini­ ature reproductions, of other works of art — a process which can be described as ‘art about art’ The correspondences between the paintings and the sculptures and the diminutive size of the latter (usually between 10 and 20 cm) suggest that the figurines were portable objects and not part of any fixed architectural decoration. This further supports the notion that they are models or ‘souvenirs’. The production of models and souvenirs to be carried away by pilgrims visiting fa­ mous religious centres is of course an ancient practice well-known in the art and archaeology of Asia. Sigiriya, however, is an exam­ ple of a site rare in the archaeological record that seems to have been visited purely on account of its secular aesthetic and ‘ar­ chaeological’ attractions. Just as the verses are mostly addressed to the ladies in the paint­ ings, the terracotta figurines seem to have been produced as sou­ venirs to be taken away by visitors who appreciated the paintings. This interpretation is preferable to one that would view the figures as decorative or iconic sculptures associated with the monastic structures among whose debris they were found. The Sigiriya torsos, like the poems on the Mirror Wall, are un­ doubtedly an expression of ‘art about art’. They interest us not only as beautiful terracotta sculptures but also as unique historical documents supplementing the insights we gain from the poems into the society and sensibilities of the period. The Sigiriya earring An extraordinary find associated with Sigiriya is a unique ear or­ nament, now in the National Museum in Colombo (acquired 1906). Made of solid gold, it is 6.4 cm in length, excluding the pen­ dant stones. Its body is formed almost entirely of a series of intricate vo­ lutes, bursting out like flames or wave-crests, most of them terminating in a small hook and nodule. The volutes are symmetrically arranged on either side of a central cluster of semi-precious stones, of which only a large, milky white quartz remains in place. Three pendant stones complete the design. At the top, a smooth loop with scrolled ends has a sliding gate device which can be opened to attach the orna­ ment to the ear of a per­ son or a statue. The ear­ ring cannot be precisely dated but probably be­ longs to the period be­ tween the 8th and 10th centuries. Its ornamental detailing and sliding gate device are evidence of its Sri Lankan origin. Exactly the same volutes appear in the Sigiriya paintings. As a piece of jewellery, however, it has no known correspondences in Sri Lanka or anywhere else in Asia. .Fig. 33 The Sigiriya ear ornament Inscriptions Sigiriya and the Sigiriya region have yielded a rich collection of inscriptions, dating from about mid-1 st millennium BC to the 18th and 19th centuries. They provide a complementary and sometimes alternative framework to what appears in the archaeological record, and throw light on changing values and the varying nature of what was considered important in social, economic and political life from epoch to epoch. The first examples of writing on stone in this area are the sym­ bols or letters engraved on two capstones from the tombs in the megalithic cemetery at Ibbankatuva, dating from around 500 BC. Fig. 34 Inscriptions: (a) Ibbankatuva, megalithic symbols: circa 500 BC; (b) Sigiriya, early Brahmi: 3rd-1st century BC; (c) Sigiriya, later Brahmi: lst-2nd century AC; (d) Unulugala, near Sigiriya, later Brahmi: 3rd-4th century; (e) Sigiriya, early cursive: 6th-7th century The earliest and most prolific writings that have meaningful texts, rather than single symbols or letters, are the donatory in­ scriptions placed near the drip-ledges of rock-shelter monaster­ ies, recording the gifting of the rock-shelters to the Buddhist mo­ nastic order. Dating between the 3rd and 1st centuries BC, they use the universal Brahmi script found throughout South Asia. They provide a great deal of information regarding the social and po­ litical conditions and the names, titles, interests and values of the social elite of the time. Less numerous than these drip-ledge records is a series of in­ scriptions, mostly from the 1st and 2nd centuries AC, indited on the rocks themselves rather than near the drip-ledges. They record donations to Buddhist monasteries by an office-holding elite in­ cluding kings, provincial rulers and army commanders. We see in them the increasing centralisation of political authority, the devel­ opment of irrigation and agriculture, and the growing wealth of the monasteries. A third category is a handful of records from the 5th to the 7th century, a period which sees the transition from the angular letter forms of Brahmi to a much looser and progressively cursive script. These inscriptions, unlike those in the previous category, are not by kings or high officials. They contain the enigmatic term ‘vaharala’, whose interpretation is a subject of some controversy. Some authorities connect it with the manumission of slaves, oth­ ers with craft occupations, the ordination of monks, or donations to monasteries. The earliest graffiti on the Mirror Wall also be­ long to this phase. From an epigraphical point of view the next group of records, the slab or pillar inscriptions of the 9th and 10th centuries, are important and extensive state documents, royal grants and legal edicts, promulgated by royal authority and erected by provincial officials. Although few in number, they give us a complex image of the legislative, executive and judicial machinery of the time. The period from the 8th to the 10th century is also the ‘Golden Age’ of the Sigiri graffiti. The largest number of well-preserved texts and the highest literary qualities are found in the writings on the Mirror Wall from this era. There are scarcely any records of significance in the Sigriya re­ gion after the 10th century, with the exception of an elaborate royal inscription at Dambulla, from the last decade of the 12th century, and the Mirror Wall graffiti, which continue into the early 14th century. The only other later period inscriptions that have been found are an incomplete royal edict of the 18th century and records of donations to the temple at Dambulla in the 19th and 20th centuries. CULTURAL LANDSCAPE: THE SIGIRIYA HINTERLAND The city, palace and gardens at Sigiriya, the Mapagala fortified site, the Sigiriya lake and canals and the immediately neighbouring mon­ asteries are all set in a rich natural and cultural landscape, impacted and fashioned by humans over centuries. Total archaeological land­ scape surv eys have given us a broad picture of the archaeology of an area of more than 1000 square kilometres around Sigiriya. This includes prehistoric sites, megalithic cemeteries, Buddhist monas­ teries, rural settlements, village tanks, and iron-production centres. The monasteries of Pidurangala and Kaludiya Pokuna and the World Heritage site of Dambulia are easily accessible to visitors, as is one of the megalithic tomb clusters at Ibbankatuva near Dambulla. The forest monastery and the natural mountain reserve at Ritigala, with its unique flora, is also open to visitors. Not too far away are the ancient capital cities of Anuradhapura and Polonnaruva, the archaeological sites at Manikdena and Madirigiriya, the great man-made lake at Kala Vava and the co­ lossal Buddha images at Avukana and Sasseruva. Fig. 36 The ‘Greater Sigiriya’ area Fig. 37 Sigiriya archaeological landscape Pidurangala and Ramakale As in most urban sites in Sri Lanka, and especially in royal cities, an integral part of the Sigiriya complex are the two outlying Buddhist monasteries of Pidurangala to the north and Ramakale to the south. The monastery at Pidurangala, like Sigiriya, dates back to the earliest phase of the Early Historical Period when it was a rock­ shelter monastery with 14 shelters, four of which have donatory drip-ledge inscriptions of the 3rd to 1 st century BC (Kodituwakku and Karunaratne, pers. comm. 2002). The shelters also contain evidence of prehistoric occupation. The complex as it is today is in three sections. At the base of the rock is the contemporary temple, whose image-house is a large ancient rock-shelter of the Early Historical Period, with a recumbent image and two standing images of considerable antiquity, but now restored and painted over. Associated with this rock temple are a number of buildings of the modem monastery, at least one, a monastic residence, dating-from the Kandyan Period in the 18th century or early 19th century. On and around the great rock are rock-shelters which were used since prehistoric times. In the Early Historic Period they became monastic dwellings and were later converted into shrines. The largest and most accessible of these is the great ‘cave temple’ more than halfway up the rock, in a rock shelter 90 m long, divided into 13 compartments. The temple has remains belonging to several periods, including one of the earliest dated colossal recumbent images of the Buddha, dating back at least to the 5th century, but now restored. The image measures 12.5m. The summit of the rock presents an extraordinary natural and cultural landscape. It is marked by a distinctive‘dry tropical’micro-environment with an unusual floral composition, and offers dramatic views of the north face of the Sigiriya rock and the Lion Staircase, against the distant backdrop of the central mountains. The remains of a votive stupa preside over the summit and the surrounding countryside. On the plain below the Pidurunagala rock is a major monastic complex of the same pabbata vihara type as the monastery at the entrance to the Boulder Garden at Sigiriya - with a stupa, a Bodhi- tree temple, an image-house and a chapter house. At Pidurangala we also have a central mandapa or shrine. The earliest phase of this monastic complex dates from at least the 5th century. The stupa itself is thought to have been built by Moggallana I (circa 495-512) over the site of King Kasyapa’s cremation. New inscriptional readings show that Pidurangala may be the ancient Dalha Vihara. There is enough archaeological evidence to accept the traditional ascription of the pabbata vihara itself and the great recumbent image to the time of Kasyapa and Moggallana in the 5th and 6th centuries. There are at least 12 inscriptions at Pidurangala dating from the 5th to 7th century period, while the style of the surviving brick-and-plaster architectural detailing in the rock-shelter housing the recumbent image indicates a 10th or 12th century date. Pidurangala therefore has a history which takes us from prehistoric times to the 18th and 19th centuries, and it still remains a living religious centre. The monastic complex at Ramakale, on the other hand, which lies to the immediate south-west of Sigiriya, is largely unexcavated and no longer in worship, but is known from epigraphical evidence to be the ancient Mahanaga Pabbata Vihara. The name is found in two 10th century inscriptions, with similar contents. They contain the text of an ‘immunity grant’, promulgated by three officials, in the name of the king-in-council, giving the monastery and its occupams immunity from taxes and from the authority of royal and regional officials. They also give the right of sanctuary to those who ‘having committed the five great crimes’ take refuge in the monastery. Today the complex consists of about 120 buried structures and an excavated and part-conserved stupa. Among the interesting finds from the monastery is a Meru stone, a symbolic representation of the cosmic mountain, depicting in registers what appear to be divine figures, devotees and scenes from domestic and rural life. The stupa is clearly visible from the approach road to Sigiriya when coming from the main highway at Inamaluva. Behind and around the stupa is the forested precinct, about a square kilometre in extent, containing a series of mounds and water pools, concealing the buildings and water features of the monastery. Environment and bio-diversity The Sigiriya area is also rich in environmental resources, intricately linking cultural and natural landscapes. The area immediately around Sigiriya is a nature reserve, highly forested, especially to the east of the rock. The Greater Sigiriya area extends from Kandalama in the south­ east to Polonnaruva on the east and Ritigala and beyond to the north-west. Much of this region is an extensive tract of Dry Tropical forest, rich in flora and fauna, and one of the most intensively studied areas in the Sri Lankan Dry Zone from the point of view of bio-diversity research. Many mountain ranges and ridges, such as the Kandalama mountains (Eravvalagala-Dikkandahena- Tittavalgolla) and the Konduruvava and Sudukanda Ridges, forming the immediate backdrop to Sigiriya, are covered with primary forests. Sigiriya has a new botanical garden, begun in the 1990s, and open to visitors on request. It contains a medicinal plant collection and rare, endangered and common plants and trees native to the region. On the Dambulla-Kandalama road lies the IFS-Popham Arboretum (belonging to the Institute of Fundamental Studies), which can also be visited with prior permission. Sigiriya and nearby Habarana are well-known for their avifauna, and are the regular haunts of birdwatchers. The archaeological park at Polonnaruva is also a nature reserve, and has been a centre for primate studies and animal behaviour research for nearly three decades. The Mahavali ‘Elephant Corridor’ lies across the Sudukanda and Konduruvava mountain ranges immediately to the east of Sigiriya, and leads ultimately to the Minneriya-Giritale Sanctuary and the Wasgomuva National Park. ACKNOWLEDGEMENTS This book is based on archaeological research at Sigiriya over the last twenty years by the University of Kelaniya Archaeological Team and the Postgraduate Institute of Archaeology, directed by the present author under the Central Cultural Fund’s Cultural Triangle Program. This has been a team effort by large numbers of people at various levels, too numerous to name individually. They all deserve due credit for the information and knowledge contained here. Our work at Sigiriya also draws on the cumulative effort of those archaeologists, technical personnel, craftspersons and labourers who worked at Sigiriya for nearly ninety years before 1982, when our program began. A special acknow­ ledgement is due to Dr. Roland Silva who, as Director General of the Central Cultural Fund, invited the present author to be the Archaeologi­ cal Director of the Sigiriya Project, which was inaugurated on 01 Janu­ ary 1982, and who continued to support, encourage and inspire our work at Sigiriya through the ensuing decades. Special thanks for assistance during the preparation of this book are due to: G.F. de Alwis, Albert Dharmasiri, Manel Fonseka, J.A.D.S. Jayaweera, Kusumsiri Kodituwakku, Sirinimal Lakdusinghe, Benille Priyanka, Raj Somadeva, Rose Solangaarachchi, I.S. Madanayake, Gamini Jayantha Mendis and Ramani Withana. The PGIAR librarians, Sanjeeva Perera, Inoka Dilhani and Rasika Murage, helped with library searches, and T.T. Souijah with coordination. Gnana Maldeniya and Chitra Dambadeniarachchi of the CCF Publication Unit saw the book through the press. Most of the photographs and drawings were done under the direction of the author. Photography is mainly by Maya Upananda with the following exceptions: Half-title and Fig. 33: I.S. Madanayake; Frontispiece: Gamini Jayasinghe (Courtesy Lake House Bookshop); Fig. 18: Robert Swarthe; back cover: Juergen Schreiber. Maps, plans and drawings are by Jinadasa Welikumbura, Asoka Perera and Resta Fernando, except for Figs. 14,15,22 and 27, which are from H.C.P. Bell’s Archaeological Survey Reports. Fig. 20 is by Malinga Amarasinghe. The graffiti poems on pages 37,49 and 50 are translations or adap­ tations from several sources. The original readings are those contained in: Senarat Paranavitana, The Sigiri Graffiti (London: Oxford Univer­ sity Press, 1956), Benille Priyanka, ‘Readings of hitherto unread graffiti on the Mirror Wall and^jana Guhava at Sigiriya’ in S. Bandaranayake, M. Mogren and S. Epitawatte, The Settlement Archaeology of the Sigiriya-Dambulla Region (Colombo: Postgraduate Institute of Archae­ ology, 1990); Benille Priyanka, ‘Sigiri graffiti: new readings’ in S. Bandaranayake and M. Mogren, Further Studies in the Settlement Ar­ chaeology of the Sigiriya-Dambulla Region (Colombo:Postgraduate In­ stitute of Archaeology, 1994). English renderings based on Paranavitana’s primary readings and translations are found in W.G. Archer and S. Paranavitana, Ceylon; Paintings from Temple, Shrine and Rock (New York and Paris: New York Graphic Society and UNESCO, 1958) and Ashley Halpe, Sigiri Poems (Peradeniya: Peradeni Prakasana, 1995).The versions in this book have sometimes been adapted or re-arranged to try to convey the poetic quality of the Sinhala verses more effectively, but the changes involved are slight. The interested reader should consult Paranavitana’s (1956) and Priyanka’s (1990,1994) original readings and translations, as well as the renderings of Archer (1958) and Halpe (1995). The sources of the poems reproduced here are as follows: page 37:Paranavitana 1956: Nos. 45,205; page 49, first column: Archer 1958:8- 9, 13; Halpe 1995: Nos. 12,40 (Paranavitana 1956: Nos. 11, 52, 17, 60, 99); page 49, second column'. Priyanka 1994: No. 101, 1990: No. 12, Archer 1958: 7-8; (Paranavitana 1956: Nos. 69, 124, 595); page 50: Paranavitana 1956: Nos. 165, 156, 376.

EVA MYRDAL-RUNEBJER, DAMAYANTHI GUNAWARDENA, SUDARSHANI FERNANDO* * Damayanthi Gunawardena was responsible for the study of rural architecture and village layout. Owing to her sudden death in the field season of 1991, this work remains unfinished. We wish however to publish the material from the initial study, hoping it can be of use for more in-depth studies in the future. Help from Sudarshani Fernando and Manjula R. Sirisena of the SARCP team enabled the Talkote and Nagalavava house plans and the Nagalavava interviews to be completed. The PGIAR cartographic unit, together with Priyantha Karunaratne of the PGIAR undertook completion of the village plan of Nagalavava. Sarath Padmakumara and Sugath Jayasinghe, surveyors attached to the Sigiriya Project of the Cultural Triangle mapped Talkote village in 1989. A few studies of Sri Lanka’s rural houses and household equipment have been published, but not specifically from our study area.1 In order to gain a better understanding of what activities could be related to a rural household in the area, and how they and their dwellings could be identified in space and from their associated raw material, tools and refuse, we initiated a study of village layout, house plans and activity areas in the house and courtyard. Plans were made of three houses in Talkote village, and of all five houses in Nagalavava gammandiya. The overall plan of the villages was also mapped. Though the sample is small, it is however possible to see significant differences in the house plans. It is possible to see a connection between relative material wealth and the house plan in the specialized functions of rooms, and building ma­ terial used. House building labour traditionally was locally based, and did not involve paid labour. The raw material for the construc­ tion was locally available. It should be further studied to what extent the size of the household and its capability of raising additional labour from the neighbouring (most often related) community influenced the size of the house. The latter aspect might also be related to the space that was allotted to each household for the construction. This could have a connection to the area of paddy land owned by the household.2 It was not possible to relate the difference in house size to caste affiliation in this small modern sample. Nagalavava is a low caste (Nakatikula) village, whereas Talkote is a Govigama village. In our example it seems that relative wealth is what determines the size and degree of differentiation of the dwell­ ing.3 No clear picture could be obtained regarding the tradition of men and women having different sleeping quarters. From descriptions of how the rooms were used, it seems that this custom was not practised in eight of the eleven households. At present Nagalavava gammandiya contains seven hou­ seholds, all related by birth or marriage and all belonging to either of two kinship groups (Millagahagedara or Komgah- agedera). Talkote, as discussed in the presentation of the study area, is mainly a paddy cultivating village; whereas Nagalavava is mainly chena cultivating. PLAN OF THE VILLAGES Talkote The present layout of Talkote village shows individual hou­ ses with surrounding home gardens dispersed mainly on the two sides of a gravel road (see fig. 18:2). The main tanks and paddy fields of the village are accessible through this road, which connects to the main road from Sigiriya to Kimbissa and Inamaluva and further, to Dambulla and Habarana. The graveyard is situated 120m to the north of the centre of the village. The former gammandiya area now is mainly a wasteland. It is situated between the two main tanks of the village. In this sense it wasn’t built in what is thought to be the traditional position: beside a tank bund. An abandoned settlement site named Pahala Talkote however was pointed out by the vil­ lagers as the former site of the village. It is situated west of the bund of the Pahala Talkotevava, which is the largest tank in the village. If this is the initial location, the former gamman­ diya (as the present position of the village) might signify a situation when additional tanks were incorporated in the vil­ lage economy. Today six tanks belong to the village, though two of them are abandoned and a third very seldom used. The breaking up of the cluster village started ca 50 years ago. We found the location of various functional features, such as storage facilities and kitchens, to vary among the houses. As the functional pattern of the common courtyard was dis­ rupted in Talkote when the houses were moved from the gammandiya, and hence also the orientation of the house, the traditional pattern might have been more uniform. No such pattern was seen though in the present day Nagalavava gam­ mandiya. The various functional requirements - that is, private space, public space, storage and food preparation space - sho­ uld have been the same however, previously too. Disposal of refuse is an interesting subject from an ar­ chaeological point of view. In Talkote only a general view of garbage handling was obtained. Refuse here is most often swept into a heap in some corner of the individual courtyard. Dogs, cats and hens (and wild boar) will eat what suits them. Now and then the heap is burnt and the ash left to blow away Figure 18:3 Plan of Nagalavava village. Nagalavava Present day Nagalavava is divided into two. A severe flood­ ing in 1977 caused most of the villagers to move to the northern end of the tank (Nagalavava). The cluster village, containing five houses, some of them partly newly built, remains at the southern end of the tank. It is this part of the village that was mapped. The houses are constructed within an area ca 39x27m west-east (see fig. 18:3). The tank bund constitutes the road between the two halves of the village. They are both connected by gravel roads to the Digampataha-Kimbissa road and hence further to Dambulla and Habarana. Footpaths also connect the village to Talkote and Pidurangala villages, to the east. Three tanks belong to the village today. Until about 25 years ago, there were no toilets in the village (important for human coprolite analyses). The jungle and scrub land outside the tisbambe (cleared area around the dwellings, about 50m wide) were used for that purpose. Pit lavatories are used today by the two households with the largest and most functionally differentiated houses. The graveyard is situated about 360m north-east of the village, in present day scrub jungle. The cemetery is common to all members of Nagalavava village, there are no separate family burial grounds. According to the villagers, it is common to have one graveyard per village. Bordering the east of the present cemetery is the old abandoned one. Our informants said their ancestors were bu­ ried there. No investigation was made as to how ancient this practice of burial is. The villagers fetch water from a common tube-well, but pne household (living in the next biggest and second most differentiated house) had dug a private well in their onion plot. In Nagalavava, the question of how refuse was disposed was put to each household. The practice was found to vary among households. An old widow just threw the garbage alongside her house, without sweeping, burning or burying (see fig. 18:4). Another household dumped the refuse in a pit and buried it once the pit was filled, after about six to seven days. Behind one house a pit had been excavated to obtain material for an extension to the house (see fig. 18:7). This pit was being filled with garbage. When full, the household wo­ uld cover the pit with soil. The household with the most differentiated house said their refuse was dumped in a pit close to the tank bund (see fig. 18:6). Another household dumped the garbage at the rear of the house (that is, not onto the former common courtyard) and burned it when considered necessary (see fig. 18:8). The Nagalavava example points to the fact that it might not be enough to identify refuse in the form of burned ma­ terial, for osteological and macro-fossil analyses in a rural settlement. Undifferentiated pits to which no clear function could be ascribed might be refuse pits, and soil samples could be taken to look for less destructible remains such as pollen and phytoliths. PLAN OF THE HOUSES Nagalavava The main features of the dwelling houses in Nagalavava are squarish rooms and at front, a roofed platform - a verandah - and pitched roofs. Considering dwelling space the atupalla must also be mentioned. This is the space under the atuva (storage facility for dry grains, see below). The platform of the atuva is 0.6-lm above ground. The space beneath, ac­ cording to our informants, was used at times as a dwelling by poor villagers who had no means to build for themselves. Usually only one person at a time occupied this space, al­ though one family of six was mentioned as having lived under an atuva in the village. The last person to have lived in a Nagalavava atupalla was a widow. The owner’s permission was necessary to utilize the atupalla as dwelling space. No constructions for storage of produce outside the houses are present today in the village. The remains of a cart stall however are seen north of the largest house. It once belonged to the house (see fig. 18:3). It is to be noted that it was sited outside (north of) the common courtyard. With extensions to the houses, due to increased material possibilities and/or a changing family life cycle, the verandahs were seen to be built into the house. Today there are two open verandahs left. One of them and all but one of the former verandahs roughly face south. The other remaining open verandah and one now enclosed are built opposite to these. One abandoned and partly demolished house is situated ca 40m north-west of the present village. Its entrance is oriented to the south-east, towards the other houses (see fig. 18:3). This indicates a former, common, central courtyard in the village. Five out of seven households had a separate kitchen. The two without, consisted of a widow living alone in one case, and a newly married couple who had built a separate dwelling, an extension of the husband’s parental home. A changed family situation had led to the largest and most differentiated house being abandoned a few months before the interviews took place. It was built in the colonial style, in 1955, bycarpenters, using cement for the flatform and tiles for the roof. But most houses are still built by the owners themselves, using material obtained locally. The present owner has moved to his parents in the main village and is planning to demolish the house. The present village layout is in a state of change. This is not a recent phenomenon, as evident from the house con­ structed by carpenters in 1955. Yet Nagalavava still resembles the descriptions of traditional cluster villages, in that the houses don’t have separate courtyards. Changing ground plan of houses To give examples of the ‘life cycle’ of the houses, they will now be described one by one. Note that all square measure­ ments are approximations, as none of the houses, apart from the carpenter-constructed one, is built with perfect right angles. The inner square measurements given, include the open verandahs. The widow’s house had contained three rooms and one kitchen, but her son had demolished three rooms (with her consent) to obtain building material for his own house. The widow, who had married a widower, said the kitchen had been built only after her marriage to him (see fig. 18:4). The single bedroom cum kitchen of the widow measures 3.2x2.75m. The open verandah measures 4.8xl.8m. The door opening is 0.75m wide and situated in the middle of the wall. The hearth, built of three bricks, is in the south-eastern comer, opposite the entrance. Previously the kurakkan gala (hand mill) was placed in the room, which has been her bedroom throughout, but after her husband’s death she doesn’t use it as often as before, and has put it in the verandah. There the miris gala (chilli-grinding stone) is also found. Rice is stored in an earthen pot kept in the room. The house previously had a continuation to the east. The decaying platform is still faintly visible. Today the room and the open verandah have an inner square measurement of ca 17.5m2. The house opposite has four rooms, a built-in verandah and a kitchen (see fig. 18:5). According to the owners, it was built around 1957. Initially the house consisted of two rooms in a row, with separate entrances from an open verandah facing south. The first addition to the house was the kitchen, built with a separate entrance from the courtyard, as an exten­ sion to the west. It measures 2.7x2.75m. The miris gala has been placed on a platform (former pila) inside the kitchen, which also has an entrance, in the form of a passage from the verandah. Later, the verandah was enclosed by building two rooms to the south. The main entrance to the house is now from the end of the former verandah in the east. It opens on to a sitting room without an enclosing wall, towards the former verandah. The other is an additional bedroom which has a doorway to the east, towards this room. The first room to the right of the 0.85m wide entrance measures 3.15x3.65m. It is used as bedroom for the housewife and her two youngest unmarried daughters (26 and 21 years old respectively). The kurakkan gala is kept in one corner of this room. The original entrance, 0.65m wide, has been retained. The other room from the original construction is now used as a storeroom for sacks of paddy, kurakkan and the like. It measures 3.45x2.95m. Its entrance, measuring 0.65m is also left in its original position. Both rooms face the built-in veran­ dah which is now lm wide. The room to the left of the main entrance is used as a sitting room, measuring 3.4x3.5m. It has thus taken over the ‘public space’ feature of the former verandah. This usage is consistent with the fact that it is not enclosed towards the former verandah. From this room a 0.65m wide doorway leads to the second additional room, to the west. It measures 3.2x3.3m. It is now used as the father’s bedroom. The katta gana gala, for sharpening iron tools, is left outside the house. No other household uses this kind of shar­ pening device any more. One interesting aspect of this changing house plan is that the principles of original usage have been kept. That is, private space, where the wife and husband don’t share a bedroom, and public space which has an open character. The extended build­ ing creates additional space for storage and food preparation. At present the inner square measurement is approximately 59m2. East of it is the house built by carpenters (see fig. 18:6). Here the common plan of aligned rooms with separate entran­ ces to an open verandah is not seen. The present mainentrance, 0.85m wide, faces the abandoned cart stall. It leads into an L-shaped sitting room measuring 4m from the entrance and 4.4m and 3m across. This room was used as a bedroom by the present owner and his brothers when they were children, alternatively with the bedroom left of the main entrance. That bedroom’s sole 0.75m wide entrance is from the sitting room. The bedroom measures 2.65x2.45m. Opposite the main entrance the sitting room has a 0.85m wide doorway towards an L-shaped, enclosed verandah. This was previously the bedroom of the owner’s sisters. It me­ asures 2.45m from the entrance and narrows to 1.7m west of the door. It measures 6.5m east-west. North of the sitting room there is another bedroom, used by the owner’s parents in the past. It measures 2.55x3.2m and its sole 0.6m wide entrance is from the enclosed verandah. In front of the verandah two later additions were made. One was used as a storeroom for paddy and such produce. It measures 2.9x5m and has two doorways. One measures 0.8- 5m towards the former verandah and the other 0.9m, towards the presumed former common courtyard. The other, later addition was a kitchen of wattle and daub, measuring 3.1x3.2m. It is now demolished. Its sole entrance, measuring 0.7m, was from the enclosed verandah. Where the cooking was done before this kitchen was built could not be found out in the interview. What is common between this house plan and the others is that the two bedrooms don’t have a direct connection, but are entered from structures with a public space function. The inner square measure of this house is approximately 67m". North-east of this house is a dwelling inhabited by two individual households (see fig. 18:7). On the first visit to the village in 1990, a family of five inhabited a house with one room, one kitchen and a verandah opening to the south. The room and kitchen had separate entrances from the verandah only. Rice and minor agricultural implements were stored in the only room. Figure 18:7 House plan. Nagalavava village. The family consisted of man and wife with their grown-up daughter and her daughter, and their grown-up son. In 1992 the son had married and an extension was made to (he house, towards the west. The son and his wife now live in one room, which is also used as a kitchen and for storage of rice and other agricultural produce. The verandah has also been ex­ tended, to the west, alongside the new room. The original room, with a 0.65m wide entrance from the verandah, measures 3.2x3.6m. The kitchen, to the east, has an 0.55m wide opening to the verandah. It measures 3.15xl.6m. The hearth consists of three bricks placed in the south-western comer of the room, just to the right of the entrance. The verandah measures 1.7x4.25m. It is built roughly up to the entrance of the kitchen. The newly built room has an opening 0.6m wide towards the newly built extension of the verandah. The room measures 3.1x2.6m. The hearth is built up by clay with place for two pots (two fires). It is situated in the south-eastern corner of the room, just to the left of the entrance. The extended verandah outside measures 1.75x3.25m. The households have neither kurakkan gala nor miris gala, nor vamgediya. Both hou­ seholds use the vamgediya of their neighbours, to the south (see fig. 18:9). North of the newly built room, at the rear of the house, there is now a pit. It was dug to obtain building material for the foundation and the walls. As mentioned earlier, it is now used as a garbage pit. The original house has a ground plan measurement of ca 23.8m2. The dwelling of the young couple has a square mea­ sure of 13.75m2. South of this house is yet another example of a dwelling that has been extended to house a newly married couple, close relatives of the original household (see fig. 18:8). The original house was built by the widow’s uncle - her mother’s eldest brother. Thus it was built in the present inh­ abitant’s grandparents’ generation. It consisted of one room with an open verandah facing north. A kitchen has been added to the house at its eastern end. The room measures 2.9x3.25m. It has a 0.55m wide ent­ rance to the former verandah. The enclosed verandah me­ asures 1.15x3.25m. The kitchen measures 2.9x1.75m. The verandah wasn’t extended to cover the kitchen; the kitchen is entered through a 0.5m wide doorway from a small enlarge­ ment of the verandah. The hearth is built up with clay, with provision for two pots (two fires). It is in the north-eastern corner of the room, just left of the entrance. The miris gala is placed outside the kitchen, not to mess the floor while cleaning the grinding utensils, according to our housewife. Together with the vamgediya, kept on the partly enclosed verandah, it is shared with the other household and also those living in the house to the north. The man living in the newly built house is the younger brother of the occupant of the original house and the woman the younger sister of the housewife. They have two small children. The main, 0.7m wide entrance to their house is from the north. No verandah has been built here. A 0.25m wide passage at the opposite corner of the room leads to the now enclosed former verandah. The room measures 2.85x3.35m. The kitchen is built east of the bedroom and has a single 0.6m wide entrance from the room. The kitchen measures 2.85x2.9m. The hearth is built of clay, for two pots (two fires) in the north-western comer of the room, just left of the en­ trance. The kurakkan gala is in the kitchen. The inner square measure of the original house is ca 1- 8.3m2. The extension measures 17.8m2. Orientation of houses and extensions of buildings In five cases a kitchen had been added to a house, but only in two cases as an extension in front of the original house, towards the presumed common courtyard. There was, lik­ ewise, no clear pattern as to the kitchen’s position in relation to the house. In three cases a kitchen had been added to the east of the house, in one case it was built as an extension to the west. Regarding the widow’s house, the now demolished kitchen had been situated to the east of the house. Room extensions had been made towards the common courtyard in three cases, and in one case towards the north, in line with the old house. Various factors probably influence where additional rooms are built. The presence of neighbouring houses is, of course, an important factor. No house however, had been extended at the rear, that is the side opposite the courtyard. Future studies should focus also on how land for dwelling houses is allotted in each case. With the extension of buildings, entrances were situated in various positions, in relation to the courtyard. All five houses had at least one entrance towards the courtyard. The two largest houses, upon extension, had moved the main entrance away from the courtyard. The same was done by the ho­ useholders who let their siblings build an extension towards the courtyard. One reflection that may be useful to future studies is that when the largest house has its main entrance on the opposite side from the courtyard, it indicates communication with the outer world (both as regards verbal communication and move­ ment and control of material things) more in the private inter­ est of this individual household than it is allowed to be the concern of the whole village. Other features that were used by individual households, like the well, the two toilets and the abandoned cart stall, were all outside the courtyard. Talkote In Talkote, the three mapped houses each had a separate kitchen. The same variation was seen regarding material and technique of construction as in the Nagalavava houses; and in two cases the houses had been extended (see fig 18:10 and fig. 18:11). The verandah, for example, had been walled and made into a room, in two cases. The entire roof in the largest house and partly in another was now made of corrugated sheet metal. The floor in part of the largest house was also lined with cement, instead of clay and cow dung. The most traditional of the three houses was built ca 50 years back (see fig. 18:10). It has two rooms and one kitchen, built in a row, with openings to the verandah only, facing south. The inner measurements of the rooms are 3x3. Im and 3x3m respectively. The kitchen measures 3x2.5m. The veran­ dah measures 9.25x1.25m. Half-walls are built around it. The inner square measure of the house, including the verandah, is ca 37.4m2. The fireplace in the kitchen is made of five bricks, placed to support cooking vessels. It is situated in the left corner, opposite the door, to the north-west. Along the wall to the left, between the door and the fireplace, there is a wooden shelf for storing earthen pots and kitchen utensils. To the right of the door are a miris gala and two vamgediyas. Along the opposite wall, beside the fireplace there are four pots for storing water. Above the fireplace, is the dummassa (wooden frame on wh­ ich, for example, fish is dried). There is also a fireplace 7m west of the south-western corner of the house, used when large quantities of rice are cooked. One of the rooms was used as a bedroom and the other for storage of food items and clothes. The household consisted of husband and wife with five children, aged between eleven and one. The verandah was used to entertain visitors. The house in its largest part is built of local raw material, though part of the verandah is now covered by a corrugated sheet-metal roof, and part of the verandah platform has been aligned with bricks. Another house consists of four rooms, with a smaller kit­ chen (see fig. 18:10). The verandah has been walled in on two sides. The fourth side, where the entrance is situated, has only a half-wall. The rooms are placed in a row, apart from the enclosed verandah which is now in a right angle to the rest of the house. The house has two entrances: one to the kitchen and the other to the former verandah, which is used as a visitors’ room and bedroom for males. The kitchen measures 3.35x1.25m and is a later addition, with wooden instead of wattle and daub walls. There is a fireplace in the north-western corner and two pots for storing water, whereas the miris gala is placed in the courtyard, just outside the kitchen, where there is also a wooden shelf for pottery. The two rooms following the kitchen measure 3.5x2.7m and 3.75xl.35m respectively. They are used for the storage of food and herbal medicines which are kept on wooden shelves, or hung from the ceiling. Pottery is also stored on a wooden shelf in the room next to the kitchen. In 1992 the largest of the rooms was also used as a bedroom. The following room is a bedroom for women and children. The room measures 3.7x2.95m. In the floor we found a vam- gediya stone, the kind on which the vamgediya is placed for pounding. The stone was said to be used even at present. Sometimes items were ground or pounded directly on this stone. Its diameter was 0.3m. The visitors’ room and bedroom for males is the enclosed verandah, in right angles to this room. It measures 2.25x3.25m. The inner square measure of the house is ca 37m2. In 1990 the household consisted of husband and wife and four children, ages 13 to 7. The most complex house plan is the house with the cor­ rugated sheet metal roof and cement floors (see fig. 18:11). It was built in 1961. It consists of six rooms and one kitchen. This results from various additions to the original building. The house is rectangular, measuring 10x8m on the outside. In 1989 it had a small (2.5x2.5m) extension in the north-eastern corner, used as the kitchen by the householder’s father and mother. The house has two entrances, one to the main kitchen, the other to the sitting room. The kitchen extension has an entrance from the courtyard, but no direct connection to the house. The main entrance leads to the sitting-room which me­ asures 3.6x4.5m. The sitting room is also used for storing agricultural produce (like onions). Right of the entrance is a room which was occupied by the householder’s parents. It measures 3.6x2.8m and its only entrance is from the sitting­ room. It was used as a bedroom and for keeping their belong­ ings. These two rooms have enclosed the former verandah. After the sitting room comes the former verandah, measur­ ing 7.55x1.35m. It is now lined by cement and likewise used to entertain guests. Two bedrooms follow, measuring 3.2x3.65m and 3.1x3.6m respectively. Their entrances are from the former verandah, but they have no direct connection to each other. One of the bedrooms has an entrance to the kitchen. Food items are also stored in this room, hanging from the wall. In the kitchen, measuring 1.85x3.75m there is a fireplace in the comer, left of the bedroom entrance. To the right is the entrance to the courtyard. Above the fireplace hangs a bag with dried fish. The grinding and pounding facilities are in the courtyard outside. The katta gana gala is also in the courtyard. Next to the kitchen is a room measuring 1.85x3.65m. It is used for storing pottery and other kitchen utensils. These ro­ oms are also extensions to the original house plan. In 1989 the household consisted of husband and wife and five children from the ages of 19 to 7. The householder’s aged parents had a separate kitchen and bedroom and could thus, in a spatial sense, be considered a separate household. As regards food supply however, they were dependent on their son. They could not, therefore, be considered an independent unit of production. This points to one of the difficulties that could be encountered in an archaeological context, in trying to define the manpower of a household. In 1992 one of the elders had moved to the house of another of her children, and the se­ parate kitchen had been demolished. The householder and his family occupied an area of ca 63m2, whereas the aged parents occupied over ca 16.3m2 in 1989. In 1992 the remaining household occupied a house with a square measure of 73m2. CONSTRUCTION OF HOUSES - THE NAGALAVAVA AND TALKOTE EXAMPLE Before building a house, the Nagalavava villagers say, it is customary to have the site examined by an astrologer. The planetary positions and the horoscope of the householder are consulted. Thereafter offerings are made to the guardian de­ ities of the earth, and the outline of the house is marked at an auspicious time. How a person obtained the right to build a house in the village and how the land was allotted, was not ascertained in this study. The basic, structural elements of the house are the founda­ tion platform, walls and roof. The platform is built of clayey soil, to a height of about 0.2m. Ashley De Vos mentions that this kind of platform both hinders damp from entering the house and insulates the building from the heat accumulating in the surrounding ground (De Vos 1977: 46). To build the roof, the four corner posts and taller gable post are erected first. They have a diameter of about 0.2m. They are sunk into the ground to a depth of about 0.6m. A mixture of sand and gravel is poured down into the post hole after erection of the post and pressed around it. Along the line of the wall, between the roof-support posts, thinner wooden posts are placed vertically (see fig. 18:12). They have a cir­ cumference of 0.1-0.15m, and are placed in post holes 0.4- 0.5m deep. Sticks are tied horizontally to the posts to support the wattle and daub walls. The walls do not join the roof.Construction of the walls commences only after the roof is thahatcched. . Figure 18:11 House plan. Talkote village. The roofs are pitched. The gable and comer posts holding the roof have crossbeams connecting them. They rest on the roof-support posts in a slot cut into their upper end. No nails were used previously (see fig. 18:13). The rafters, placed in right angle to the crossbeams, were joined to them by kirival. In right angle to the rafters, wooden sticks are placed to fa­ cilitate the thatching of the roof with iluk straw, or cadjan (woven coconut palms). This is fastened to the roof by coir rope, or tied with the palm leaf itself. Cadjan is most often bought from itinerant merchants arriving before the start of the monsoon. According to our informants, the cadjan roof will need replacement at least every second year. If it is attacked by termites it has to be replaced annually. In Nagalavava we found the carpenter-built house to have a tiled roof. In Talkote, the largest and most differentiated house had its roof entirely covered by corrugated sheet-metal, whereas another house had part of its verandah covered by corrugated sheet-metal. This kind of roof cover necessitates less maintenance work, and that was the reason given for such material being preferred. The walls are built with clayey soil. Fine clayey soil is pounded, the stones and gravel removed. Water is mixed with the soil and the mixture is trampled underfoot (see fig. 18:15). It is left for 24 hours, then made into balls and placed in the skeletal structure of the wall. Three days later a second pl­ astering is done. To fill in cracks resulting from the drying of the walls, a plaster of clay mixed with fine sand and water is daubed on the wall several times. This final coating is not applied by poorer villagers. No trowels or other implements are used to apply the clay to the wattle and daub walls. To give a neat appearance, the outer wall can also be daubed finally with clayey soil mixed with cowdung, up to about 0.6 m from the outer platform (yatapild). The largest houses in both Talkote and Nagalavava were white-washed. The walls of the dwelling were said to need repair once a year. White ants and the monsoon rains are the main cause. In Talkote, this was done before the New Year festival. We have also seen that the walls of wattle and daub houses left unin­ habited and unattended, fall into heaps of mud within a few years. The floor is made by flattening out the clayey soil (on the platform). Thereafter the soil is mixed with cow dung and applied to the floor. The practice varied as to how often this coating had to be replaced. One housewife in Nagalavava reapplied it once a week. In Talkote, one housewife said she did it once a month. Even the largest house documented in Talkote had cowdung-coated floors in the latest built rooms. Around the outer walls there is often a narrow platform­ like construction, the vatapila, about 0.3m high. This is built to prevent the supporting wall posts washing away. The clay that falls to the ground while plastering was used in Nagalavava village to construct this platform. In Talkote we found parts of outer platforms built of bricks, or sharp-angled stones. In the process of constructing the wall, space is left for doors and windows. The doors are narrow. In Nagalavava, the average breadth varied between 0.65 and 0.9m. In Talkote the breadth varied between 0.7-0.85m. A wooden door-frame is placed in the opening, fitted firmly by creepers. The door has kudutnbi (tenons) at one comer, above and below. These are fitted into corresponding tauva (sockets) in the door frames. Halamba, teak and palu wood are used for doors. In Talkote we found a discarded yota (water-lifting device) used as a door, without tenons and sockets. In the average house, the doors are crudely finished. In Nagalavava, a simple flower motif about 0.1m wide is sometimes carved on the upper crossbeam of the door frame. This was not found in Talkote. Coomaraswamy noted of the beginning of the century that this beam was "...generally decorated." (Coomaraswamy 1956 (1- 908): 133). In Nagalavava we were told that "in early times" door openings did not have attached doors. When the owner went out they would cover the entrance with a winnowing fan (kulla). No outsider would move the winnowing fan and enter the house. Later this was replaced by door leaves, but no latch (agula) was used. The closed door was tied with a rope fas­ tened to a post in the verandah (pilaf The purpose of closing the entrance was to prevent serpents and other wild animals entering the house from the jungle. The windows, where they exist, are very small. The neces­ sary ventilation is provided by the gap between the roof and the walls. RURAL HOUSES FROM AN ARCHAEOLOGICAL VIEWPOINT Knox says of the Kandyan region in the late 17th century: "Their Houses are small, low, thatched Cottages, built with sticks, daubed with clay, the walls made very smooth...They employ no Carpenters, or house builders, unless some few noble-men, but each one buildeth his own dwelling. In build­ ing whereof there is not so much as a nail used; but instead of them everything which might be nailed, is tyed with rattan and other strings, which grow in the woods in abundance; whence the builder hath his Timber for cutting... The poorest sort have not above one room in their houses, few above two, unless they be great men." (Knox 1981:235). The houses we documented, with two exceptions, are still built in this traditional way. We have found no rural settlement site remains that indicate a previous rural architecture using less destructible building material. We also see that the more prosperous villagers invest in larger and more differentiated buildings and modern building material. Figure 18:12 Close-up of wall construction. The thin vertical sticks are bound to the wooden post with wickers. Talkote village. Photo: Eva Myrdal-Runebjer. Figure 18:13 Joint of roof-carrying corner post, Talkote village. No nails are used in traditional rural house construction in the study area today. Photo: Eva Myrdal-Runebjer. Figure 18:14 Preperation of clayey soil used to repair dilapidated wall in the background. Diyakapilla village. Photo: Eva Myrdal-Runebjer. Figure 18:15 Applying cowdung to the kitchen floor, for neat appearance and smooth surface. Nagalavava cluster village. Photo: Eva Myrdal-Runebjer. Only two families in Nagalavava had more than one room at their disposal, apart from the kitchen and verandah (see fig. 18:5 and fig. 18:6), and only five out of seven households had a separate kitchen. In Talkote we found separate kitchens in all three houses, and all three families had more than one room at their disposal. In all cases where the building history of the house could be established, it seemed the kitchen constituted a later addi­ tion. Whether this is a development of the latter half of the 20th century only, or if it was a practice previously too, to extend the house with separate kitchens or additional storerooms, we don’t know. According to Knox, it was not common in the 17th century Kandyan region. This should be further studied, together with the question: is there a difference between now and then, as to how land is allotted for building houses? In Talkote, the inner square measure of the free-standing documented houses ranged from 37m2 (6.16m2/individual) to 73m2 (10.43m2/individual). The aged parents had occupied an area of 16.3m2 (8.15m2/individual), though of course they had access to the rest of the house also. Unfortunately we don’t have information about the total number of inhabitants for all the houses in Nagalavava. The largest house in this village had a square measure of 67m2, but we don’t know how many lived there before the son, the present owner, took it over. The next largest house measured 59m2 (11.8m2/individual). The smallest free-standing house with one occupant measured 17.5m2. The young couple oc- cupied an area of 13.75m (6.87m /individual). Previously the whole family of five had occupied an area of 23.8m2 (4.76m2/individual). The young couple who had built an ex­ tension to their sibling’s house lived with their two children on 17.8m2 (4.45m2). Of course, no conclusions can be drawn from this small sample. One trend however, was that the larger houses (73m - 59m“ in our sample) were also those that gave a larger area of individual-living in the house - the widow’s house being an exception. It should be examined whether the size of the house could be a useful indicator of household wealth. Probably the number of fireplaces is the best indicator of the number of households. Most problematic still is finding the actual site of the house at all, in an area where surface pottery indicates a settlement. Unlike many regions of South America, where several archaeological rural sites have been studied with the focus on household wealth, Dry Zone Sri Lanka doesn’t have a rural architectural tradition of substantial stone platforms and foundations which make outlay and size easily visible.4 The earthen floor of compact soil and the aligning post holes would be the indicators of house size, and this is not visible (if at all) until the surface soil has been removed (see fig. 18:16 and group piece). This is true of course also of countries in temperate climates like Sweden, but soil properties, soil de­ position through monsoon rains, and lack of machinery, make large scale openings of surfaces in Sri Lanka’s Dry Zone a time-consuming task. Given the difficulty of identifying the individual house, it is likewise difficult to follow the other approach in studying household wealth - that based on household possessions.' However, relative difference of wealth among different settle­ ment sites might be possible to distinguish by following the latter line of research. To identify and document houses in rural archaeology re­ mains a most important task. The solution is probably a gr­eater (relative to the large elite structures excavated) investment of time and manpower in rural excavation. investment of time and manpower in rural excavation. Activity areas, tools and auxiliary devices EVA MYRDAL-RUNEBJER Eight families were questioned about material equipment and practices relating to household maintenance: three families of Talkote, one family from Diyakapilla, one from Pidurangala and one family from Kosgaha-ala. A family in Kayamvala was also questioned about mi oil processing, and a household in Nagalavava about the use of the kurakkan gala (hand mill) and storage of ground kurakkan. In Nagalavava, a total inventory was begun of household equipment and locations. This interesting study is still not complete and will be reported separately by the researcher Sudarshani Fernando. In the following we will try to give each item a place and date of origin, describe its specific physical form and function and compare it with the descriptions given by previous first hand sources. We must not think of the present ‘traditional’ set-up as something timeless and static, even if we find paral­ lels from the early modem era. PROCESSING OF FOODSTUFF The processing of foodstuff includes various labour proces­ ses such as pounding, husking, grinding, scraping, mincing, meat-cutting, boiling, steaming, baking, roasting, frying and preserving (through a variety of methods like smoking, drying, freezing, salting, preserving in honey, fermenting, churning, cheese-making etc.) depending on the cultural set­ ting under study. For archaeologists, food processing techniques are inter­ esting also because it is such labour that produces most of the orgsinic refuse from the settlement site - as well as our most chenshed datable settlement feature: the hearth! Ethno-botani- cal studies help us to understand in what way given species could be used in the household, and to know what to look for and where Fireplace The fireplaces seen in Talkote and Nagalavava were con­ structed either of bricks, usually three in a rough triangle, placed on their long ends, to stand the earthen vessel for cooking, or a clay foundation built to form a 3/4 circle, with the opening facing the room, the fire burns below, on the earthen floor (see fig. 18:17). It is to be noted that the burning wood is usually a single branch, burning at one end and pushed further into the hearth as it burns. Such a technique leaves less charcoal and soot than if the fire is also used to warm the dwelling, when more wood has to be used. In one household in Talkote, we found an additional fir­ eplace built in the same manner, 7m west of the south- we­ stern corner of the house. The housewife explained that she used this when cooking large amounts of rice in the big rice cooking vessel. Husking and winnowing place In Talkote, a palm roofed structure - the maduva - was built as an extension of the dry grain storehouse - the atuva. The roof construction was common with the atuva, the gable a height of 3.1m. Wattle and daub walls 1.3m high surrounded the earthern floor on the outer sides, except the entrance. The floor, measuring about 2.75x2.6m, was plastered with a mix­ ture of mud and cowdung. This structure was said to be a Traditonal additon to the "Atuwa" and used as shelter when husking wih the "vamgediya" and winnowing the husked seeds. in this household, the walls were also used for hanging fishing nets, and variety of other items were also hung or leant againsts the walls, in line with the floor level we siscoverd a very instersing feature: a flat stone with a diameter of 25 cm (see fig. 18.18} No name could be given to it, but such stones were according to the housewife, meant to place the vamgediya on while pounding, so as not to damage the floor. We found a similar stone, its diameter 30cm, in another dwelling house, in the floor of the bedroom of the housewife and children (see ab­ ove). The housewife said the room was previously used as a maduva. It couldn’t be established if this meant the same floor was used for an extended room, measuring 3.7x2.95m, and once functioned primarily as a maduva. The room was at the opposite end from the kitchen, but the stone was still some­ times used. It had a slightly concave surface and the ho­ usewife told us that pounding was sometimes done directly on the stone, without the vamgediya. In the former gammandiya of Ilukvava village (it was abandoned in 1942) we found a broken stone with a flat upper surface that had been used as a ‘vamgediya stone’ according to the villagers. It was not found in situ however, but thrown out into a former courtyard. The area is used for swidden cultiva­ tion nowadays, much favoured for the fertility of its soil (see fig. 18:19). Robert Knox writes, based on his experiences in Sri Lanka between 1660 and 1679, that: "They lay the Rice on the ground, and then beat it, one blow with one hand, and then tossing the Pestle into the other, to strike with that. And at the same time they keep stroke with their feet (as if they were dancing) to keep up the Corn to­ gether in one heap. This being done, they beat it again in a wooden mortar to whiten it...” (Knox 1981:240). The stones used in these operations are much larger how­ ever than those referred to above7. So far, I’ve found no explicit reference to these smaller stones in the literature. If their function could be ascertained, we could take it as a food-processing indication when discovered in an archaeolo­ gical context. Pounding facilities In Talkote, both paddy and kurakkan and other dry grains are pounded in the vamgediya - a large mortar and pestle-like device - partly for the sake of husking the grain (see fig. 18:20). Knox writes when listing household furniture: "Also some Ebeny pestles about four foot long to beat rice out of husk, and a wooden Morter to beat it in afterwards to make it white,..." (Knox 1981:235). The vamgediyas we saw were all made of wood, but John Davy writes that "the mortar is occasionally made of stone, but more generally of wood". He refers to it as a "rice-mortar or pestle, for pounding paddy, and depriving it of its husks." (Davy 1983 (1821:208). Figure 18:17 The hearth is the archaeologist’s most cherished datable remains from household activities. The present method in the study area however leaves a minimum of charcoal. The logs are slowly burned at their ends and subsequently pushed into the hearth. Talkote village. Photo: Eva Myrdal-Runebjer We found it to be of importance also in the dry-grain economy, well integrated into the harvest and storage techni­ ques of ear-by-ear harvested dry grains. Davy might refer to another geographical setting, or might have overlooked this other use of the vamgediya, as we have no indications that the present practice derives from a time after 1819. Robert Knox mentions that, as regards kurakkan, they "gr­ ind to meal or beat in a Mortar" and that "Tana" (Tanahal - Italian Millet) "is beaten in a mortar to unhusk it." (Knox 1981:112-113). The vamgediya we documented was made by the present senior householder, who had worked also as a carpenter. We don’t know the age of this particular piece. The informants said such a device could be used for a very long time - "for 200 years", which if we don’t take it literally, at least indicates a possible usage of over two or three generations. The mortar was made of alangu wood. It was 47cm high in all, with a base diameter of 28cm. The diameter of the inner orifice was 25cm, the walls were about 3cm thick and the carved depth of the mortar 15.5cm. The pestle was made of kaluvara wood. The pounding end was encircled by an iron ring. The pestle was, in all, 110cm long. The pounding end had a diameter of 5.5cm, whereas the upper part was 4cm in diameter. It weighed 2kg. In another household we saw a similar device, worked by a seven-year-old girl who husked the paddy while the gra­ ndmother winnowed the husked grain. We have to consider children’s work within the household in discussing, for ex­ ample, household size and household wealth. This kind of partly needed, partly playful work carried out by children, with growing responsibility as they grow older, is evident the world over (even in the peripheral rural districts of a highly industrialized country like Sweden) and is, of course, difficult to measure in the sense of labour input and quality. What is important though, is that it represents reproductive work of life-supporting necessity, carried out under the watchful eye and instructions representant of a grown-up. As far as the individual household is concerned, the next generation’s sur­ vival depends on this, not at all glamorous, task of supervising and learning by doing - the only archaeological trace of which we see as chronologies. Grinding facilities Various devices have traditionally been used by the indi­ vidual household for grinding purposes, both rotary hand mills and grinding stones. In the villages we found grinding stones for chilli and medicine, but not for grains (see below) Figure 18:18 Rice is husked in the wooden mortar (vamgediya). Often it is husked and winnowed inside the shed (maduva) adjoining the dry grain storehouse (atuva). Level with the floor of the maduva is a flat stone on which the vamgediya is placed. Example of an archaeologically visible food-processing feature. Talkote village. Photo: Eva Myrdal-Runebjer. Figure 18:19 In the former courtyard (present-day chena field) of the abandoned gammandiya ofllukvava village a broken "vamgediya-stone" was found. Broken and not left in situ it would have been difficult to identify in an archaeological context. Photo: Eva Myrdal-Runebjer. Henry Parker, from his and civil servant Hugh Nevill’s observations in the field, in the latter half of the 19th century, mentions that the Vaddas (probably in the Vavuniya region) ground millet into flour on "a flat stone, or in a quern, by those who possess one...” (Parker 1984 (1909):52). As regards gr­ inding seed for flour, we might consider the difference in capacity and required labour investment between the grinding stone and the hand mill. We could also look into the labour invested in the making of the grinding stone as compared to the hand mill. The dating of the first appearance of rotary hand mills in Sri Lanka is still obscure. Figure 18:20 The grandchild pounding rice in the wooden mortar (vamgediya) placed in the coutryard, Talkote village. To teach the younger generation practical tasks is an important part of reproductive work in any society. Photo: Eva Myrdal-Runebjer. In Talkote village, we found no kurakkan quern in use. The lower part of a kurakkan quern was found in the site of the former gammandiya, one such part was used as a step below a fence; and a piece of stone found outside one of the houses formed an almost finished upper part of a kurakkan quern. The cutting had been started by the householder about 10 years ago, according to the housewife, but was now abandoned. Kurakkan is still grown by several of the visited households, which might indicates a shift in kurakkan preparing by the household, from grinding to pounding. None of the visited households sent their kurakkan out to be milled, as far as we could find out. In Nagalavava gammandiya, several kurakkan querns we­ re found still in use (see fig. 18:21). The one we documented was typical, consisting of two nearly cylindrical pieces of stone placed one above the other. The lower stone, called the yati-gala, was sunk into the verandah of one of the traditional houses. It had a diameter of 25-27cm and was 11cm above the floor. Its upper surface was smooth and flat, though the sur­ face towards the edges was slightly worn. In the centre of the lower stone, a conically shaped peg was fastened in a hole. The peg was ca 3cm at base, thinning towards the top. It was made of gadumba wood. It was fitted; but with a lesser dia­ meter than the centric hole of the upper rotating part, called the matu-gala. The underside was smooth, but had the con­ cave form of a flat dish, probably through wear. The stone was cut so that two knobs protruded over the lower part, in 200° angles to each other. In one of them was a hole, fitted with a wooden stick made of gadumba wood, used as the handle, to draw the quern. The upper stone was 27x40cm in diameter, and 11.5cm thick. The conically formed hole in the middle had an upper diameter of 6cm and a lower diameter of 3.5cm. The kurakkan seeds were poured in th­ rough that hole, the conical peg of the lower part helping to force the seeds to the sides. When pulled, the quern rotated round the wooden peg but also moved slightly sidewise, as the centric hole was wider than the peg. This particular quern was made by a traditional Tamil stone-cutter from Digampataha village in 1935. It was transported to the village by bullock cart. His family carries on the stone-cutting tradition. Parallel to this traditional craft of cutting querns, ordinary male villagers in Nagalavava gam­ mandiya also dressed stones to be used as kurakkan querns. An example was the partly dressed lower section of a quern that was found outside one house. The wife of the house said an ordinary peasant of the village had worked the stone. He had learned the skill from the craftsmen. The quality of this layman work was not considered as good as that of the sp­ ecialists. Robert Knox, from his 17th century horizon, says of kur­ akkan processing that "When they are minded to grind it, they have for their Mill two round stones, which they turn with their hands with the help of a stick." (Knox 1981:112). Part of the population however still used grinding stone and roller at the end of the 19th century, as we noticed (above). John Davy, based on his own observations between 1817 and 1819, describes a similar item which he named the "cor- rican-galle, or stone hand-mill". It was used to grind kurakkan and other small grains (Davy 1983 (1821):209). Figure 18:21 A rotary hand-mill (kurakkan gala) in the verandah of a house in Nagalavava. The rotary hand-mill is a technical improvement of the grinding stone cum roller. The latter was in use for grinding flour as late as the beginning of the century, in some places of the island. Photo: Eva Myrdal-Runebjer. Figure 18:22 Chili grinding stone and roller (miris-gala) outside a kitchen in Talkote village. Note also the pottery stored on wooden frame to the right - a common sight in villages. Photo: Eva Myrdal-Runebjer. The grinding of kurakkan might be linked to a more ge­ neral discussion of social development. Take, as a starting point, that the kurakkan quern probably represents a labour- saving device within the dry-grain economy (this could be investigated). It also represents a larger amount of invested labour in the making of the quern. This is an interesting aspect when comparing different communities and also individual households, using either the quern or the grinding stone and roller. The relation of household wealth to household size has, of course, to be taken into consideration. The positive benefits of a labour-saving device such as the kurakkan quern are greater in a large household. What the large household makes pos­ sible perhaps, is to be able to feed household members not only when they are directly involved in food procurement activities, but investing labour in technological work such as cutting stone for making a kurakkan quern. It shortens the time women spend in grinding, giving them more time to take part in productive labour. It may be at this very basic level that we should seek the answer to what made the emergence of a stratified society in a given area possible. It is also at this level that we find the circumstances in which new techniques or cultural patterns are and can be adopted by a community. In our study area chilli and other ingredients for curry making are ground on the miris gala (chilli-stone). The miris gala we studied was a rectangular cut stone made from a piece of locally available rock (see fig. 18:22). The upper side was smooth by use. According to the owners, any stone can be used. This was 55cm long, 45cm wide and 14cm thick. It was placed outside the kitchen on a wood trunk 34cm high and stabilized by small pieces of wood. The roller was of stone, 18cm long with a diameter of 9cm. In other households the miris gala was placed inside the kitchen. Coomaraswamy gives the name "curry stone (kudu miris gala)" and describes it as "a plain slab and roller" (Coomaras­ wamy 1956 (1908): 147). John Davy writes that "a couple of smooth stones are used, one small and the other large, for mixing together intimately and grinding finely all the different ingredients of their curry." (Davy 1983 (1821):210). Robert Knox describes it 140 years before Davy: "A flat stone upon which they grind their Pepper and Turmeric, &c. With another stone which they hold in their hands at the same time,..." (Knox 1981: 236). For grinding various products used in traditional medicine, the behet ambarana gala (‘medicine grinder’) is used in Ta­ lkote (see fig. 18:23). Not every household makes traditional medicine, yet practising traditional medicine doesn’t neces­ sarily involve high specialization. The behet ambarana gala we measured was used in the household by the householder’s old father. He had also been practising a little carpentry for the neighbours, had previously acted as the vel vidane of one of the local tanks, in his capacity as one of the large landowners in the command area of that tank, doing paddy and chena cultivation and owning cattle, hens and previously, goats. The stone had been cut from a locally available rock. This particular stone had been in use for 30 years. It was placed on the ground inside the karatta maduva (shelter in the courtyard where the bullock cart, the plough and various other larger tools and implements are kept). It was irregularly trough­ shaped, its convex underside stabilized by a small stone, when used. It was 44cm long, at its broadest upper side 28cm and its narrowest, 13cm. It was at most 7cm thick. The roller me­ asured 9x6x5cm and was described by the users as an ordi­ nary stone, picked for the purpose. Figure 18:23 Medicine-grinding stone (behet-ambarana-gala) under a shed in a home garden in Talkote village. Most people gather leaves and plants which are dried and used for medicinal tea. Several knowledgable people of the village also make herbal medicines, using such grinding stones in the preparation. Photo: Eva Myrdal-Runebjer. It differs markedly from the behet gal from Balangoda, used in the preparation of herb and mineral medicines and for the grinding of colours by painters, described by Coomaras- wamy. The illustrations show two behet gal which have a square-cut outline and a marked and even oval depression for grinding. Below the protruding, square-cut grinding section four cobra heads are carved, one in each corner. The top of the pestle is carved into a head. The form of the pestle indicates a vertical grip by hand, in contrast to the horizontal grip dem­ onstrated on the roller of the Talkote behet ambarana gala, indicating a pounding rather than a grinding movement (C- oomaraswamy 1956 (1908): 147, Pl. LII.,7). We don’t know, however if such stones come from an ordinary rural household, or from secular or religious elite surroundings; and we know nothing of dating. The illustrated items do represent a larger investment of labour and a slightly different grinding technique than the Talkote stone, but does this simply mean a different socio-economic origin and tradi­ tion, or does it also indicate that they were used for different purposes? If left behind, all these grinding devices would of course be preserved in an abandoned settlement. Production of cooking oil Information was gathered about one type of cooking oil, namely oil from the znz-tree fruits. It was found that oil processing devices are not present in every household, but in two or three places in several of the villages under survey. People will come and use their neig­ hbours’ or relatives’ oil pressing device, if they don’t have any of their own. Only one family was questioned regarding the labour process, but their description was verified by others who knew. The work was described as follows: in September or October, one or two will go out and collect wz-seeds under the znz-tree in the village. The tree belongs to the village at large. They will do this from three to five days. In total, 5-10kg of znz-tree seeds are gathered. The seeds are left to dry in the sun for a week. They can be kept from three to five months. When the women have time, they will thresh the seeds in the vamgediya. It would take one woman about one and a half to two and a half days to thresh this quantity of seeds (probably only part time work is referred to). The following day she will winnow the seeds. They are subsequently boiled by one person and pressed by two of them. This will take approximately one and a half days. From this quantity of seeds five to ten glass bottles of oil (containing approximately 0.9-11.) can be made. It can be kept for three or four years. Oil for medical purposes could be kept for five years. This household consisted of eight people. Ten bottles were not sufficient for them, so they had to buy addi­ tional coconut oil. The oil-pressing device, named mi tel hindina pattalaya, is a log from the milla or buruta tree, divided lengthwise and pierced with holes at one end, in right angles to the cut (see fig. 18:24). A stick is thrust through the holes. The log is placed with one end in a natural bifurcation of a tree, about 0.8-lm above ground and the stick at the other end is sup­ ported by two forked branches, sunk vertically into the soil. Between the halves, a long, flat, basket-like device is placed, named a paha. The znz-tree seeds are put into the basket. A third stick is stuck into the ground, close to the tree. A string is pulled over the stick, wound twice around it. The other end of the loop is pulled under a fourth stick, which is put under the log with the string stretching over it. When the fourth stick is pulled downwards, towards the third stick, the two halves of the log are pressed together, thereby squeezing out the oil of the seeds into an earthen pot laced under the paha. The mi tel hindina pattalaya keeps from five to seven years. A similar device is described by John Davy at the beginning of the 19th century (Davy 1983 (1821):210). Preservation of meat Goat’s meat, was the only meat from domesticated animals preserved in the household. Surplus meat was also preserved from the wild animals caught - buffalo, elk, deer and wild boar. Regarding goat’s meat, it was reportedly cut into slices and dried above the fire, on the dummassa, for a few days. The dummassa is a rectangular or circular frame, made out of "any kind of wood", the same material forming a net in the middle, similar to a loose two-shaft fabric weave. After being dried and smoked above the fire, the meat was put into a bag and hung above the fire. The smoke from the fire prevented insects from attacking the meat, which could be kept for five months. If the meat was put into a bag filled with ash, it could be kept for a year. Another informant said the meat of wild animals was cut into slices, dried above the fire on the dummassa, and either put into bags and hung over the fire, or kept in pots filled with ash, where it could be kept up to one year. One former hunter, 81 years old, said he had met Vaddas in the Polonnaruva district 60 years back. They used honey for preserving meat. Parker writes of the "Wanniyas" and hunters in Sri Lanka at large, that they used to dry surplus meat on "a rectangular stick frame, over a slow fire,...” (Parker 1984 (1909):50). The pot of ash might make the storage of meat arch- aeologically visible. The most important fact however, is that it is possible to keep meat for up to one year, in the prevailing climate, through methods that are easily practised in a ho­usehold. This is considered to be mostly women’s work. Preservation of fish Surplus fish is dried for two or three days above the fire, on the dummassa, as is the meat (see fig. 18:25). It is then put into a bag and hung from the roof above the fireplace. With Figure 18:24 A m-oil pressing device (mi-tel-hindina-pattalaya) in a home garden in Kayamvala village. When the stick in the forefront is pressed down, the two logs press out the oil from the mi seeds in the basket. Photo: Eva Myrdal-Runebjer. larger fish, the head and entrails are cut away. In this way the fish can be preserved from three to six months. One informant said that 50 years back (when he was yo­ ung) dried fish was kept, mixed with ash, in large earthen pots. The fish could be kept for one and a half years. This again is considered mostly women’s work. Preservation of milk Nowadays the milk is mostly sold. What is kept is most often given to the children, for immediate consumption, or used for cooking sweetmeats. It was stated that a buffalo would give three to four litres of milk per day during the rainy season, but not more than one and a half to two litres during the dry season (if it could be milked at all). Two milk processing techniques, the making of clarified butter and curd, are still being practised. Clarified butter en­ ables storage for a longer period. The household that used to make clarified butter kept buf­ faloes. They described the process as follows: The buffalo is milked and the milk boiled. When the milk is cool, the fat at the top is removed. It is kept in the nabiliya (rice-cleaning, earthen basin) for three days. Then a wooden, pestle-like de­ vice, the mate, is used to stir it into butter. The butter is melted and the liquid poured into the clay vessel, or glass bottle, taking care to leave the protein containing lees. The clarified butter is used in curries, for example. The mate in this case was fashioned out of one piece of gansuriya wood. It was made in the household, using a knife. It was 0.34m long, the handle end 0.04m in diameter and the stirring end 0.07m. Curd was also made out of buffalo milk. The milk is boiled, then poured into an earthen basin, the kiri hattiya, and left overnight. It has to be eaten during the following day. This is considered women’s work. Preservation of chilli and wild plants, for beverages Chilli is dried in the sun in the courtyard, directly on the ground. Wild plants collected for preparing beverages are hung to dry from the roof inside the house, or in the verandah. Preservation of deer, elk and goat skin The skin of the deer or elk is used for making footwear - sandals, which are worn in the swidden field after the burn­ ing, to protect the feet. The skin is cut from the flesh with a sharp knife. It is put out in the sun to dry for three days. It is fastened to the ground by wooden pegs, so that it stays flat. The inner side of the skin is not cleaned in any special way, but used as the sole of the sandal. This is considered men’s work. Goatskin is used for the women’s New Year drums. It is dried in the household, like elk skin. It wasn’t made clear whether the processing of the hair and fleshy part of the hide is different. Deerskin is also used for drums. Figure 18:25 Fish preserved by drying on a wooden frame (dummassa). STORAGE We will first consider storage of primary products from the vegetable kingdom. Most important, from our point of view, is the storage of grains. A range of grain processing stages in storage have been documented in different cultural and en­ vironmental settings: as sheaves, as ears, as unhusked grains, as husked grains, as pre-boiled, husked grains, as flour, as bread, or other processed items such as beer and malt.8 Starting by identifying the various purposes for which grains were stored, four very obvious factors come to mind: for household consumption and further processing; for the next season’s sowing; for selling or barter; and for payment of taxes or rents. One quality which might concern the buyer, or tax or rent-receiving institution or households, is the germinative potential of the grains, which makes their further use a matter of choice - either for sowing, or further processing and con­ sumption, or for selling back to the cultivator as seed, or food, late in the season. During this study we found no indication of different storage for seeds meant for sowing. In Talkote we were told the grains were stored for the household’s own consumption and for sowing. Sometimes part of the produce was sold. This wasn’t because of a surplus, but because cash was urgently needed, and grains most often had to be bought at the end of the season. The selection of seeds for sowing, and if the seed corns were specially treated before sowing, remains to be documented. To estimate how much of the harvest a cultivator needed to store for the next season’s sowing, we must have a notion of return per area unit. This varies with different grains and seed quality, and also because of different techniques, water av­ ailability etc. Based on accounts of the amount of seeds sown and the return on a given acreage, we could conclude that the ratio of sown to harvested paddy seemed to vary considerably with different households. Regarding the traditional purana vi (pa­ ddy) varieties that one informant grew fifteen to twenty years earlier, the ratio of sown to harvested paddy was 1kg: 18-20. Then neither manuring nor transplanting was practised, ne­ ither insecticides or weedicides used. The Local High Yielding varieties (L HYV) he now uses give him a 24-26 return, he said. The return from LHYV of another informant was 10-13.3, and according to him this was little different from five years back, when he used the purana vi. Their respective statements might be in accordance with the factual reality, but the varying results must be caused by some difference in field practice, quality of the soil, or water availability; or even a difference in storage practices, influenc­ ing the germinative potential of the seeds. No indications of a difference in grain storage techniques have come to our notice however. This is probably a good illustration of the limitations of a questionnaire, as compared to actual field participation. That traditional wet-rice cultivation techniques do give a varied return is indicated also in the literature/ For us, it is most important to establish the return of the purana vi, in contrast to the hill paddy and millet. Millet poses a special problem: sown in grains, but harvested and stored up to food preparation in the form of ears. In Talkote, paddy is cut low, with much of the straw taken from the field. The whole harvest is threshed by buffalo-tram­ pling. It is stored as grain. The final husking however is done ‘meal-wise’, by pounding in the vamgediya. We saw very small portions of husked paddy being stored, and then only for a short time in flawed earthen pots which here found a secon­ dary use. On the other hand, kurakkan and similar dry grains such as idal iringu (sorghum) and tanahal (Italian millet) are cut just below the ear, and stored in this form. ‘Threshing’ and husk­ ing takes place at the same time, ‘meal-wise’, in the vam­ gediya. Meneri (Panicum millet) is harvested with the sickle. Whether threshed before storage, and where stored, has to be further studied. These observations are in line with Francois Sigaut’s remark that dehusked grains or groats of millet and rice can be kept for only a few hours, making it necessary to process them daily (Sigaut 1988:5). It is in fact, a question of wider significance, as Sigaut claims. The difference in storage between the processed millet and rice, and the wheat, barley or rye flour is of primary importance for the history of machinery. The possibility for every household "in temperate climates" to keep the milled flour made European development of bigger mills profitable, says Sigaut; whereas the fast deterioration of millet and rice prevented a similar development in areas dependent mostly on these grains. He further mentions parboiling, as practised for example in India, to make husked or groated rice more du­ rable (Sigaut 1988:5). The extremely short time limit given by Sigaut must be questioned. The women in Nagalavava gammandiya said the freshly ground flour of kurakkan tastes better, but that it is possible to store ground kurakkan up to three months in an earthen pot covered by a cloth, before insects (like the gulla) destroy it. As an argument favouring the fundamental difference be­ tween the grains, Sigaut states that millet continued to be pounded with mortar and pestle in western France well into the 19th century - though water-and-windmills were as com­ mon there as in other areas of France (Sigaut 1988:5). We could find an alternative explanation if we assume that millet cultivation wasn’t integrated with the market economy. Would it have paid to send the millet to the mechanised mills? Or did millet fit into the individual household’s yearly cycle of labour, as one of many minor, combined ways of finding a living? This seems to be the case in the swidden cultivation of millet in the Sigiriya area (see above). The wheat growing regions of northern India further indicate that the big Eu­ ropean mills are not just a result of the crops grown. Vi bissa The traditional structure for paddy storage in Talkote is the vi bissa (see fig. 18:29). The terminology of grain storage struc­ tures is a bit confusing also when consulting what has been written by people who have seen the structures they describe. Coomaraswamy talks of the bissa as a special type of small granary, found in the Kegalla district - "a cask-shaped wattle and daub erection on stone pillars, whitened with kaolin and covered with a heavy thatch, standing near the house but not forming part of it (Coomaraswamy 1956 (1908): 116). Coomaraswamy states that otherwise the grain was stored in the atuva - a term that in Talkote was used only for dry­ grain storage structures. Ashley De Vos, too doesn’t make any difference between paddy and dry grain storage facilities, but talks of them as storage bins named bihi - the plural of bissa (De Vos 1977:43). Coomaraswamy writes: "The granary (at­ uva), always raised on low stone pillars as a protection against white ants, was if small, built over some part of the inner verandah of the house; if large, as in the case of a royal store (maha gabadava), a pansala, or the home of a well-to-do or powerful man, it formed a separate building" (Coomaras­ wamy 1956 (1908):116). That such superstructure storage could involve large, or many structures, is indicated by a statement in the Sammo- havinodini, to which W.I. Siriweera calls our attention. There it is said that in the Tissamaharama monastery "were always" stored a quantity of grain that could feed twelve thousand monks, for three months (Siriweera 1989:84). We found no granary built on the verandah of any house in Talkote. In Nagala vava, the larger houses used rooms in the house itself for storage and the undifferentiated houses, of one room only, kept their produce in that same room. Previously, the village had storage facilities for dry grain, according to the informants. As Coomaraswamy’s statement indicates, as a way of distinguishing not only between superstructure and village level storage, but also the socio-economic position of different households within the village, this line of study is worth continuing. In the following we keep to the terminology used by the villagers in our study area. The vi bissa consists of a floor built of wooden planks on six horizontal wooden posts, resting ca 15-30cm above ground on six stone slabs (one in each comer and one at the two opposite sides). The stones are sunk into the ground only to level the floor. In case wooden pillars are used as a foundation, they will be sunk about 30cm into the ground. Separate from the floor, there are in all six wooden roof-supporting posts: one in each comer and one on the two opposite gables. They are dug down into the ground to a depth of 45cm. They are connected by horizontal posts resting in a natural crutch which constitutes the end of each vertical post. The roof is made of palm leaf or paddy straw. In the latter case, it is termed ‘pandaleya\ Underneath it paddy straw meant for fodder is sometimes stored, on a loosely laid floor of sticks. The floor plan of the vi bissa measures ca 2.4x2.4m. The pillar diameter is ca 18-20cm. The height of the construc­ tion is 2.9m. Under the roof, the paddy is stored in a 1.6m high cylindri­ cal construction, with a diameter of ca 2.3m. It is built of bundles of paddy straw horizontally piled one above the other, bound together by wickers (ibikatuval), forming a circular space which is successively filled with threshed, but not final­ ly husked paddy. When the height is considered sufficient, the vi bissa is sealed with a roof of bundled paddy straw - and thereafter plastered with a mixture of mud and cowdung. Its storage capability should be measured. For future archaeological identification, let us hope that foundation stones were used and left in situ. The only other possibility of identifying a free standing structure like the Talkote vi bissa, in a non-burnt archaeological context, seems to be finding the postholes, where their spacing could tell us the difference from a dwelling house - but maybe not from the traditional hen coops on poles, which are also there in the village. Figure 18:26 Rice in Talkote village is stored in the vi bissa. This is built with bundles of paddy straw and coated with a mixture of soil and cowdung. Photo: Eva Myrdal-Runebjer . Atuva In some of the villages a certain construction named the atuva was used to store the ear-by-ear harvested swidden grains (see fig. 18:27). The one documented was built on a wooden platform, measuring 2x2.6m. The*platform was ra­ ised on wooden posts, ca 0.6m above ground, without func­ tional connection to the roof posts. There was one vertical Figure 18:27 Millet is stored in the atuva, an airy granary with palm-leaf walls. Talkote village. Photo: Eva Myrdal-Runebjer. roof post in each corner, and one at each gable. There was no stone foundation for the wooden posts. The gable posts were ca 3m high. The horizontal roofing posts continued over the adjoining maduva. The walls were built of vertical and ho­ rizontal sticks, the horizontal sticks bound to the roof posts. Inside, this frame was covered and bound with palm leaves. The only access to the atuva was through a ‘gate’ at the upper corner of the gable, facing the inside of the maduva. Here a naturally bent beam, its diameter ca 0.1m, resting on the wooden corner posts and crossing the gable ca 2m above ground, constituted the end of the palm leaf wall. This left an opening ca 0.5x1 m wide, through which people had to crawl to get inside and take out the ears of grain. The roof was made of plaited palm leaves. In the literature we will find several meanings given to the word atuva. Coomaraswamy only says that the atuva is a granary. As noted above, he says it was only when built in the home of the well-to-do or powerful man that it formed a separate building, otherwise it was built over some part of the inner verandah (Coomaraswamy 1956 (1908):116). Coomaraswamy also states that the granary was always raised on stone pillars, to protect the construction from white ants. Our informants however said this was not necessary (unfortunate for the archaeologist). By digging 0.1-0.13m do­ wn around each wooden post and pressing ash and kohomba leaves around it, the white ants would be kept away. Accord­ ing to our informants the ants don’t like the smell of these leaves. As for the rats, they were considered no big problem for grain storage. They would (for a reason unknown to us) not come closer to the house than 3-5m. Most households had cats, though. Moreover, as the atuva and vi bissa walls were not connected to the ground, the rats would not be able to climb inside, according to the informants. Gelbert gives a different meaning to the word atuva-. he says that in the Anuradhapura area it is the place in the smoke above the kitchen fire (Gelbert 1988:162). Storage of chilli Dried chilli was kept in bags inside the house. Storage of dry fish, dry meat As stated above, the dry fish or meat could be kept either hanging in a bag, with or without ash, above the kitchen fire, or mixed with ash in an earthen pot and kept in the kitchen. Storage of dried wild plants Leaves for making beverages are dried and stored, hanging from the ceiling of the verandah, for example. Figure 18:28 The only access to the atuva is through this opening, facing the adjoining seed-processing shed. The bent beam is deliberately chosen to ensure easier access. Talkote village. Photo: Eva Myrdal-Runebjer. Storage of honey Honey can be stored for one or two years. If the honey is to be kept for a longer time, it is left in the honeycombs. These are put into an earthen pot, covered with a piece of cloth and placed near the fire, to protect the honey from ants etc. Storage of oil Mi oil nowadays is stored in glass bottles. It can be kept for three to four years. Oil for medical purposes could be kept for five years. It is stored in the kitchen. Storage of clarified butter Clarified butter from buffalo milk (duntel) is now most often stored in glass bottles. Previously it was stored in clay vessels (kiri muttiya) in the kitchen. Water storage facilities Water was carried from the well and stored in plain earthen pots with round bellies, short and comparatively narrow ne­ cks and flared rims. The pots closely resembled those des­ cribed by U.A. Gunasekera, P.L. Prematilleke and Roland Silva, in their article on pottery forms in Sri Lanka, as the present common water pot, the kale, or kalagediya. The pots were kept in numbers ranging from two to four, in the kitchen, or immediately outside. The orifice is protected by a saucer, or similar device. TOOL REPOSITORY AND MAINTENANCE Tools repository Smaller tools like the kurakkan katta and the ordinary sickle are kept inside the house, on a rack, or hanging from the wall. Larger implements such as the ard or plough, field levellers etc. are kept under a roof shelter in the courtyard - if they are not put under the vi bissa or atuva floor. Tool sharpening facilities When the individual household forms a separate working unit, with tools for clearing and harvesting, we could expect tool-sharpening devices to be among the household equip­ ment. In Talkote village however, we found only the katta gana gala (tota-grindstone) in some of the courtyards (see fig. 18:29). This is a device to which I have, so far, not found any reference in literature. The day-to-day sharpening of the kattas (long shafted billhook/axes) and poro (axes) is done on this stone, though the implements are taken once a year to the local blacksmith for professional treatment. The stone is taken from locally available material. No special kinds of rock are preferred for the katta gana gala, according to our informants. Not all households using kati and poro have a katta-gana-gala, instead the neighbour’s grin­ dstone is used. Which was the practice during the ‘gamman- diya time’ is not known. In Nagalavava only one household still used the katta gana gala. The grinding stone is sunk into the courtyard, to keep it stable, with only a minor part above the ground. We could find no pattern in its location except that it was placed outside the dwelling, within a maximum radius of 3-5m. We measured one item only, as it had to be dug up for us to see it entirely. The grinding part of the stone was lengthwise, concave, slig­ htly s-shaped. It was 27cm long and 10cm broad, which seems to be the ordinary size. The marks of sharpening operations were visible on the smooth upper side - thin, lines parallel to the long side. The height of the stone was 23.5cm on one side and 17cm on the other. It was irregularly six-sided, with len­ gthwise, flat sides, as if cut from a larger piece. Figure 18:29 The whetstone is sunk into the ground in the garden. The old man is sharpening his "bill-hook" (katta), Talkote village. Photo: Eva Myrdal-Runebjer. Figure 18:30 A pit for boiling tar for medicinal purposes. An example of many types of pits used in a household context. Leaves, twigs and sand has blown into the pit. Ilukvava village. Photo: Eva Myrdal-Runebjer The kurakkan katta (diminutive sickle used by the women for the ear-by-ear harvesting of dry grains) and the pan katta (an even smaller sickle with a long, iron tang, without wooden handle, used for cutting, dividing and flattening the pan straw, and sometimes used instead of the kurakkan katta in the dry grain harvesting) are sharpened with a pira (a small iron file) at home, when needed. Whetstones have not been used within the memory of our informants. The dakatta (sickle for bunch-cutting operations) has a serrated edge, and is taken to the blacksmith once a year for sharpening. In Hatampola-village, in the Ratnapura district, we saw a blacksmith fashioning the serrated edge with a tiny chisel and a hammer, sitting on the ground and stabilizing the sickle blade by holding it between his toes. This kind of work could then be carried out without any functionlly-needed associated structures. Pits in a household context On excavating a settlement site you will always find diffuse pits that have been filled in. During our work in the villages we found three different activities which would result in these pits. They will be presented here just as three examples of activities which leave such traces. Before the north-east monsoon has started, the soil is very hard. During that period, in Talkote in 1990, we saw a hous­ ehold about to remove a roof shelter in front of the verandah. To be able to pull out the roof posts, they dug around the post and watered it, waited for the water to soften the soil, and dug again. The result was a pit 0.4m wide around every wooden post. Wide pits in a row might then be indications of a house construction, though they don’t show the typical post-hole profile. In Ilukvava village, one informant boiled tar for medical purposes (skin diseases) from chopped up branches of wood. He burned them in his courtyard, in a pit ca 0.8m wide at the top and 0.4m deep. The tar would drip down into a pot placed under the heap of sticks. The pit was not in use when we visited the house in 1991, and its base was covered by dry leaves and twigs and soil that had blown in (see fig. 18:30). The last example has already been discussed regarding Nagalavava village: that is, various pits used for garbage. The pits had various origins. One, for example was dug in the course of obtaining building material for a house extension nearby. The garbage was not burned, but seemed most often covered with soil once the pit was filled. 1. Regarding household equipment and organization of space see MacDougall, R. D. and B. G. 1977. Reg­ arding house construction see for example De Vos, A. 1977 and De Silva, T.K.N.P. 1979. 2. See description by Codrington 1938:63-65; Leach 1961:53 ff. 3. For a discussion of the relation between caste and class affiliation in a Lankan context, see for example Gunasinghe 1990:105-135. 4. See for example Webster, D. and N. Gonlin. 1988. 5. See for example Smith, M.E. 1987. 6. See for example Messer, E. 1979. 7. Amy de Silva, PGIAR, personal communication Ap­ ril 1990. 8. See Sigaut, F. 1988:3-32, 46 for examples and bibli­ ographical references. 9. See for example Ellepola, D.B. 1947:97. 10. See Gunasekera, U. A., P.L. Prematilleke and R. Silva 1971:166-192. A similar description is also given by Coomaraswamy 1956 (1908):220 Pl.XXV 1-4.

Eva Myrdal-Runebjer and Anura Yasapala This presentation has two aims. One is simply to present the traditional hunting, trapping, catching and fishing techniques we documented in our Dry Zone study area during the field seasons 1989-92. The other aim is to widen the scope of archaeological interpretation. These techniques for food pro­ curement would to a large extent be invisible in an archa­ eological context. By knowing the technical possibilities we shall be able to view thTe osteological and palaeoecological material with additional questions. This presentation would preferably have been done as a result of a thorough ethnographical study. Unfortunately our archaeological fieldwork schedule left us no chance of con­ ducting such a study. No thorough presentation of Lankan hunting and trapping techniques has, however, been published previously. We also found the evidence fast disappearing, the­ refore we considered it urgent, from the point of view of cultural history, to present what has been documented, though the picture might not be complete. Fifteen male informants from different households in Tal- kote, Diyakapilla, Pidurangala, Ilukvava and Digampataha vi­ llages were interviewed regarding their own experience of hunting, trapping, catching and fishing. Eleven gave informa­ tion about hunting and trapping, three about fishing, three about the catching of wild buffaloes and three about hunting dogs. All but one were born in the area. The newcomer came from Anuradhapura district ten years ago. That district has an environmental situation similar to our study area. As the trapping season is later than our field season, most of the traps were built by our informants after the interviews, at our request. They were built in their true environmental setting, but one or two months earlier than usual. The excep­ tions are one functioning and two abandoned pit traps that we documented in one home garden, one abandoned pit trap in wasteland, one newly built deadfall in a chena fence, rabbit snares and a wild boar snare in a chena field and fence respec­ tively, and one abandoned arched snare, a monkey trap. Only one of the fishing devices had to be made, at our request. The reason of course is that in contrast to the animal traps, the fishing devices are movable and most of them can be used for several seasons. The traps built on request have parallels in other cultures, of which the informants could have hardly any knowledge Figure 17:1 Trap locations in cultural landscape. (South East Asia, Africa and north-west Europe, for example). Some of them are also mentioned by other field workers from Lankan evidence, and some of them are mentioned in the topographical literature or mediaeval texts (see references be­ low each heading). The survey by Michel Gelbert, reported in his book on swidden cultivation in Sri Lanka and two other socio-eco­ nomic surveys quoted by him, carried out during the 1980’s, presents a similar picture of where and when trapping and hunting is carried out. The techniques however are not des­ cribed in detail. It seems that hunting and trapping are more integrated in the day to day work of our study area than in the field survey area of Gelbert. No references to religious beliefs were made to us when asking about hunting. Bryce Ryan presents a sim­ ilar picture from Uva Province in the 1950’s. Though his informants are devout Buddhists, ’’hunting for meat is not considered sinful." (Ryan, Arulpragasam, Bibile 1955:157). Gelbert however states that slaughtering is diminishing in im­ portance in his study areas in the 1980’s, due to the increased influence of orthodox Buddhist emphasis on vegetarianism (Gelbert 1988:112-113,151-153). Regarding fishing, we also find references in pre-modern Lankan literature and in topographical works. It is further noteworthy that the hunting and trapping pat­ tern closely intertwined with cultivating practices, as we found in our study area, has been observed as typical of sedentary subsistence farmers in other parts of the world (Kent 1989). The close connection between small-scale subsistence farming and hunting and trapping is also documented in India. The basic outline of what animals and fish are caught, of techniques used, seasonality and environmental setting of the various activities, could therefore be considered set. What is still lacking is a documentation of the number of animals and amount of fish caught, and their relative importance as food in the household. What is presented here regarding this are only estimations by the interviewed households. Hunting and trapping techniques and what animals are sought are not socially neutral facts. Already the written sour­ ces give us some hints of hunting and trapping in different social environments. In Wilhelm Geiger’s translation of the Culavamsa, King Parakramabahu (1153-1186) goes hunting, the whole forest "surrounded by hunters with spears in their hands and nets..." (Culavamsa 1953. 70:35). Traps, bows and arrows and nets were the main equipment of the hunter, says the 13th century text Saddharma-ratnav- aliya, according to Ariyapala (Ariyapala 1968:340). From 17th century Kandy, Robert Knox tells us of the common people hunting with bows and arrows, and of trap­ ping from the point of view of agriculture and husbandry: trapping by placing sharp poles behind the fence to the "corn grounds", upon which the deer will land after jumping over the fence; by self-triggering crossbows to kill cow-eating leo­ pards; pit traps with a sharp pole in the bottom and deadfalls with bait for wild boar. He also mentions the catching of "Birds and Vermin". His concluding remark is true also of today’s situation: "And all of them (the traps) they make onely by the help of their Knives with green sticks and withs that grow in the Woods." (Knox 1981:137-140.). What we have documented then is hunting, trapping, cat­ ching and fishing as part of the common people’s economy in our Dry Zone study area. We have not tried to follow up what became of hunting from King Parakramabahu’s point of view. We further noted that trapping, hunting, catching and fish­ ing today are male undertakings. Our informants had learnt the techniques from their fathers and grandfathers. HUNTING, TRAPPING AND CATCHING, FOR FOOD; ANIMAL POWER AND PROTECTION OF CULTIVARS The main animal proteins consumed by the dominantly sub­ sistence- swidden cultivators came, according to themselves, from trapped or hunted wild animals. This statement is cor­ roborated by several socio-economic surveys (see Gelbert above). Skins and horns of certain animals also bring in a supplementary income today, as they are cleaned and sold to craftsmen. We found that 27 different land species were tra­ pped or hunted. As in the case of foraging for wild plants, we can see that trapping is closely connected to swidden cultivation. Traps are built on jungle paths leading to the swidden field, in openings in the swidden field fence at the end of jungle animal trails, on the fence itself, from the fence, behind the fence in the swid­ den field and in connection to crops grown (see fig. 17:1). The area round the swidden field is also searched for the den of the porcupine and the pangoline. Our informants were unanimous that protection of the swidden field and the additional food provided by the animals were equally important reasons for building the traps. Traps are also built in the jungle without a connection to a contemporary swidden field. Then it is preferably built in animal and wild fowl jungle tracks, or on previous swiddens where the open land would entice the animals, or close to waterholes during the dry season. Hunting and trapping are therefore mainly associated with the jungle and swidden cultivation. Two exceptions to this were found. One, the catching of wild buffaloes close to tanks and open spaces. The other, protection of home garden crops. Traps built close to the swidden fields are, of course, con­ nected to swidden cultivation. Hunting and trapping in the jungle would, however, be possible for the wet-rice cultivator to take up. We found it to be mainly among a few new settlers and young boys in the wet-rice village that this was actually done. Various explanations for this are possible. Meat from wild animals might be less needed, as tame animals provide some of the animal protein (milk products for example); the jungle proper is farther off from the wet-rice cultural landscape than the swidden-based village; less jungle area is available within the traditional village territory; the peasant household concentrating on subsistence-swidden cultivation, honey col­ lection and foraging, will have better knowledge of the forest and its hunting possibilities; and the additional income from selling skins and horns might be more urgently needed in such households, to mention a few possible reasons. Bryce Ryan notes of the Uva Province in the 1950’s that paddy cultivators seldom hunt, in contrast to swidden cul­ tivators. He gives no material reason for this difference but refers it to a different "frame of mind" of paddy cultivators (Ryan 1955:157). The environmental surroundings of different traps and hu­ nting and fishing techniques are given in fig. 17:1.The hunting and trapping season extends from June to April. Hunting with gun and dogs, in jungle or wasteland, can be carried on the year round. The informants stated however that the dry season - July and August especially - is the best hunting time. Then it is easier to move in the jungle and better for tracking and lying in wait, as the animals will more certain­ ly come - for example, to a waterhole. The swidden cultivating season is the main trapping season. Trapping can be carried out in the home garden the year round. The seasonality of various activities is presented in fig. 17:1. A description of the various traps and hunting techniques follows. As an appendix to this article is a list of the animals and fish caught. TRAPPING Fences As will be seen, the protecting fence around the home garden or swidden field is closely related to trap building. Several kinds of fence construction were noticed, using living, fast growing trees, or tree trunks and branches, or branches and twigs; but no special study of fences was made. Regarding the general question of time taken in making the fence, it was stated that fence building around that part of the swidden field cultivated by one household would engage three people for 10-12 days. Probably only part-time activity is meant. Figure 17:2 Pitfall with pit-spear (boruvala). Dug at our request. Approximately 0.4m shallower than usual. Digging was difficult as monsoon had not yet started and soil was as hard as cement. Diyakapilla village, October. Photo: Mats Mogren. Killing traps Pitfail with pit-spear (boruvala) The pitfail with pit-spear is dug after the rains have started, in October (the soil is too hard otherwise, as was the case when our pitfall was dug). Traps of this kind are used from the end of October to February. They are dug 2m deep, with a slight funnel shape (owing to lack of time, the pitfall built for our documentation was only 1.6m deep). The upper diameter is about 1.3m and at the bottom about lm. A sharpened wooden pole, about 0.08m in diameter, of mora or danduvaha wood, is dug into the bottom of the pit, to a height of about 1.4m above the pitfall base. The earth dug up is heaped at some distance, as it is said that some animals, especially the por­ cupine, will otherwise be suspicious and avoid walking over the trap.The pitfail is covered with thin branches, bark and over that dry leaves and some sand from the surrounding ground. It is dangerously well hidden below this cover. An axe and a vak pihiya are used to cut the wooden material. A hoe and an iron lever are used for digging. This kind of pitfall is usually built in animal trails (which are often the same as the footpaths men use) leading to chena fields or, as in our case, in a footpath in scrub jungle. The chena field is protected by the pitfall, while it also attracts animals which can then be more easily trapped. It will take two persons two or three days to make this trap. Because of the elephant threat, they will work in the jungle only between 9 am and about 3-4 o’clock in the afternoon. Our informants used to build one or two pits outside each chena field every season. The traps are used for four or five months and they estimate 10-15 animals will be trapped in their pitfalls during one season. The animals are usually killed by falling on the pit-spear. According to our informants, it is common practice for two or three persons to share the work and the bag from one pitfall, especially if they also work within the same swidden clearance. The trap is visited every morning and the trapped animals are divided between the people who took part in making the pit. The meat is said not to be destroyed although the animal is pierced by the pole. Porcupine (ittava), wild boar (val ura), elk (gona) and deer (miminna, titmuva, valimuva) are the most common quarry, but leopard (kotiya) was also mentioned. Figure 17:3 Covering the pitfail, dug in the middle of footpath in scrub jungle. Diyakapilla village. Photo: Eva Myrdal-Runebjer. Figure 17:4 The pitfall is covered and not easily detected (in front of the trap-builder's feet). Diyakapilla village. Photo: Eva Myrdal-Runebjer. Pitfall without pit-spear Deep rectangular Two deep pitfalls were found. One was dug in former chena land. It had been dug five years previously and was now abandoned. It measured 1.5x1.5m in circumference, and was 6.3m deep. When in use it was covered by thin branches and leaves. This kind of trap is mainly used for catching elk (gona) deer (miminna, titmuva, valimuva) and wild boar (val ura). The animal will either die by the fall or will not be able to climb up again. The men climb down a ladder to bring up the beast. If it is too heavy to carry, a special hoist is built. This device was only described, not documented. Two wooden poles from forked branches are dug in, with the fork turned upwards, on either side of the pit. Another wooden pole, pierced at one end, is placed in the fork. A peg is put through the hole and a rope is fastened at the middle of the horizontal pole. The rope is bound round the animal and by turning the horizontal pole, the animal is hoisted. The other pit was in use, dug in the home garden about 10m from the dwelling house. It was dug to protect the home garden from wild boar and deer. It was 1.7x1.7m in circum­ ference and 4m deep. Thin branches were laid above the hole and covered with a thick thatch of straw (sesamum indicum). This kind of trap is in use from October to March. According to the informant, the family didn’t consider it a danger to children or domesticated buffaloes and cows to have such a deep, hidden pit just behind the house. Cattle are kept out of the home garden and children will recognize the danger below the camouflage. Figure 17:5 Rectangular pitfail 4m deep, without pit-spear, in home garden ca 20m from dwelling house. Used to catch intruding wild boar and deer. Cattle are kept outside the home garden. Ihala Talkote village. Photo: Eva Myrdal-Runebjer. Figure 17:6 Deadfall (habaka) built into the chena-field "fence", Pidurangala village, late October. Photo: Eva Myrdal-Runebjer. Deadfall Habaka This kind of deadfall is built in an opening in the swidden field fence, by the farmer who built that part of the fence, that is, the farmer who cultivates that specific part of the field. It is in use from the end of October to March. The habaka is built where the swidden field crosses a jungle footpath. The purpose of the trap is to protect the swid­ den as well as catch animals. On either side of the opening there is a heavy, inclining platform, placed lengthwise along the fence. The platform is made of 10-15 branches bound side by side onto three branches across. The upper ones protrude outside the platform. Upon this about 10 wooden branches, ca 20.05-0. Im in diameter, are placed. The platform might also be loaded with stones, to give it additional weight. A vertical branch with forked top is dug down on each side, just behind the fence. The platform supports are placed in the fork. A kalaval rope is pulled over each platform support and the protruding stick of the platform’s upper end, thereby making the platform incline. The inner platform support holds the outer one down with the help of the key-stick. In front of the opening a ‘net’ is hung, made of kirival strings and horizontally placed sticks. It is held up by the key-stick. This is a wooden stick bound to the inner platform­ supporting stick. Through the platform weight the key-stick is pressed upwards, to a horizontal stick bound across the open­ ing above the net and towards the first horizontal stick in the net. When an animal tries to enter the (ca 0.3x0.3m wide) opening in the fence, it will be caught in the net, pulling it down. The key-stick is let loose and the platform will fall on the animal, usually crushing it. The habaka is examined at least three times a day. If the animal is not killed by the platform, the chena field watcher will kill it. It belongs to the person who built the trap. Wild boar (val ura) and rabbits (hava) are the most com­ mon prey, though bear (yalaha) and leopard (kotiya) were also mentioned by one informant. Mihabaka This deadfall is a rat trap. The one we documented was built as a protection against rodents in an irrigated bean field. It is also used in home gardens. About 25 sticks of about 0.01m diameter are bound by a kirival string, one above the other, to three sticks crossing them at right angles. A stick is placed at each end and one in the middle. The two outer sticks are longer than the platform. The stick in the middle of the platform is longer than the platform’s width. The platform is made to incline in the following way: a forked stick is sunk into the ground, bifurcated end upwards. A kalaval string is bound to the two outer, longer sticks of the platform. On top of the platform four stones (about 0.1m in diameter) are placed, to make it heavier. A stick is placed vertically in the fork. The kalaval string is pulled over the stick. The other end of the stick is bound to a kirival string fastened to the key-stick. Beside the platform, at mid-point a smaller stick is sunk into the ground. By the weight of the platform the key-stick is pressed upwards, against the longer horizontal stick in the middle of the platform and towards a stick placed horizontally under it. This in turn is held in posi­ tion by the smaller vertical stick outside the platform and under the platform end, resting on the ground. The bait - usually some kind of vegetable - is bound to the horizontal stick beneath the platform. When the rat slips in under the platform and tries to take the bait away, the key-stick is let loose and the rat is crushed. Rats are not eaten by man, but will be given to the dogs or cats, according to our informants. Figure 17:7 Deadfall (mi habaka) - a rat-trap built in irrigated bean field, Talkote village. Rats are not eaten by humans, but given to the dogs or cats. Photo: Eva Myrdal-Runebjer. Figure 17:8 Self-triggered gun (bandina tuvakkuva) in abandoned chena field. This one is adjusted in height to shoot deer. Talkote village. Photo: Eva Myrdal-Runebjer. Poison Rats (miyari) today are also killed by poison. Self-triggered gun (bandina tuvakkuva) Technically, this is a modernisation of the self-triggered spear or crossbow. The bandina tuvakkuva is located in jungle foot­ paths, or former chena land that will entice the animals, close to waterholes in the dry season, or where a footpath crosses a chena field during the chena season. A thin, metal cylinder is tied horizontally to two vertically placed poles. The height above ground is dependant on which kind of animal it is meant for. The cylinder is open only at one end. At the opposite end is a spear-like metal object, tied at its ‘tang-end’ to the cylinder, the ‘spear-point’ in the direction of the closed end of the cylinder. The object is bent upwards by placing a sturdy wooden stick under the ‘§pear-point end’.About 2cm beneath the ‘spear-point end’ there is a small hole in the cylinder, upon which a slightly funnel shaped metal orifice is soldered. This trap also functions through force and counter force on a key-stick. Two strings are fastened to a small tree, which is bent. One string is bound to the key-stick, the other to the sturdy, wooden stick beneath the ‘spear-point end’. The key-stick is placed at the vertical post, at the open end of the cylinder. By means of the string it is pressed towards a stick fastened horizontally to the vertical post. It is held in position about 4-5cm above the cylinder, by a string run over the cylinder and the key-stick. This string in turn is fastened to a thin, but strong thread. The thread crosses the footpath or some place where the animal is supposed to pass, and tied to a branch, or something the like, on the other side. The cylinder is charged with gunpowder and bullet. When the animal passes, the thread pulls the string off the key-stick, and the bent tree will be released. Straightening up, it will pull the sturdy stick supporting the spear-like object. This metal object will hit the orifice of the hole. The gunpowder explodes and the bullet shoots out; it might kill, or wound the animal. The trap is usually built to kill elk (gona), deer (miminna, titmuva, valimuva), civet cats (urulava), porcupines (ittava) or wild boar (val ura), but bear (yalaha) and leopard (kotiya) were also mentioned. Figure 17:9 Fixed spear (uladamilla) inside chena field fence. Height of spear and distance from fence depends on which kind of deer is the target. This is decided from footprints seen in the field. Ilukvava. Photo: Eva Myrdal-Runebjer. Fixed spear (uladamilla) The fixed spear is made of hard wood such as kirikon, seru or danduvaha. The pole is sharpened at the upper end. It is placed inside the chena field behind the fence, inclining tow­ ards the fence at about 25°. It is sunk 0.09-0.13m into the ground. It is placed where footprints of deer or elk have been noticed. The height of the pole and the distance to the fence depends on which animal left the footprints. The one docu­ mented was aimed for elk. It was about 0.05m in diameter, lm above ground and placed at a distance of 1.5m from the fence. The pole is blackened by charcoal from velam or damba wood, to be less visible. When the animal jumps over the fence it will be pierced by the sharpened pole. The animal might run some distance to get rid of it, thereby causing a more severe wound. Those watch­ ing the chena field will then kill the animal, if it isn't killed by the spear. Five or six fixed spears could be placed along the chena fence. The number depends on how many places deer and elk try to enter. Eight to ten deer (miminna, titmuva, valimuva) or elk (go- na) caught in one swidden season is said to be a normal tally. An axe is used to cut the pole and apihiya (knife) is used to sharpen it. Figure 17:10 Arched snare for squirrels (lena ugula), for protection of home garden crops in Diyakapilla village. Snare is built on a pole resting on the fence. Squirrels are sometimes eaten, sometimes given to the dogs or cats. Photo: Eva Myrdal-Runebjer. Snares Arched snare Len ugula Squirrel traps are built in home gardens and swidden fields, where squirrels are often seen to run. In one swidden field of six to eight acres, divided between two or three households, up to ten len ugula will be built in grain ripening times. There are several different, but closely resembling types, of which one is described below. A wooden pole is bound horizontally from the fence to a tree, for example. The trap is placed between the fence and the bait, consisting of kurakkan or tanavi, bound to the pole. The trap is an arched snare, resting on a key-stick. It con­ sists of one bifurcated twig placed with the forked end downwards. The two ends are bound on each side of the pole, leaving a large opening ca 0.15m between the pole and the twig. The snare is bound to a branch, which is bound to the tree (or naturally growing). A thin stick is pulled through the open­ ing of the snare. The stick is bound horizontally across the upper end of the twig. The key-stick is bound to the branch. The branch is bent and the key-stick is pressed upwards, towards the upper stick, held in place by a thin stick placed in the bifurcation about 0.03m above the pole. The squirrel will run along the pole, when it smells the grain. While running through the bifurcated twig it will touch the lower stick. The stick will fall, the key-stick is let loose and the branch, bending upwards, catches the squirrel between the snare and the upper stick. Some people will eat the squirrel, others will give it to the dogs and cats. They could not say how many squirrels are usually caught in one growing season, the event being less spectacular than, for example, the catching of elk and deer. Figure 17:11 Monkey trap (rila ugula) in swidden field. Snare is built on a pole resting on the fence. Diyakapilla. Photo: Eva Myrdal-Runebjer. Rila ugula I Monkey traps are of two types. Both are used in home garden and swidden fields. One type is meant only for the rilava monkey. Like the squirrel snare, it is built on a pole bound horizontally from the fence to two forked wooden posts in the field. Vertically in the ground between the posts, a long branch is also stood. It is bound to the posts and the pole. A twig is bent into a ring, the ends fixed by a kalaval string. The ring is bound to the pole about 0.5m from the fence, forming a kind of platform. Another twig is pulled under the pole, bent upwards, the overlapping ends strongly bound by a kalaval rope. The branch is bent towards the platform and the twig is pulled over it. A natural bifurcation, or similar, protrud­ ing element holds the twig in place. A corn cob or some other grain growing in the field, much liked by the monkeys, is strongly bound to the twig. The snare, nowadays usually made of metal wire, is connected to a wooden stick about 0.04m in diameter and 0.25m long. The snare is rounded and laid on the platform. At the other end of the stick another, stronger wire (an electric wire could be used) is fastened. This wire is bound to the bent branch. The monkey will come and sit, or stand on the platform, enticed by the corn cob. When it takes the corn cob it also lifts the twig from the bent branch. The branch bends back and the snare catches the monkey. The wooden stick and the strong wire prevent the animal from cutting the snare and running away before the watcher of the swidden, or home garden, comes along. The monkeys are eaten, though no estimation could be given of how many are caught in a season. Figure 17:12 Close-up of triggering mechanism in monkey trap. Snare rests on the twig bent into a circle (right). Diyakapilla. Photo: Eva Myrdal-Runebjer. Rila ugula II In this kind of trap no bait is used - it simply catches monkeys running on the fence around home gardens or swidden fields. This means that both rilava and vandura monkeys are caught. Also, this is a trap resting on a key-stick. Two wooden posts of velam are sunk vertically in the ground, just outside the fence. Two velam sticks are bound horizontally together with kirival strings, at either side of the upper end of the wooden posts, connecting the two. A small platform about 0.3-0.35m long and 0.2m broad is made with twigs of kuratiya, about 0.01m in diameter, bound together with kirival strings. It is stabilized by three twigs crossing them at right angles. The platform is balanced on the upper, horizontal fence post. Figure 17:13 Monkey trap (rila ugula) being set in swidden field fence. Talkote village. Photo: Eva Myrdal-Runebjer. Figure 17:14 Close-up of triggering mechanism in monkey trap. The animal is caught between the snare and the horizontal sticks. No bait is used. Photo: Eva Myrdal-Runebjer. A branch is placed vertically in the ground about lm from the wooden structure (or a small tree naturally growing there might be used). The key-stick and the snare, made of kalaval string, are bound to the branch. The snare passes between the two horizontal sticks of the wooden structure. The branch is bent towards the fence. The platform is held in position by the key-stick from the bent branch. The key-stick is pressed upwards, towards the upper horizontal sticks of the wooden structure and its lower end is placed between two platform twigs. The snare is made circular and laid on the platform. The key-stick is let loose when the monkey walks on the platform, as the platform will fall. The branch will bend back. The monkey is caught between the snare and the upper ho­ rizontal sticks of the wooden platform. Both kinds of monkeys are eaten. An estimated 10-12 monkeys are caught in this way during the four-month swid- den season. Up to six traps will be built along the swidden field fence. When a monkey has been caught in a trap, the swidden field watcher will have to change the position of the trap, otherwise other monkeys will avoid it. Figure 17:15 Porcupine and pangoline trap (kaballa ugula) in scrub jungle, outside a porcupine den. The snare is sprung when the animal treads on the small, leaf-covered horizontal stick. Talkote village. Photo: Eva Myrdal-Runebjer. Kaballa ugula This trap is also called the itta-ugula. It is built outside the den of the porcupine (ittava) and the pangoline (kaballava). They often stay in the same den. The area around the swidden field will be searched for the den. It is often situated beneath huge stone boulders where natural cavities are formed. Outside the entrance of the den a fence is constructed of wooden sticks up to 0.5m long. They are positioned vertically in the ground and horizontally placed sticks are bound to them. The natural vegetation or topographical features are used as much as possible, as part of the fence. At a suitable point a passage is left open. Two sticks are bound together horizontally over the upper end of the fence, on either side of the opening, about 0.2m above ground. Figure 17:16 Wild fowl trap I (valikukul ugula) being set in scrub jungle. Photo: Mats Mogren. The key-stick and the snare are bound to a small tree, or low branch close to the fence. The snare (nowadays made of electric wire) passes between the two upper horizontal sticks of the opening. The key-stick is placed resting on the two horizontal sticks, held in position by a third stick about 0.03m above ground. This stick is pressed towards the sides of the fence by the key-stick. Thin twigs are laid, inclining from the lower, loose stick, towards the ground. The lower twig is covered by dry leaves, and the opening is sometimes covered by a newly cut, leafy twig. When the animal treads on the lower twig, the key-stick is let loose, the bent branch rises and the animal is caught be­ tween the snare and the upper horizontal twigs. Often two traps of this kind are built by one householder during a single season. In one season he might catch up to 10 animals. The meat of both species is eaten and the quills of the porcupine are sold. Valikukul ugula This wild fowl trap can be built in two slightly different ways. Both are built during the swidden season. They are built where wild fowl are known to walk (in ‘wild fowl paths’, or feeding grounds). In the first variant, a fence about 0.2m high is built of horizontal twigs tied to vertical sticks, which are put down on both sides of the path. Any kind of tree could be used. The opening in the fence is about 0.25m wide. A key-stick is bound to a low, growing branch, or small tree nearby. The snare, made of kalaval, is tied to the same branch. Hook-shaped sticks are positioned on either side of the opening. A stick is placed horizontally under them. The branch is bent and the key-stick placed with one end beneath the horizontal stick, the other end resting on the ground. Thin sticks are laid, inclining from the ground towards the horizon­ tal stick. The snare is opened and laid upon them. When the wild fowl treads on the sticks, the horizontal stick is pushed down, the key-stick is released and the branch swings up. The wild fowl is caught in the snare. There are two, very similar ways of adapting the key-stick. Just one of them is described here. The other kind is made especially to catch the neck of the fowl. The fence is built only on one side and the snare is left hanging above the horizontal stick. Otherwise the construction is similar. Scraped coconut is strewn on the ground, and the horizon­ tal and inclining sticks are covered by dry leaves. The fowl, in eating the coconut, will trample on the horizontal stick and might get caught in the snare when the key-stick is released. The wood pigeon (kobeiya) and peacock (monara) are also trapped in this way. The branch will now be bent like a bow. A bait is fastened to the peg. When the parrot alights on the peg, trying to take the bait away, the peg moves. The snare is released and the branch swings up, catching the parrot’s leg between the snare and the branch. Figure 17:17 Wildfowl trap II (valikukul ugula) of a somewhat different kind being prepared in scrub jungle. Here minced coconut is used as bait. Both traps are built on "wild fowl paths" in the jungle. Diyakapilla village. Photo: Mats Mogren. Figure 17:18 Rabbit-traps (hadala) being laid between the jungle and a field in the home garden. The snares are tied together with grass straw, not to leave any passage between them. Diyakapilla village. Photo: Eva Myrdal-Runebjer. Simple snare Hadala This is a rabbit trap, built in openings in the home garden or swidden field fence, where rabbits pass to eat the crop. It is in use throughout the swidden season, for four months. A string is bound between two wooden sticks, in an open­ ing in the fence about 0.25m above ground. Simple snares made into a ring are bound to the string, one next to the other, hanging just above the ground. Where they touch, they are bound together with a grass straw so that there is no opening between them. The rabbit is caught in passing the snare. Nowadays thin metal wire is used, blackened with charcoal to be less visible. Previously kalaval strings were used. The traps we documented were strung on an 8m long wire, 56 snares in a row. This was considered a big hadala. Hadala traps are often built in two or three places along the fence. It was said that in all 20-25 rabbits could be caught in one month; sometimes two animals were caught during one night. The meat is eaten and skin thrown away, or given to the dogs. Figure 17:19 Simple Snare with camouflaged hindrance (urumanda) in a "wild boar path" in scrub jungle. Twigs from plants growing there are bound into a circle, and the snare tied to it with grass straw. The heavy chunk of wood tied to the snare will ultimately stop the wild boar. It is hidden to the right in the picture. Talkote village. Photo: Eva Myrdal-Runebjer. Madu yedima This is a snare simply laid on a jungle path. The other end of the string is bound to a tree, or rock in the vicinity. The leg of the animal becomes entangled in the snare and the harder it pulls to free itself, the tighter the snare knots. The animal is slaughtered by the trapper. Porcupines (ittava), wild boar (yal ura) and iguanas (tal- agoya) are the usual prey. Simple snares are also placed hanging from the hadala, on jungle footpaths. They hang to reach a porcupine’s head. This kind of porcupine snare was used by the trapper in question especially in January and February. None of these simple snares was documented. Snare with hindrance (chunk of wood) Urumanda This snare is laid to catch wild boar. The trapper is guided by the footprints of wild boar to determine where a ‘pig path’ occurs in the jungle. Often it is close to the swidden, or home garden fields. The use of the snare is not confined to the swidden, or growing seasons, though, and may be established in the jungle at any time of the year. The snare nowadays is made of metal wire; kalaval strings were previously used.Theendof the wire is fastened to a heavy log of wood - often a forked branch, or stem, is used. The one documented was lm long and each branch was 0.1m in diameter. This is hidden in bushes beside the path. Thin twigs of a growing plant along the path (kuratiya for example) are bent over the path and tied together, forming an arch. The snare is bound with grass straws to the upper part of the arch, hanging over the path. Thin twigs with leaves and/or grass straws or various herbs, are used to cover the rest of the snare, though leaving an opening in the middle. When the wild boar passes, his neck is caught in the snare. He will try to escape, but the heavy forked log of wood will soon become entangled in the vegetation and the boar will be unable to move very far. The trapper then slaughters it with an axe, or other sharp implement, or stick. Eight to ten such snares are usual outside a swidden field. Our trapper used to have five or six urumadu at the same time in the jungle, when it was not the swidden season. In one year five or six wild boar could be caught in this way, according to our informant. There is also a goymanda functioning in the same way, but adapted to catch the iguana (talagoya). This trap was not docu­ mented. Figure 17:20 Abandoned shallow pitfall without pit-spear (boruvala), in the opening of a "living fence" round a small mannokka field in a home garden. It protects the field from wild boar. The animals are caught in the pit and later killed by the home garden owner. Ihala Talkote village. Photo: Eva Myrdal-Runebjer. Figure 17:21 Inside an abandoned kraal for buffalo catching (gala) in scrub jungle south of Polattava. Photo: Eva Myrdal-Runebjer. Figure 17:22 Wild buffaloes are scarce in our study area today. Catching them is considered an undertaking for experts only. Here an old expert talks about his experiences to the team. Talkote village. Photo: Eva Myrdal-Runebjer Gonamanda The gonamanda is a snare attached to an antler. The snare is placed in a small opening of the chena field fence and the antler is placed beside it. When the animal passes, its head is caught in the snare. In tries to escape and the antler becomes entangled in the fence, or the vegetation outside. The animal is tracked and killed. This snare is aimed at wild boar. Pitfail catching without pit-spear. Narrow rectangular pitfall for wild boar. In one home garden we found two abandoned pitfalls, pro­ tecting enclosures for mannyokka cultivation. They were ma­ de for trapping wild boar, which are especially fond of mannyokka. On three sides around the small mannyokka field a ‘live fence’ (giniseeriya) is planted, taller than a man, with twigs horizontally twined between the closely growing trees. The opening in the fence is on the fourth side. In front of it a rectangular pit is dug. The pits documented were 1.7x0.7m in circumference and about 1.6m deep, and 2x0.8m and 1.3m deep respectively; and slightly decayed. The pits are covered with thin branches. The boar falls into the pit when trying to enter the field. As the pit is deep and narrow the animal will not be able to get out. The boar is then killed by the owner of the home garden. The pit is in use when the mannyokka is ripening. HUNTING METHODS, KILLING AND CATCHING DEVICES CATCHING Kudaya The wood pigeon (kobeiya) is caught with a basket (kudaya). The basket is placed in the courtyard outside the house. It is made with the opening inclining downwards and a rope tied at the upper end, while someone sits inside the house, holding the rope. Food liked by the kobeiya is strewn on the ground, under the basket. When the bird comes to eat the basket topples over it, as the rope is released. Similarly, the seasonal visitor, the avicciya, is caught with a basket in the jungle. The bird is either eaten or sold as a pet. None of these catching methods was documented. Net The wood-pigeon (kobeiya) is caught in a net cage placed in the home garden. This was not documented. Gala (wild buffalo pen) Buffaloes are caught in a kraal, in open wasteland, or not too dense jungle. Nowadays wild buffaloes are not common in our study area. One abandoned, decaying gala (wild buffalo pen) was however found in scrub jungle. It was situated close to former swiddens where the vegetation had not recovered.It consisted of an enclosure about 2.5-3m wide, 5m long and 1.5-1.7m high. It was built on three sides with wooden poles, about 0.1m in diameter, put down vertically in the ground, two by two, with 0.5-0.75m distance between each pair. Between the two rows of wooden poles thinner branches had been laid horizontally, one above the other. They were tied to the poles by kalaval strings. About 2m inside from the open (fourth) short side, a gate was built of the same material. On the outside of the enclosure leafy branches had been stuck as a camouflage. Such wild buffalo pens were built in paths they were kn­ own to frequent. Sometimes the animals were driven into the gala by battue. A strong net, kambadala, is also used to catch wild buf­ faloes. The net is placed crossing a footpath, and the buffalo is driven by battue towards the net. The buffaloes are either tamed or eaten. They are tamed by being tied with their heads put through a wooden frame fas­ tened in the ground - a hilinguva - for about 20 days; then they are tied to a tame animal for three or four months. Thus tamed they can be used as draught animals. If the buffalo dies of exhaustion while struggling to free itself, it will be eaten. None of these catching methods was documented. Figure 17:23 Device by which wild buffaloes are caught, a varamanda. Advancing behind a tame buffalo the catcher ties a rope with a deer antler attached to it to a leg of the wild buffalo. The antler will ultimately become entangled in the vegetation, trapping the animal. Anuradhapura Folk Museum. Photo: Eva Myrdal-Runebjer. Stalking with decoy animal In contrast to hunting and trapping in general, wild buffalo catching by stalking with a decoy animal is considered to be work for an expert. The catcher advances towards the wild herd, hiding behind a tame buffalo. The buffalo is caught by tying a rope bound to a sambhur horn (a varamanda) around one of its hind legs. When the buffalo tries to run away the horn is caught in the vegetation, hindering escape. The animal is then tamed as described regarding the catch in the gala. The old expert buffalo catcher we talked with was also a former vel vidane of one of the Talkote tanks. He was known to the surrounding villages too, as an expert buffalo catcher. If he caught buffaloes within his own village territory he would sell the animal and keep the money. Sometimes he was sent for, from other villages which had a herd on their territory. Then he would share the money from the sale with the people of that village who accompanied him on the catch - after he had met his expenses for the rope. This was not documented. KILLING Smoking out Porcupines (ittava) and pangolines (kaballava) are sometimes smoked out from their den and killed with sharp metal imple­ments or wooden sticks, while fleeing. This was not docu­ mented. Lying in wait Two methods are used for hunting: the man hides with his gun (with or without dogs) close to a waterhole. This is effective in the dry season only; he uses the dog to track the animal. John Davy writes in the first quarter of the 19th century regarding the "village Weddahs" he met in Uva: "They have dogs, but they do not employ them in hunting, excepting the Talagoya (iguana). They lie in wait for their game, or steal upon it when feeding, and kill it with their arrows." (Davy 1983 (18-21):88). Today no arrows are in use in our study area. The Ceylon yellow-legged green pigeon (batagoya), wo­ od-pigeon (kobeiya), wild fowl (valikukula), peacock (mo- nara), elk (gona), mongoose (hotambuva), porcupine (ittava), different kinds of deer (miminna, titmuva, valimuva) and wild boar (val uro) are shot by the hunter while lying in wait. Bear (valaha) and leopard (kotiya) are also hunted and, according to our informants, sometimes trapped. Their skins are sold. Their meat is not eaten, but partly used for medical purposes. Figure 17:24 Hunting dogs with their owner, Digampataha village. They are descendants of his father's pack. Almost every household has a watchdog, but not everyone owns good hunting dogs. Photo: Eva Myrdal-Runebjer. Figure 17:25 Sling-shot (galdunna) most often used to throw stones at intruding cattle, to scare them away. Small animals resting on tree-tops can also he killed by this method. An old, well known hunter demonstrates how to handle the device. Pidurangala village. Photo: Eva Myrdal-Runebjer. Tracking with dog Dogs are used to track the animals. The hunter will scour the jungle with his dogs, two or three days a week. Most hunters prefer the dry season as it is easier to move in the jungle at that time of the year, and lying in wait will also be more rewarding. The dogs know to which house they belong, as they always receive their food from there. One informant also bathed his dogs once a week. The hunter will select promising young puppies from ‘his’ bitches. He will take them on hunting trips with the older dogs. One informant however, a share-cropper and swidden cultivator owning three grown dogs and three puppies, said all dogs are good hunting dogs and no selection is necessary, but he preferred to keep bitches as they are con­ sidered more faithful to the house, and because they give birth to the next generation of hunting dogs. According to him, dogs usually live 10-12 years, but can be used for hunting only up to their eigth year. He will keep a dog until it dies a natural death. When asked if he would kill a dog too old to hunt, he strongly protested. The very idea seemed strange to him. Neighbours and relatives are given puppies; also non-hunting households usually have at least one dog, to watch the house. His father had kept dogs and his own dogs were their descendents, he said. Figure 17:26 Close-up of the galdunna. It was made by the user and needs repair approximately every third month. Photo: Eva Myrdal-Runebjer. Dogs are trained in hunting by older dogs. They are trained to track an animal and drive it towards the hunter, or to hold it and bark until he comes. The animal is sometimes killed by the dogs before the hunter arrives, especially the smaller animals, such as the big lizard (iguana) mouse deer and pangoline. When an animal is caught the dogs are given the parts not eaten by man. This gives them zest for the hunt. Some hunters have a personal relationship with their dogs, in the sense that they give them individual names - Ant (Ka- diya, that is hardworking), Long Hair (Bula), Bathiya and Tomia (claimed simply to sound good as a dog’s name) and "Pig-puppiess" (Urukikki, because her face resembled a pig’s when she was a puppy). Others called their group of dogs by a collective "name", a sound: Aiy-Aiy. The same and similar names are mentioned by the Seligmann’s, as given to dogs by Vaddas in Sri Lanka at the beginning of this century - Kadiya, Taniya and Komiya for example (Seligmann, C.G. and B.Z. 1911:117). When at home, the dogs are always watching. If of fierce character, they will be tied during the day but let loose at night. This is far from Sir James Emerson Tennent’s urban-biased statement of the mid-19th century: "There is no native wild dog in Ceylon, but every village and town is haunted by mongrels of European descent, which are known by the ge­ neric description of Pariahs. They are a miserable race, ac­ knowledged by no owners, living on the garbage of the streets and sewers, lean, wretched, and mangy, and if spoken to unex­ pectedly, shrinking with an almost involuntary cry. Yet in these persecuted outcasts there survives that germ of instinctive af­ fection which binds the dog to the human race, and a gentle word, even a look of compassionate kindness, is sufficient EVA MYRDAL-RUNEBJER AND ANURA Y ASAP ALA foundation for a lasting attachment." (Tennent 1977(1- 859):119). Other authors however have provided us with a less biased account of the use of dogs. Ariyapala quotes the thirteenth century text Saddharma-ratnavaliya, which refers to dogs us­ ed in hunting (Ariyapala 1968:340). Henry Parker further says of the dogs of the Vaddas and Vannis: "The dogs are trained for hunting, and will track any wounded animal, or follow up unwounded ones, through the thickest of jungle;...Well-trained dogs of this kind, of no par­ ticular breed, sharp-snouted, pointed-eared, little bigger than an Airedale terrier, in colour commonly yellow-brown or bl­ ack, the ordinary non-descript dogs that are seen in every village, are wonderfully intelligent in the forest." (Parker 19- 81(1909):90). The giant squirrel (dandu lend), mongoose (Jiotambuva) and one kind of deer (miminna) are tracked and killed by the dogs; or shot. The polecat (kalavadda) is killed by dogs at night. The iguana (talagoya) and the civet cat (urulava) are also killed by the dogs. The elk (gona) and two kinds of deer (titmuva, valimuva) are tracked and shot. Galdunna The galdunna is a bow-like device (a kind of slingshot) by which stones are shot to scare off intruding cattle in fields and home garden areas. However, it was said that the galdunna could also be used to kill smaller mammals and birds resting on the outer end of branches, or in tree-tops. This was not documented. The documented bow was lm long, made of katapila wood with a string of kalaval. It was made by the householder, and would be repaired every three months. FISHING In our study area, fishing was traditionally carried out the year round, with the exception of September. Dry-season fishing is conducted during the time when tank-based foraging of stalks and seeds is least rewarding, and the cows and buffaloes can’t be milked. We found 15 different species were trapped or fished. Catching with karaka, kudaya, or by hand The most important tank fishing season starts when the tank is drying out, in July, and continues into September, with a peak season in August, often the driest month. Fishing in tanks with hand manipulated basket devices, or by hand, was till very recently a collective undertaking of ten or twenty males. They would decide when to go fishing, start­ ing about 9 o’clock in the morning on the day agreed upon, Figure 17:27 Traditional fishing techniques. walking in line over the tank bed with their fishing devices. They would stop for the day at 3 or 4 o’clock in the afternoon. The catch was divided among them. We were told that earlier this century, when the village was more homogenous, the fish would be shared also with households not taking part in the catch, but having land below the tank. It was also said that anyone "belonging to the village" could take part in the fishing - he did not need to have land below the tank. As we noticed regarding Digampataha, it is mainly swid- den cultivating villagers with access to water tanks who also utilize the tanks for fishing. With the karaka, bigger fish were caught by placing the funnel shaped basket over the fish and removing it through the opening in the upper end. Siriweera notes that the fourteenth century literary work, Saddharmalankara, mentions the karaka as a fishing device (Siriweera 1986:17). Knox states that fishing is conducted in the dry period and mentions the use of the karaka during the latter half of the 17th century, in the then Kandyan state (Knox 1981:141-142). Sir James Emerson Tennent describes fishing with a similar device in the lowland, as he saw it travelling on the road from Colombo to Kandy in the mid-19th century. Here, according to Tennent, fishing started with the rains. Whether it is fishing in tanks that he describes, however, is unclear, as he terms the fishing ground "the hollows on either side" of that road (Tennent 1977(1859): 172-173). The kudaya, a bowl-shaped basket used also for other pur­ poses, is still used, dragged through the little water that is left, ’sieving’ it to catch the smaller fish. Both the karaka and the kudaya were made by male household members from locally available trees and plants.No reference to the ‘water-sieving’ kudaya has been found in the literature. Using bare hands to catch fish is mentioned in the fifth century Samantapasadika, according to Siriweera (Siriweera 1986:16). Bailing water out from small ponds and catching fish by hand is said by Henry Parker to be a typical fishing technique in the jungle villages of Vaddas, Tamils and Sin­ halese (Parker 1981(1909):51-52). In 1991 we observed this fishing technique at Sigirivava. A low bund was formed in the tank bed where water was still found. The water was then bailed out from the enclosure, by hand and bucket. Small cuts in the tank bed outside the enclosure were also used to drain off the water. The fish were caught by hand, lifted out of the mud and deposited in buckets. It must be noted that these techniques for fishing in tanks can’t be performed where the tank has a connection to peren­ nial water sources. Figure 17:29 Bailing out water from the almost dry Sigirivava, to catch the fish hiding in the mud by hand. Late August. Photo: Eva Myrdal-Runebjer. The use of poison The poisoning of water is said by Parker to be another typical fishing technique, in the Vadda-Sinhalese-and-Tamil jungle villages. Siriweera quotes the fifth century text Samantap- asadika, where reference is made to the use of poison in the water in times of drought, in order to stupefy the fish and thereby catch them more easily (Siriweera 1986:16). In our study area, the root of the kalaval is used for poison­ ing the water of tanks, natural ponds and streams that are drying. The root is crushed and then rinsed in the little water left in the tank, pond or stream. The fish come floating up. Figure 17:30 Creel (kemana) demonstrated by owner and builder, Talkote village. This type of stationary fish trap is used in flowing water during the rainy season. Photo: Eva Myrdal-Runebjer. Kemana With the rain in mid-October, fishing in the Sigiri Oya (a seasonal stream) begins, using creels (a kind of stationary fish trap - kemana). The kemanas are made by menfolk out of locally available plant material. They are stationed and sear­ ched singly, in contrast to the collective fishing in the tank. The kemanas will be searched every forth or fifth hour during the day. If many fish are caught, most of them will be dried in the sun and later smoked above the fire, in the stove. Figure 17:31 Lui kuduva, another stationary fish-trap for catching lullas of various sizes watching their roe. Talkote village. Photo: Eva Myrdal-Runebjer. Figure 17:32 Placing nets in water-filled tank, using a raft (paruva). Vannigamayavava Kosgaha-ala village, late October. Photo: Eva Myrdal-Runebjer. The technique of building stationary traps for fish in the streams seems to have been known in Sri Lanka long back. There is no archaeological evidence to cite, but Samantap- asadika, Buddhagosha’s 5th century commentary on the Vm- aya Pitaka, mentions among other fishing devices also the kumina, which according to Siriweera is the Sinhala kemana. The medieval Saddharmalankara also refers to the kemana (Siriweera 1986:16-17). Our 17th century source, Robert Knox, doesn’t explicitly mention the kemana, although he mentions the construction in the Mahavali ganga of some kind of bar­ rier, leading the fish into "fish-pots" placed between rocks (Knox 1981:142). Nearly 200 years later, Tennent mentions the creel in Sri Lanka, in a footnote, but without indicating its geographical distribution at that time. Stationary fish traps of the creel type are examples of a universal technique. Creels have been preserved from the Me­ solithic Kongemose-culture in Denmark (60007-5200 BC), north-west Europe, and they were used by several ethnic gr­ oups in northern Laos, South-East Asia, in the 1940s (Izikowitz 1951:175). As in Sweden (north-west Europe) in the 20th century. Lui kuduva Another stationary fish trap is the lul kuduva. It is used in the wet season when the lullu are spawning, and after the roe is laid. It is made out of divided kirival branches bound to a coni­ cally spiral, basket-like device, open at the base, closed at the top. Two sizes were shown us, one 0.2m long and 0.06m wide at the base; the other 0.4m long and 0.15m wide at the base. They were, of course, meant for fish of different size. When a spawning pair is seen, the lul kuduva is placed at the bottom of the tank. The fish swims into it but cannot turn back as gills and fins are caught in the sides of the lul kuduva when it tries to back out. The male lula will watch over the roe. Where a male with roe is observed, sticks are put vertically into the tank bed around the roe, leaving a small opening where the lul kuduva will be placed. The male, trying to save the roe, will swim into the lul kuduva, and is caught as described. Our informant stated that he usually makes eight to ten lul kudu for a season. He will catch 20-25 lullu in this way. No reference to the lul kuduva has been found in published material. Net Nets (data, vis id ala) are used in the water-filled tank. The nets in Talkote are of nylon, bought ready-made. They are stationary nets, as compared to the cast-net of the coasts. During the north-east monsoon they are placed in the tank, in water five to six foot deep, fastened vertically by two wooden posts. Sometimes a small raft (paruva) is used to go out to the tank. One net was said to catch a quarter to half the amount of fish given to each partaker in one day’s work with the karaka. Siriweera calls our attention to nets being mentioned in the 5th century Samantapasadika text, and in the 13th century Saddharma-ratnavali. Whether cast or stationary nets are indi­ cated, or which part of the country is referred to, is not made clear however (Siriweera 1986:16-17). According to Knox, fishing with nets was not important in the 17th century in the part of Sri Lanka he knew (Knox 1981:142). There is no reference to fishing with nets in the mid-19th century text by Tennent. Samuel Baker writes how­ ever from his mid-19th century horizon : "Every little boy of ten years old along the coast is an adept in throwing the casting-net." (Baker 1983 (1855):199). Parker mentions that it is only among the (east) coast Vaddas that nets are used, around the turn of the century (Parker 1981(1909):51). It is not known however whether stationary nets are a quite recent introduction in our study area. Hook Steel-hooks (bilikokka) on strings are used in tanks and in the stream/canal the year round, where there is sufficient water. The hooks are bought. Angling is an individual undertaking and the fish caught is meant for immediate consumption. Siriweera refers to the 5th century Samantapasadika and the medieval text, Saddharmalankara, as mentioning hooks for fishing (Siriweera 1986:16-17). Knox has no reference to hooks. Tennent states in the mid-19th century : "As anglers, the native Singhalese exhibit little expertness..." (Tennent 19- 77(1859): 173). According to Parker, it is at the turn of the century that only the Low-country Sinhalese and the Tamils use the fish-hook, whereas it is not used by the Vaddas, the Vannis or most of the Kandyan Sinhalese (Parker 1981(19- 09):51-52). We do not know if angling was introduced in our study area in modern times. Kastana patiya The kastana patiya was another fishing device used in the water-filled tank in Talkote. It is a flat iron blade ca 0.75m long, fastened to a wooden shaft. The fish approaching the surface are knocked and killed by the iron blade. No reference to the kastana patiya has been found in the literature. Appendix: Animals Used in the Household Domesticated animals eaten Buffaloes: sold to Muslims, meat bought from them, horns used. Cows etc: sold to Muslims, meat and hides bought from them. Goats: slaughtered by male family member or good hunter from the village, meat eaten, skins for womens’ New Year drums. Poultry: sold or slaughtered by male family member, meat eaten, eggs sold or eaten. Animals caught: (usage and trapping/hunting devices) Avicciya (Indian Pitta, Pitta brachyura brachyura (Linna­ eus): Comes during Maha season. Eaten, or sold as a pet. Kudaya. Batagoya (Ceylon Yellow-legged Green Pigeon, Trero p- hoenicoptera phillipsi): Meat eaten. Gun, galdunrta. Dandulena (Giant squirrel, Genus Ratufa): Meat eaten. Gun, or caught by 3-4 dogs when running. Girava (Parrot fam. Psittacidae): Kept as pet or sold. Gira tonduva. Gona (Sambhur, Elk, Cervus unicolor unicolor)'. Meat eaten, hides (drums) used, horns used or sold. Gun, bandina tuvakkuva, boruvala, uladamilla. Hava (The Ceylon Black-naped Hare, Lepus nigricollis singhala): Meat eaten. Habaka, hadala. Hotambuva (The Ceylon Ruddy Mongoose, Herpestes smithii zeylanicus): Meat eaten. Gun, dogs. lbba (tortoise): Meat eaten. By hand. Ittava (The Indian Crested Porcupine, Hystrix indica): M- eat eaten, quills sold. Kaballa ugula, smoked out from den killed with a wooden stick or axe, boruvala, bandina tuvak­ kuva, gun, madu yedima. Kaballava (Indian pangoline, Manis crassicaudata): Meat eaten. Kaballa ugula, smoked out from den; killed with wo­ oden stick or axe. Kalavadda (Palmcat or Polecat, Genus paradoxurus): M- eat eaten. Killed by dogs at night. Kobeiya (pigeon, Fam. Columbidae)'. Meat eaten. Gun, net, valikukul ugula, kudaya, galdunna. Kotiya (The Common Indian Leopard, Panthera pardus fused)'. Dried meat for medical purposes (to cure "adum skin sold. Gun, bandina tuvakkuva, habaka, boruvala.Lena (squirrel, Genus funambulus): Meat eaten or given to dogs/cats. Len ugula, galdunna. Miminna (Mouse-deer, Tragulus meminna): Meat eaten. Gun, bandina tuvakkuva, boruvala, dogs, uladamilla. Miya (rat): Given to cat/dogs. Mi habaka, poison. Monara (Common Peafowl, Pavo cristatus): Meat eaten. Valikukul ugula, gun. Rilava (Red Monkey, Macaca sinica sinica): Meat eaten. Rila ugula I and II. Talagoya (iguana): Meat eaten. Killed by dogs, eaten by men, goymanda killed with wooden sticks, madu yedima kil­ led with wooden sticks. Titmuva (deer, Sub-family Cervinae): Meat eaten, hides used (drums). Gun, bandina tuvakkuva, boruvala, uladamilla. Urulava (The Ceylon Small Civet-cat, Viverricula indica mayori): Meat eaten. Bandina tuvakkuva, dogs. Vandur a (The Ceylon Grey Langur, Semnopithecus pr- iam thersites): Meat eaten. Rila ugula on fence. Valaha (The Ceylon Sloth Bear, Melursus ursinus inor- natus): Fat used for medical purposes, skin sold. Gun, bandina tuvakkuva, habaka. Valikukula (Ceylon Junglefowl, Gallus lafayetti)'. Meat e- aten. Valikukul ugula, gun. Vai miharaka (Water Buffalo, wild, Bubalus bubalis bu- balis)'. Tamed or meat eaten, horns used. Caught in kraal {gala) or by tying the varamanda, or by net made of rope. Vai ura (The Indian Wild Boar, Sus scrofa cristatus): Meat eaten. Gun, bandina tuvakkuva, boruvala (of several kinds), urumanda. Valimuva (Barking deer, Muntiacus muntjak malabar- icus): Meat eaten, hides used (drums). Gun, bandina tuvak­ kuva, boruvala, uladamilla. Fish caught: (fishing devices used) Ada (Anda? eel, Genus auguilla): Kudaya, kemana, kastana patiya. Ankutta (Macrones keletius): Kudaya, hook. Hirikanaya (Labeo dussumieri (Valeniiennes)): Karaka, kemana, net, hook. Humga (Stinging catfish, Heteropneustes fossilis): Kud­ aya, hook. Kanaya: Kudaya, kemana. Kokasa (Ompok-bimaculatas (Bloch) Siluridae): In Aug­ ust Karaka. In rainy season net, kemana, kastana patiya. Koraliya (small) theppiliya (big) (Pearl spot, Etroplus su- ratensis): Karaka, kemana, kastana patiya, visidala. Lula (Snakehead, Ophicephalus striatus): "Baby lulla" by the kudaya, big lulla by the kemana, karaka, net, in Septem­ ber: lul kuduva (caught while watching the spawn). Madara: "Baby" Madaras by the kudaya, big Madaras by the kemana, karaka, in September: lul kuduva (caught while watching the spawn). Magura (Clarias teysmanni brachysoma (Gunther)): Ku­ daya, kemana, hook. (Poisonous sting). Mai koraliya (Etroplus suratensis (Bloch)): Karaka, ku­ daya, hook, kastana patiya, net. Petty a: (Puntius, Barbus (-Puntius) Spp.): Karaka, ku­ daya, kemana, net, kastana patiya. Teliya: (Spine eel, Mastacembulus). Kudaya, hook. Tittaya (Panchax melastigma (Maclelland)): Kudaya, kemana. Valigouva (Lepidocephalus jonklasi Deraniyagala): K- araka, kudaya, kemana or net. Net: April to July. Karaka, Kudaya: July to August. Hook: Year round. Kastanapatiya: Rainy season (north-east monsoon). Kemana: Rainy season (north-east monsoon). Lul-kuduva: Rainy season (north-east monsoon). Poison (kalavalmul-xQQt): Dry season (approximately Ju­ ly to September). Animals which are not eaten as food according to the informants Bear (used for medical purposes) Dog Domesticated cat Elephant Kabaragoya (Monitor lizard) Leopard (used for medical purposes) Rat Snakes REFERENCES Fernando, C.H. 1990. Freshwater Fauna and Fisheries of Sri Lanka. Colombo: NARES A. Phillips, W.W.A. 1980. (2nd rev. ed.). Manual of the Mam­ mals in Sri Lanka. Part II. Colombo: Wildlife and Nature Protection Society. Phillips, W.W.A. 1984. (2nd rev. ed.). Manual of the Mam­ mals in Sri Lanka. Part III. Colombo: Wildlife and Nature Protection SocietyREFERENCES Fernando, C.H. 1990. Freshwater Fauna and Fisheries of Sri Lanka. Colombo: NARES A. Phillips, W.W.A. 1980. (2nd rev. ed.). Manual of the Mam­ mals in Sri Lanka. Part II. Colombo: Wildlife and Nature Protection Society. Phillips, W.W.A. 1984. (2nd rev. ed.). Manual of the Mam­ mals in Sri Lanka. Part III. Colombo: Wildlife and Nature Protection Society

EVA MYRDAL-RUNEBJER "Thus in a sense, the history of the Island has been viewed in the main as the history of the activities of kings, while the life of the ordinary people and their economic pursuits have only attracted cursory attention. Not surprisingly, this over­ emphasis on political and religious history has led to a neglect of social history in general and economic history in par­ ticular. " (W I Siriweera, Department of History, University of Peradeniya).1 Irrigation and swidden: two technological complexes It is a well known fact that Dry Zone subsistence peasants of Sri Lanka, dependent upon village based irrigation - that is, using rain-fed tanks - have to combine paddy cultivation with swidden cultivation, foraging wild plants, fishing and, to so­ me extent, hunting and trapping, to support their households. To define the Lankan Dry Zone economy, focusing on the combination of small-scale paddy cultivation and swidden cul­ tivation, as seen today, does not however give us sharp enough tools for analysis of changing land use and subsistence pat­ terns in the past. This is not to say that the combined economy is a modern phenomenon. But in focusing on the combination, we lose sight of the varied implications of each cultivation activity as regards yield and organization of labour. Hence due attention is not given to the potential social implications of one activity dominating the other. Viewing each cultivation activity separately, from the vi­ ewpoint of the cultural and social landscape created - to­ pographical setting, crops, traditional views on the disposition of land, seasonality, mobilization of labour (number of pe­ ople), organisation of labour (relation between those working the land and those having it at their disposal), labour invest­ ment (man hours), implements and helping aids, productivity area-wise, storage techniques and combined activities - differ­ ing practices and returns of irrigated and swidden cultivation respectively, are seen today in the study area. This difference is, of course, also what creates the peasant’s capability of combining the two cultivation techniques.2 It is also clearly apparent that the cultural landscape of wet-rice cultivation rules out some of the activities of the subsistence-oriented swidden landscape. Figure 16:1 Aerial photograph, Talkote village. Scale 1:3500. Note the mixed wet-rice and swidden cultivation landscape. Photo: Survey Department This implies that two factors with a direct bearing on the foundation of social organisation - that is, organisation, mo­ bilization and investment of labour, and productivity area-wise - differ in the two cultivation activities. At the same time, the ethno-archaeological study gave reason to hope that these fac­ tors could be archaeologically traced. Ecofact and artefact material from a modern settlement site of the kind we saw, as well as structural remains associated with cultivation activities, would indicate whether irrigated or swidden subsistence cul­ tivation had been practised. The case of commercially oriented swidden cultivation might also be posible to single out. The presentation below should not be seen as an attempt to develop a static model, or a closed, neatly balanced system. It is based on an exploration of a tiny fragment of our material world and is an attempt to understand its limitations and op­ portunities - an understanding that is useful for a future discus­ sion of the archaeological material. The Dry Zone, a spatial and temporal outlook Presenting the villages under study, we might start with a brief temporal and spatial outlook of the Dry Zone. John Davy’s description of subsistence cultivating "village Weddahs" of the Uva region in the second decade of the 19th century, gives (besides the biased view of a colonial civil servant of the 19th century) a picture of a materially im­ poverished people: "Their dwellings are huts made of bark of trees; their food, the flesh of deer, elk, the wild hog, and the iguana, with a little Indian corn, and corican of their own growing, the wild yam, and the roots of some water-lilies. They use, besides, honey and wax; and in times of scarcity, decayed wood, which, mixed with honey, and made into ca­ kes, they eat, not so much for nourishment as to distend the empty stomach..." (Davy 1983 (1821):88). Parker gives his account and that of another Ceylon civil servant, Hugh Nevill, of the hunting/gathering/subsistence s- widden cultivating economy, chiefly in the Eastern Province and the Vanni area in the late 19th century: "The food of the Forest Vaeddas consists of fruits, roots of wild yams, and especially honey and the flesh of any animal they can kill, which are chiefly Iguanas, Pigs and Deer. All the Village Vaed­ das, and the Tamil-speaking Vaeddas (with the exception of a very few who are solely fishermen), and the Wanniyas eat the same food, and have in addition the small millet above-men­ tioned, called Kurahan by the Sinhalese the Indian Ragi(Eleusine coracana). This is grown in temporary clearings (termed hena in Sinhalese) made in the forest, all bushes and grass being cut and burnt off, but not the larger trees. After one crop, or sometimes two have been taken off the ground, the clearing is abandoned and allowed to be overgrown once more with jungle, and is not re-cultivated until from five to seven years have elapsed. In their clearings, which are exactly like those of the Sinhalese, are also grown a few red Chillies and Gourds, and sometimes a little Indian Com, and a small Pulse called Mung (Phaseolus mungo). A very few Village Vaeddas and Wanniyas who live in suitable places for it grow and irrigate a little rice, which the Forest Vaeddas are now learning to cook and eat when they can procure it”. Dried and ground seeds of the Tree-fern (Cycas cirinalis) and the bud of unopened leaves at the crown of the wild date (Phoenix zeylanica) are also mentioned as eaten by the Vad- das, Vannis and Sinhalese villagers (Parker 1984 (1909):50- 52). Reading A.C. Lawrie’s A Gazetteer of the Central Pr­ ovince indicates that the situation in our study area, at the end of the 19th century, was far removed from a prosperous system of combined swidden and paddy cultivation. Of the 22 pur ana villages of Inamaluva Korale (where most of our field work so far has been concentrated) mentioned by A.C. Lawrie in 1898, three were abandoned villages at the time of his survey and two were explicitly mentioned as not doing paddy cultivation. Of 17 villages associated with functioning tanks, only one was said to cultivate paddy every year, whereas 13 were explicitly mentioned as not doing so for lack of water (three "seldom” or "infrequent”, one "once in ten years", nine "once in three or four years"). Swidden cultivation is explicitly stated to be the support of two of these nine villages cultivating paddy "once in three or four years".3 We should of course be careful not to take it for granted that the pre-modem situation for subsistence cultivating fa­ rmers was similar to the colonial. We might, for example, ask if there was an economic decline in the Dry Zone, from the end of the 17th century to the mid-19th century. If we read Robert Knox’s description of the one-time centre of the Ra- jarata civilization - the area around Anuradhapura - we get an impression of habitation and irrigated agriculture, though at a lower level. Knox says about the situation in 1679: "To Anarodgburro therefore we came,...which is not so much a particular single Town, as a Territory. It is a vast great Plain, the like I never saw in all that Island: in the midst whereof is a Lake, which may be a mile over, not natural, but made by art, as other Ponds in the Country, to serve them to water their Com Grounds. This Plain is encompassed round with Woods, and small Towns among them on every sid- e...Being come out thro the Woods into this plain, we stood looking and staring round about us, but knew not where nor which way to go." (Knox 1981:353). About the health situation he says: "The Northern parts are some what sickly by reason of bad water, the rest very health­ ful." (Knox 1981:96). In the mid-19th century, Tennent gives a different picture of desolation reigning over the plains (Tennent, J. E. 1977 (1859):361). Until a careful investigation is carried out regarding this late historical period, nothing definite can be said; but when handling our ethno-archaeological material we should again think about what is materially given (in this case, what security a subsistence economy could have, given the environmental and technical situation) and what is historically specific (in what way the social set-up and forces from outside might have influenced the situation). Presentation of the study area Three villages around Sigiriya formed the core of our study, chosen to cover different aspects of rural life that might have been of relevance also at the time of the formation of the (historical period) archaeological record. The choice was ba­ sed on interviews by the SARCP Village Inquiry Unit in 1988 (Wickremesekara 1990:161-166). One is Talkote village. In 1981 the village had a population of 581 persons.4 The cultural landscape in and around the village is formed by irrigated agriculture, and still much of the seasonal work is organized around paddy cultivation and what was found to be related to it - that is, fishing in tanks and cow and buffalo farming. The earliest known mention of Talkote is on the Mirror Wall, in an inscription from the late Kandyan period, the 18th century. It has various minor monastic remains within its present boundaries and a minor test excavation was conducted at one of the sites in 1989. The former village site is now abandoned. The desertion started 50 years ago. To develop our ability to gain information of settlement sites without large scale excavation, sites such as these might be used to detect activity areas. Intensive field-walking could be done, armed with previously culled knowledge of housing and storage facilities and having the villagers to inquire from; and coring for soil samples; or using radar and mapping pot­ sherd concentrations to detect garbage pits, or dwelling house floor areas. If just field-walking is done, the researcher must be conscious of the fact that the habitation site has gradually shifted a short distance, without haste or disorganization. That is, whatever could be useful for the new house moved away without cost of great energy, including, perhaps, grinding st­ ones and foundation stones for storage bins. Figure 16:2 Site location map. This fact has a significance beyond the pedagogical prob­ lem. There is a possibility that many of the pottery sites that have been found might be the result of such short-distance shifts. So the artefact material and structures that give ho­ usehold-level information on production processes, patterns of storage and household wealth, might have an extremely de­ lusive and fragmentary character, consisting mainly of wasted pottery and floor areas invisible on the surface. This is an important difference, compared to abandoned elite-structures in the area, where more durable and not easily movable build­ ing material has been used. A very large settlement site is situated in an area named Tammanagala, at the southern end of one of the larger village tanks (Ihalavava). This settlement site was partly excavated in 1990. Talkote, then, is the most important village from an educational point of view, where we can work in parallel with ethnographical and archaeological studies. Another village is Diyakapilla, dependent mostly on swid- den cultivation and related activities, such as foraging wild plants, collecting honey, hunting and trapping and, to a lesser extent, paddy cultivation. In 1981, 132 persons lived in the village (Wickremesekara 1990:162). The tank is situated in the neighbouring village Kosgaha-ala, with which the Diyakapilla cultivators share the irrigated area. Some paddy is also cul­ tivated by building weirs in a little seasonal stream, directing the water to neighbouring fields. The village is situated in scrub jungle and according to Lawrie, was a new settlement from Sigiriya, with two and a half acres of paddy land, registered in 1878. There is no mention of the frequency of paddy cultivation at that time (Lawrie 1990 (1898): 170). The dispersed village, which is now typical of the new settlement areas based on large tank irrigation schemes and which is a recent development in Talkote, could not be treated simply as a modem phenomenon. Ashley De Vos mentions the type, in his article on rural architecture, as typical of the tradi­ tional rain-fed village (De Vos 1977:43-44). That is, a settle­ ment more dependent on direct precipitation than on stored water for its cultivation activities. In his description, this type of village has the houses located along the village path, or road, in large and scattered home garden plots, the paddy fields located close behind and the chena fields farther away. We might ask if the dispersed village pattern is the tradi­ tional form of new settlements in areas outside those intensive­ ly cultivated by irrigation, even before re-colonization started in the 19th century.5 Diyakapilla village might then be a modem example of a dispersed traditional village of this type in the survey area. With past experience of field survey, we can formulate the question: to what extent before the modern era was it at all possible for a central authority to keep pace with settlements of this kind? Probably only as long as locally based admi­ nistrators were ready to support the central authority. With fragile remains of a humble settlement and a very minor change in topographical features (although, if swidden cultivation was practised, there probably were significant cha­ nges in the original vegetation) it is difficult to trace these settlements in a field survey. We shall now discuss a tech­ nological complex that might fit into this settlement pattern and way of life. The third more intensively studied village in our survey area is Nagalavava. It had a population of 330 in 1981 (Wick­ remesekara 1990:164). The village is situated in a cultural landscape established by irrigated agriculture, in close pr­ oximity to a cave monastery. The villagers nowadays however are mostly dependent on swidden cultivation. Lawrie states that paddy was cultivated here only once in ten years, at the end of the last century. The villagers traditionally served as messengers and drum-beaters (Wickremesekara 1990:164, a- nd Lawrie 1990 (1898):171). That is, in contrast to Diyaka­ pilla, they have been closely related to the paddy-cultivating economy, but in contrast to Talkote, they are integrated into the social network as a serving community. A small part of Nagalavava village (seven households) still retains features of a traditional cluster village, and many of the traditional techniques of food preparation and storage are still practised here. One theme of study was traditional rural house building techniques, organization of space for various activi­ ties within the house and compound and the village as a whole, and its setting in the cultural landscape. The burial ground was considered part of the cultural landscape. Another theme was traditional medicine - the collection of material and techniques of preparation, as well as storage. Traditional measurements were documented, including the m- aking of various measuring devices. This characterization is to be viewed only as a help to focus three different rural possibilities. To cover other aspects of rural life as well, we went to six other villages in the area. Kayamvala village was visited to study traditional mz-oil pressing for household consumption. According to the 1988 village survey it is a new settlement, formed in 1953 by people from Sigiriya. In 1988 there were 304 residents (Wickre- mesekara 1990:163). Pidurangala village, with 247 inhabitants in 1981, was visited to discuss hunting, trapping and the collection of wild plants. A well-known elder, a former hunter, is now living there. Pidurangala village itself is mainly a paddy cultivating village. At the end of the 19th century it cultivated its fields once in three or four years. A monastery dating back to the 5th century is situated nearby (Wickremesekara 1990:165 and La­ wrie 1990 (1898): 171). Kosgaha-ala village, situated below Vannigamayavava, has no motorable road. It was visited to discuss questions of transport. According to our informants, the present village was established in the 1940’s-50’s by reoccupation of a former settlement site of unknown origin. It is not mentioned in La­ wrie and not discussed by Wickremesekara. The villagers sh­ are Vannigamayavava with Diyakapilla. Paddy cultivation is combined with cattle rearing, herding cattle from Pidurangala, commercially oriented swidden cultivation, some hunting and trapping and various minor entrepreneurial ventures. Ilukvava village, recently named Mahasengama, was visi­ ted to discuss the collection of wild plants with an elderly woman; and a specific trapping device with a swidden cul­ tivator. We were also able to visit a swidden field during the kurakkan harvest. In 1981, before the new village was fou­ nded, it had a population of 69. The village is not mentioned by Lawrie. The Village In­ quiry Unit states that paddy cultivation, swidden cultivation, hunting and to some extent cattle rearing, constitute the basis of subsistence (Wickremesekara 1990:163). In 1991 we got the information that the original gammandiya of the village had been abandoned in 1942. A new location was chosen, closer to Sigiriya village. In the 1980’s the new village of Mahasengama was founded by governmental initiative. The original gammandiya of Ilukvava is now used for swidden cultivation. It could, like the abandoned gammandiya of Tal­ kote, be used for ethno-archaeological studies. Udavalayagama was visited to study the aba (mustard) harvest in a swidden field. Lawrie states that paddy cultivation is infrequent and that the inhabitants suffer much from parangi (Lawrie 1990 (1898):173). The Village Inquiry Unit mentions that paddy cultivation is supplemented by chena cultivation. The village had 219 inhabitants in 1981 (Wicremesekara 1990:166). Digampataha village was visited to interview a Tamil fa­ mily of traditional stone-cutters (the only traditional stone-cut­ ters known in the area today) and the owner of hunting dogs. Lawrie mentions that the fields of the village, at the end of the 19th century, were "seldom sown, from want of water.". The Village Inquiry Unit states that hunting and swidden-subsis- tence cultivation constitute the most important means of li­ velihood in the village today; the tanks are utilized for fishing. The village had 898 inhabitants in 1981 (Lawrie 1990 (18- 98): 170; Wickremesekara 1990:162). Sigiriya village was visited to interview a stone-cutter who learned his skill by working for the Archaeological Depart­ ment, and who replaced the Digampataha stone-cutters as sup­ plier of grinding stones etc. in villages around Sigiriya. Lawrie describes the end-19th century situation: "The people suffer much from want of nourishing food and from parangi and aramana." (Lawrie 1990 (1898): 171). Sigiriya was not visited by the Village Inquiry Unit. We worked within a radius of 5km from Sigiriya rock. THE QUESTIONS AT ISSUE The household, a unit of reproduction. In the ethno-archaeological inquiry we wished to define the basic unit that would ensure the physical reproduction of rural society (here we use the word ‘reproduction’ in the restricted physical sense). By this we don’t just mean raising children, but acknowledging the importance of human labour in a non­ mechanized society, the support of every individual. This would, of course, also be a unit of consumption. The basic unit of reproduction was defined as the hou­ sehold.6 That is, people who live together and contribute pro­ ductive and reproductive labour to the unit’s survival. Human labour was seen in the context of the individual household. Our questions were: what does human labour in a household accomplish? By whom? When? Where? With what equipment? And for whom? To understand the whole production process of wet-rice cultivation, for example, the household in most cases is too small a unit. There are examples of a single household build­ ing a small seasonal weir in a seasonal stream, to release water to minor wet-rice fields; but even the smallest tanks (termed vav kotu by the villagers in our study area) had at least three families as builders and maintainers. At least three labour processes relating to wet-rice cultiva­ tion usually involve the exchange or hiring of labour, harvest­ ing, stacking and threshing. In most cases therefore, we could assume that wet-rice cultivation of one field directly involves several households. Also, regarding reproduction in its primary meaning of procreation, the individual household is, of course, dependant on a larger social unit. But to understand what constraints and material possi­ bilities exist in relation to non-mechanized wet-rice or swidden cultivation in the area, focus on the household as a unit of reproduction is rewarding. The household in an archaeological context The question of identifying a Lankan household in the ar­ chaeological context has not yet been discussed or attempted in the field. As a starting point therefore, we should take the modern social anthropological studies of rural Sri Lanka. What then constitutes a Lankan household today in the ‘traditional’ rural context? And if we test this pattern on an archaeological site, what traces would be significant? Nur Yalman has defined the basic unit in the Kandyan rural highland as the commensal unit, the nuclear family. He em­ phasizes the economic independence of the nuclear family - where food preparation, consumption and storage are or­ ganized by the wife and husband, and where each married adult is also a single unit in relation to the rewards of labour (Yalman 1967:100, 102). We can take this as our starting point, as it gives us two archaeologically visible features: the hearth and storage facilities. But what about the house as such? There are examples of a house being inhabited by several ‘commensal units’. In Nur Yalman’s study of a highland village, it is pointed out that newly formed couples may choose to live in either parental home during the first years of their union (Yalman 1967:104, 118-119). This is not to be confused with the fact that the couple could choose to work the land of either family - the wife’s (binna marriage) or husband’s (diga marriage). They need not live in the same house, or even in the same compound as the parents. The arrangement has significance where land-owner­ ship and inheritance are involved (Yalman 1967:126). E.R. Leach points out that in his study area, the young husband will pool labour with his father/father-in-law while he lives with them. He notes: "Individuals who lived under the same roof nearly always worked in the same team; but in­ dividuals who lived in the same compound in different houses very seldom worked in the same team." (this relates to the exchange labour teams that are formed) (Leach 1961:265). He further notes that in a diga marriage it is likely that there will be no separate storage of grain, though the couple might live in a separate house in the compound (Leach 1961:270). If they stay in the same house, we shall find two hearths and one granary there. The other possibility is that the son or son-in-law will receive grain on a share-cropping basis from the land still in possession of the householder, and keep it in a separate store (Leach 1961:270). Here the number of hearths within the house indicate a household with additional labour capacity. Each couple con­ sumes its own share, but from a common source of labour investment. Nur Yalman says of families not possessing land, that the different commensal units in a house might act independently in the production process and can have a quite different ec­ onomic status (Yalman 1967:119). In this case, the house as such would neither constitute the physical base for a unit of consumption, nor a unit of produc­ tion. He has not approached the question from the standpoint of reproduction however. It is possible that this arrangement could also liberate the young woman from certain child caring duties and hence give her a better chance of taking part in productive work outside - increasing the food supply. Even a house with separate consumption and production units (several hearths and granaries indicate this) might still have a common, enhanced production potential. This, of co­ urse, has implications from a reproduction point of view. It seems possible, although it should be studied in the actual Lankan context. In our study we also met an extended household of a different kind. Two aged parents were living with their son and his family. It was not a permanent arrangement, as they later moved to another son in the village. As in the case of the new young couple, they had a separate hearth, but they could con­ tribute very little to the productive labour of the household, although they were fed from the land their son cultivated. We could however follow the earlier argument and assume that this arrangement could have an overall effect on the hous­ ewife’s productive capacity. Unmarried, grown-up sons and daughters contributed to household labour, but they would not be visible in the ar­ chaeological record, unlike the newly married couple. As the size of a household is constantly changing with the family life cycle, it might be argued that the number of hearths and granaries in one house would have no significance from a structural, socio-economic point of view. The individual hearth is the feature that signifies the basic unit of reproduction. But, as noted above, several households living under one roof might stand in a different supportive relationship, compared to families living in different houses. Though we might not be able to draw any conclusions in terms of the labour capacity of ‘a house’ based on the number of hearths, these different possibilities should, however, be kept in mind. To sum up, it would be possible to find the modern unit of reproduction in an archaeological record. Focusing on a house, we could also distinguish between different reproduction units which support one another. The case of double hearths and granaries would indicate two households (separate consumption and probably separate production units), whereas two hearths and one granary would indicate two consumption units but one production unit (the case of the aged parents, for example). The latter should then be treated as one household (one unit of reproduction), as the basis of reproduction (productive work and its reward) is com­ mon. No direct conclusions could be drawn however regarding labour capacity or organization of labour, as the work teams for exchange labour, land ownership and share-cropping ar­ rangements are not directly visible in the archaeological ma­ terial, at least not at the single habitation site. It might be inferred from wealth studies of an entire village site, of house space and household goods. Food Procurement Labour Processes The various food procurement labour processes were divided into 11 groups, with a separate form drawn up for each of them. The groups were: irrigated low-land, home garden, swidden and permanent dry land cultivation; cattle, buffalo, goat and poultry farming; hunting and trapping, fishing and foraging for wild plants and other forest products. Added was a form concerning the use of buffaloes and bullocks as dr­ aught animals, focusing on ploughing, transport, threshing and drawing water; and another form focused on building, maintenance and management of irrigation facilities. The in­ tention was to document each stage in the labour process by means of the questionnaire. Agriculture Regarding agricultural work, the study defined each stage from the selection and preparation of the field, to long-term storage of the produce - as regards area of land, time of year, number of work participants, their age and sex, their relation to the household, work days per year and the tools and other devices used in the process. The location of threshing and storage was also sought. We discussed yield per unit area and the amount sown, as well as the total amount of long-term storage for consumption; and sowing and selling patterns in a two to three-year time frame. The reliability of interview data concerning these last mentioned questions is discussed below. Livestock tending Turning to livestock tending, our focus once again was on human labour in terms of location, work days etc. (as outlined above and equipment for herding and tending the animals in their daily and seasonally varying life cycles, as well as their individual life-cycle (from calf to milch cow to meat, for example); what labour goes into keeping them alive, extract­ ing work or produce from them, and their returns. As regards produce extracted, we recognized eggs, milk, manure, meat and hides. In the field, we found that animal horns were also used. We asked especially about further processing of the produce both for immediate consumption and preservation and storage. From kiri (milk) to di kiri (curd) or gitel (clarified butter from cow milk), or from fresh meat to sm­ oked meat hung above the fire, or kept in ash. Also what storage vessels or other devices were used in the process, and where they were kept. Hunting, trapping, fishing and foraging The questions were the same for hunting, trapping, fishing and foraging: which species were caught or collected, pr­ ocessed and stored - where, when, by whom and with what equipment. Documentation of tools and implements Our second line of study was the documentation of tools and implements and other devices used in the labour process. We focused on describing the physical properties of the item - that is, the material, size, weight and shape. We also sought answers to such questions as by whom it was produced, ap­ proximate cost, if bought, or time spent if made in the household, by whom and how often repaired, how old the given item and its general span of usage, also by whom the particular item is used and for what purpose. In fact, all those additional questions that will not be directly answered while studying archaeological material.One aim of this study was to get an approximate notion of the amount used of the various raw materials necessary for food procurement - for example, iron. At the present stage we consider our source material too scarce to allow for any ge­ neralized conclusions. To make a comparison between present day consumption and that of pre-modern settlements to be excavated, a com­ parative metallographic analysis of present day and archaeo- logically found material must also be undertaken. In focusing on a given item, we happened on some extra information: the karaka used for fishing was also used as a protective cage for the chickens at night; the pan katta could be used instead of the kurakkan katta in dry crop harvesting - although our informants didn’t think of referring to the pan katta when asked about tools used in the kurakkan harvest; a malu muttiya (large pot) that we documented gave rise to the information that such pots were previously used for storing dried fish mixed with ash, but that nowadays they were used for cooking food for almsgivings. This particular pot was broken and used for short-term storage of husked rice. This is a salutary exercise for archaeologists who think they can once and for all determine the detailed function of a given item. At the same time, the study underlined the fact that form and function are interrelated - you don’t carry water a long way from the well in a dish, if you have a choice; if you don’t have access to a kurakkan katta, it is the smaller pan katta, not the larger dakatta that you prefer to bring to the chena, for the ear-by-ear harvesting of kurakkan. And usages totally alien to the purpose for which the item was originally made could, probably, in the majority of cases, be classified as marginal - but more important, as indicators of changing con­ ditions of production. By chance we found the last surviving yota (water-drawing device) in the village of Talkote. It was functioning as a door to a dwelling house at night and as a see-saw for the children by day. Since the introduction of water from the large irrigation scheme, theyota hasn’t (yet) been necessary, and hence made useful in another context. It could be argued that this points to the regularity and logical coherence of social and material life, thus the possibility of interpreting and generalizing from an archaeological standpoint. Figure 16:3 Karaka (fishing device) used as a chicken coop at night. The birds range freely picking food during the day. Talkote village. Photo: Eva Myrdal-Runebjer. Documentation of the setting of the labour process Our third line of research was the site of the labour process. Labour here is to be understood both as reproductive work (as, for example, food processing) and productive work (such as ploughing). Therefore we recognized 10 major areas of study: the dwelling house with its compound and home garden, the ani­ mal pens, herding areas, irrigated fields, tank bunds and other irrigation facilities, chena (swidden) fields, hunting and forag­ ing grounds, water sources for fishing and stone quarries. None of the iron production sites has any parallel in the ‘live’ area. The purpose of this study was threefold. First, we wanted a notion of what traces the various activities could possibly leave in and on the ground; secondly, we sought to recognize the topographical setting of each activity, where any such pat­ tern emerged; third, we wished to get a more general under­ standing of what determined the location of the labour process, what was ecologically and topographically its optimal setting, so to say. It is obvious that the various labour processes will in­ fluence the natural surroundings in different degrees - from the construction of the tank and paddy fields, creating a new landscape, to the foraging of wild plants, which activity might not be traceable in the terrain. Many locations will be similar for different activities. Chena land, for example, is also used for grazing, foraging and trapping. It is also important to note that the material can’t be interpreted in a static way. Expanding irrigated agriculture, for example, will reduce the hunting ground. So the labour processes have to be seen in relation to other contemporary activities. Land use, activity areas and archaeological visibility During the field survey campaigns in 1988-1991, a complex pattern emerged as regards the location of settlements and land use through time. Present day Talkote has six at least partly functioning tanks within the present village boundary. The gammandiya site is not by the bund or below the tank, but beside a crossroad west of the main water course between the two largest tanks within the present boundaries. This might mark a settlement pattern answering to the inclusion of several tanks within the economy of the village. Pahala Talkote (mentioned by Lawrie as an abandoned village) west of Pahalavava was, according to the village inquiry 1988, the site of the village before the gamman- diya of yesterday (Lawrie 1990 (1898): 173 and Wic- kremesekara 1990:165). Figure 16:4 Seasonal land-use pattern. There is a complex pattern of changing land use through time, pointing to various possibilities of utilizing the environ­ ment. Within the Talkote-Pidurangala boundaries there now is a chena on an abandoned paddy field, in a former tank bed (the abandoned Halmillavava) in the command area of another tank still in use (the 20th century Pidurangala temple tank). There is also a chena on a former minor tank bund built on a previous settlement site (SO .47). The bund itself is made up of cultural layers from its earlier existence. The present temple tank bund is built on a previous settlement (SO .56), and also bears the residue of earlier living. The site of Pahala Talkote village, mentioned in Lawrie in 1878 as an abandoned settle­ ment, is now turned into chena land (Lawrie1990 (1898): 173). The various erections made for the religious superstructure (SO .30) sited on abandoned paddy fields below a tank bund, and the minor tanks themselves, some of which are still re­ membered in the oral tradition as being used by three to five families (as the Kotalahimbutugahavava) are other examples of a cultural landscape long in use and constantly changing. Though it is possible to arrive at a relative chronology for some of the sites, stratigraphies and accurate dating of the settlement sites, fields and tank bunds would be necessary for a detailed knowledge of how utilization of the environment has changed through time (for a thorough discussion of the Talkote settlement pattern see Mogren this volume). In this context we might consider Robert Knox’s descrip­ tion - probably from the kingdom of Kandy : Figure 16:5 Modern example ofchanging land use. Former Ilukvava gammandiya, abandoned in 1943. Now used as chena land for aba (mustard). Land known to be very fertile. Photo: Eva Myrdal-Runebjer. "And as I said before of their Cities, so I must of their Towns, That there are many of them here and there lie deso­ late, occasioned by their voluntary forsaking them, which they often do, in case many of them fall sick, and two or three die soon after one another: ...Whereupon they all leave their Town and go to another... Thus relinquishing both their Houses and Lands too. Yet afterwards,...some will sometimes come back and reassume their Lands again." (Knox 1981:353). A situation which would result in a very complicated set­ tlement pattern for the archaeologist to analyse from a field survey only! Focusing on the different areas where food procurement activities are carried out, we will discuss the archaeological visibility of those activities in field surveys and excavations. Irrigated paddy cultivation, livestock tending, fishing In villages like Talkote, where paddy cultivation today is, and according to our informants has traditionally been more do­ minant than swidden cultivation, we found livestock tending and fishing to be important associated activities. There are many descriptions of wet-rice cultivation prac­ tices in Sri Lanka, more or less detailed as regards cultural landscape, equipment, labour investment, yield and division of labour between the sexes.7 Therefore we shall only present a brief overview of the technological complex of rice-cattle-fish, with focus on the labour process as seen in the study area. Figure 16:6 Main seasons for various food procurement activities. They vary in time in relation to start of the rainy season and amount of precipitation. With the Mahavali irrigation scheme giving greater year round access to water in some tanks, the High Yielding Va­ rieties (HYV) of paddy totally dominant and irrigated cash­ crops like onions and chilli becoming more important, the yearly cycle is now changing. A difference between the field­ systems of owners with and without access to the new irriga­ tion facilities has also been created. Therefore the following description, based on interviews with two households and the three former vel vidanes of the village, on documentation of material equipment and by field-walking, is a reconstruction of past usages in a fast changing environment. In the modem period before the Mahavali irrigation sc­ heme was introduced, only locally based water sources were used and the local precipitation was crucial for all cultivation activities. This naturally determined the planning of the ag­ ricultural year. The most reliable cultivation season, according to our villagers, was the Maha season, starting in mid-October with the North-East monsoon. Preparation commenced in August, with the maintenance work of tanks and field canals. Ideally, this work was collec­ tively undertaken by the cultivators of the fields below the tank, each cultivating household contributing labour in relation to their cultivated acreage. Fields were cultivated by the in­ dividual household (with additional hired labour, or exchange labour) in their capacity as owners, share-croppers etc. Figure 16:7 Overview of the small-scale irrigation landscape in late August, showing almost dry modern Pidurangala Temple tank with adjoining paddy fields. Smoke from swidden field burning is seen to the left in background. Photo: Mats Mogren. In Talkote village (depending on the variety of paddy used and the timing of the monsoon) sowing in the Maha season was done sometime before end-November, harvesting in Fe­ bruary-March. If there was water left in the tank, the peasants could also cultivate during what they termed the Madakanne season, sowing in February and harvesting in May. The Yala Figure 16:8 Maintenance of the tank bund, late August, Talkote Pahalavava. Almost dry tank bed is seen to the left. Traditionally, each cultivator contributed work in relation to the amount of land he tilled below the tank: Photo: Mats Mogren. season (during the less rainy South-West monsoon) started in May, with harvesting in July or August, depending on the paddy used. As the agricultural year was wholly dependent on when the rains started and how much rainfall there was, the practices would vary much between different years. Hoes (udalla), and long-shafted bill-hooks (katta) were used by male household members for maintenance work on tank embankments and field canals. Hoes were also used to prepare field bunds and boundaries. Buffalo-drawn wooden ards with iron bills (nagula) were used for ploughing. Today the ard has been replaced by a light plough, with a mouldboard turning the soil. One farmer said the difference in field capability between the ard previously used and the present day plough, is that water is necessary to prepare the soil sufficient­ ly if using the ard, whereas the plough turns the soil, making water at this stage dispensable. According to him, it is possible to cover larger areas per day with the plough. This statement couldn’t be tested. Manuring (when practised) was conducted thrice at most, between and after the first and second ploughing, by a male household member. Cow or goat dung was used, carried out to the field in a basket. Buffaloes and hand-drawn levellers (po- ruva) were used before sowing, to even the surface, obtaining an even water level throughout the field. Figure 16:9 Ploughing of paddy field now watered by the Mahavali scheme. Talkote village, late October. Photo: Eva Myrdal-Runebjer. By making an opening in the field bund, water was sup­ plied through gravitation after the first and second ploughings and continuously after the seeds had sprouted. If water was scarce, a trough-shaped, swinging wooden device (yota) was used by the male cultivators in turn, at the end of the cultivat­ ing season, to draw water from the tank into the field canal and irrigate their individual fields. For 10-15 days after sowing, the whole family, or at least the children, went to the fields every morning and evening, to scare off seed-hungry birds by screaming and clapping etc. Weeding was done by hand, by the women, twice - first after one month, then after one month and ten days. Often labourers had to be hired for weeding a two-acre field, to supplement the household women’s workforce. The paddy field was watched every night, from late Dec­ ember till harvest, by a male of the household, who used crackers and dried coconut palm flares to scare away elephants and other animals. The watching of swiddens was simul­ taneously undertaken. See below. At paddy-flowering time in mid-January, the rice plants were often attacked by insects. Until about ten years back, a man of the household used to light oil lamps in a comer of the field, in the evening, to lure the insects away. This was con­ tinued for a week. The threshing ground (kamata) was prepared in uncul­ tivated highland close to the fields, in February-March, before harvest, by either men or women of the household.8 Paddy was harvested with sickles (dakatta) by men and women tog­ ether. Often harvesting required labour to be brought in from outside. As the peak harvesting season was common to all peasants, labour might have to be hired. Often however the work was carried out through mutual help, with relatives and neighbours.9 Bundling and carrying these bundles to the th­ reshing ground was done by men. After one day of drying, the harvest was threshed by buf­ faloes trampling the grain. They were led by men. This re­ quired both trained buffaloes and skilled ploughman. Not every household owned buffaloes. Additional labour is some­ times hired for this work. This may be a stage of work where power relations are manifest and re-established, although we could not document such aspects, given our few informants.10 The grain was stored in a closed container (vi bissa) built Figure 16:10 The last yota (water-drawing device) found in Talkote village. Owner now uses it as a door to his house. Photo: Eva Myrdal-Runebjer. of paddy straw, clay and cow-dung - provided sufficient yield was obtained to make building it worthwhile. Before cooking, the rice was hulled by the women of the household in a wooden mortar (vamgediya) and winnowed with a plaited winnower. In paddy cultivation, the majority of tasks were considered male work. Weeding was exclusively done by women. Scaring the birds, preparing the threshing ground and harvesting could be carried out by either men or women. Figure 16:11 Sickle with serrated edge used to harvest paddy. Talkote village. Photo: Eva Myrdal-Runebjer. Archaeological visibility of paddy cultivation The irrigation constructions themselves, in terms of earth bunds, are of course wet-rice indicators. There will, however, be a problem of relating a given construction to a given settlement if the topographical location of the site does not give a clear indication. To calculate the capacity of a tank during the settlement’s time of occupation (and thereby infer possible returns) will be a further difficulty, as the tank bund may have been enlarged. These aspects, as well as identifying the fields and determining their acreage, will be dealt with in greater detail in the coming report on the archaeology of irrigated agriculture. We noticed that rice is also grown without irrigation fa­ cilities. Paddy fields are sometimes introduced in low lying, flat areas, where water naturally collects during the rains, the bunds of the paddy field themselves acting as water storage devices. Ecofact remains from rice-plants (macro-fossils, pol­ len and phytoliths) therefore don’t by themselves indicate ir­ rigated agriculture. With better knowledge of the paddy field weeds, however, we might be able to tell from pollen and other ecofacts of a settlement site if the rice-plants co-existed with plants thriving in long term, waterlogged conditions - conditions that would not be present in a rain-fed field. The yield from rain-fed cultivation compared to irrigated agriculture should be discussed not only for a given acreage, but over time, as this is a form of cultivation even more vulnerable to local precipitation vagaries than the rain-fed tank. Bryce Ryan describes the technique used in field prepara­ tion by paddy cultivators in Bulupitya village, Uva Province, in the 1950’s. Here, no ard is used in field preparation. Earth is turned by hoes and the fields are ‘muddied’ by driving buf­faloes over the land, after flooding it. Here, paddy cultivation comes second in importance to chena cultivation, and the peasants are totally dependant on the local precipitation for irrigation (Ryan, Arulpragasam, Bibile 1955:155). No similar practice was reported from our study area. Macro-fossil paddy could, therefore, indicate various le­ vels of cultivation techniques. The ard-share (as compared to the hoe) would indicate the possibility of larger areas, as tilling is a faster process with ards than by turning the soil manually (Sathasivampillai 1967:198-199). The ard-share is a probable indicator of extended areas of paddy cultivation, when found in an archaeological context. As discussed below, we shall probably find ard-shares only in connection with ‘muddied’ fields. Iron sickles are less certain rice-indicators, as they could have been used for harvesting certain dry grains as well (see below, under swidden). Home gardens What was similar though, in most of the predominantly wet­ rice or swidden cultivating villages in our study area, were the individual home gardens with their wild edible plants, coconut palms and fruit trees. Various minor crops were also grown here. At any time of the year there would be something to pluck, or harvest, from the garden. The individual home garden is a recent phenomenon how­ ever in Talkote, which until 50 years ago was a cluster village with a common compound. In Nagalavava gammandiya there are still no individual home gardens. Livestock tending and archaeological visibility Poultry is kept to provide eggs and meat. The birds are sl­ aughtered by any male of the household, for festive occa­ sions. Hens and chickens receive very little attention in households which keep merely a few birds. A protected place is sometimes prepared for broody hens. Chickens are often protected under the karaka at night. A few, small, hen coops of palm leaves, built on wooden posts, are seen in Talkote. During the day the birds walk around freely, finding their own food. They are given water in the dry season. The picture is changing now, with richer peasants keeping somewhat larger runs. The osteological material left on a settlement site is pr­ obably the sole indicator of poultry keeping. The availability or not of water also determines the season of fishing and the use of cattle, and their maintenance. Fishing is discussed in a separate chapter (see below). Cows, buffaloes and goats face their most difficult time in July-September, when food and water are scarce. Usually co­ ws and buffaloes are not milked during this period. After they have calved, they can be milked regularly for four months, from October to May. They are milked by either male or female household members. The milk is further processed by the women. Goats are not milked. They are kept only to produce meat, and hide for the women’s New Year drums. They are slau­ ghtered at home, but often by someone in the neighbourhood who is known to be a keen hunter. Cattle flesh is also eaten, though the animals are not slaughtered at home. They are sold to traders and the meat is bought, if the household can afford it. According to the villagers, wild buffalo flesh is eaten if the animals, when caught and tied, die while struggling to free themselves. The animals graze in the dry tank bed, on harvested paddy fields (adding manure to the soil), along the roads and on the previous year’s swidden fields. The official truth is that the animals are herded, most often by boys. Yet in the village surroundings this is very seldom the case, at least not thr­ oughout the entire day. During the dry season, when they are not milked, they are often not even taken in for the night. Young and weak animals are sometimes given additional food during the driest season: paddy-straw preserved from last season’s harvest, or leafy branches, or freshly cut grass ga­ thered by the children, day by day. In the dry season it happens (less often nowadays, according to the villagers) that a male of the household takes the herd north-east, to the unoccupied area around Gallinda, where water and grazing grounds are more abundant. The open grasslands at Gallinda might be the result of previous human activities, an assumption strengthened by the presence of tanks in the area. Figure 16:12 Cattle-branding iron. Brands indicate ownership. Talkote village. Photo: Eva Myrdal-Runebjer. Food Procurement: Labour Processes and Environmental Setting We found, as a contemporary practice, that wealthy buffalo and cattle owners in Pidurangala send their cattle to ho­ useholds in the remote Kosgaha-ala village, during the dry season. The recipient household is responsible for the animals and given one cow buffalo per year in return. Today, ownership of cattle is shown by branding the an­ imals. The branding device is an iron rod (suruttu kolaya). The documented piece had a diameter of about 0.5cm, was 49cm long and bent into a half-circle at one end. No other special features associated with the practice could be documented. Even if similar devices were used previously and are still to be found among settlement site material, it would be difficult to identify them as branding devices. The wet-rice environment then, does not provide for ideal grazing during the cultivation season, for the animals have to be kept away from the fields and the tank is full. Also the livestock population seems too large - at least as it is presently organized - for a paddy cultivating village to support it in the dry season. What is obvious, however, is that animal power is much more important in wet-rice cultivation (for field preparation and threshing) than in swidden cultivation (where no animal power is utilized).11 Bullocks and buffaloes are also used for transport, as draught and pack animals. To infer cattle keeping the osteological material is probably the most important, though the animal pen should be identifi­ able in case larger surfaces are opened up during excavation. In the cattle pen there will be a considerable accumulation of dung, which might be possible to trace through phosphate analyses. The enclosure, built of wooden posts, might also leave post-hole traces as indicators. A community in settled conditions, using the animals also for milk, should have had an interest in keeping the animals penned, at times. Goats are kept in smaller enclosures. For example, in Di- gampataha village, the goats were seen occupying a hut built on a low, wooden platform raised on wooden posts. Here identification might be more difficult. Talking of the wet-rice landscape proper, there are few species available for collection, compared to the swidden field or jungle paths. Certain stalks and seeds and mushrooms are available in and around the tanks, mainly during the wet se­ ason: November to January. Concluding discussion. To sum up, wet rice cultivation requires a certain degree of coordination and cooperation between the cultivators below a given tank. Each field has its specific owner and cultivator (which need not be the same person). Animal power is impor­ tant in field preparation and threshing. More animals for draught-and-trampling power, milk, meat, manure and, to some extent, their horns and hides (goat) are kept in the predominantly wet-rice cultivating village, as compared to the subsistence-swidden dominated village. Figure 16:13 Largest cattle-pen in Talkote village, with brand-marked cattle; mid-September. Posts in background once carried a thatched roof. Previously, most of the dung was sold to Tamils from the north. Now thrown outside enclosure. Photo: Eva Myrdal-Runebjer. Additional labour has often to be hired for weeding, har­ vesting and threshing, tasks which are conducted in a con­ centrated period of time. On the very tentative data gathered in our interviews, we get an indication of a larger amount of man hours and a bigger yield area-wise in wet-rice cultivation, as compared to swidden cultivation. Important food-procurement activities are collective fish­ ing in the tanks during the dry season, followed by individual wet-season fishing in tanks and streams. Fishing probably will be seen through osteological material only. Hooks and the iron blade of the kastana patiya might have been used, however, in the pre-modem situation as well (see below, under fishing). Finally we shall observe that the combination of wet-rice cultivation with other water-related activities is not exclusive to Sri Lanka. The elaborate system of rice-fish-silk production practised in east China and south Japan for centuries, is a well known example of this.12 Chena cultivation, foraging, trapping, hunting, fishing Swidden cultivation in our study area obviously can’t be viewed as static, either spatially, through time, or as regards crops grown, or degree of subsistence orientation. The study area also has to be related to the overall Lankan picture. The main problem is that no (published) field studies have been carried out focusing on the history of swidden cultivation. Figure 16:14 One type of goatshed in Digampataha village. There are no walls; goats are tied in the shed. A construction with low archaeological visibility. Photo: Eva Myrdal-Runebjer. Regarding the environmental setting in an islandwide pe­ rspective, De Rosayro states in 1949 that in Sri Lanka, in contrast to other Asian countries, flat land in dry, mixed ev­ ergreen forest which is not under irrigation, is preferably ch­ osen for swidden cultivation, instead of hilly areas of wet evergreen forests (De Rosayro 1949:51). The lack of shifting cultivation in the up-country, however, developed during the 19th century, depending on the establishment of plantations as pointed out, for example, by Michel Gelbert (Gelbert 1988:38- 39). Regarding spatial and chronological development, we kn­ ow nothing of the beginning of swidden cultivation. As po­ inted out by R.A.L.H. Gunawardana, the first mention of dry grains on swiddens occurs in inscription material not before the 9th and 10th centuries (Gunawardana 1971:15). This, of course, reflects the interest of those who ordered the inscrip­ tions and is no indication of the beginning of swidden cultiva­ tion. Regarding crops and location, we can discern four stages of development based on written source material and today’s changing situation. One stage - which we don’t have to think of as the first - testified to inscriptions regarding taxation, which mentioned sugar cane, sesame and cotton as important swidden field crops, each the base of important cottage industries. We don’t have to think of them as the only crops grown on the swidden field. What the inscriptions tell us is only what crops were interesting from the taxation point of view. Then there is the possible expansion of swidden cultiva­ tion after the decline of the dry-zone civilization, as indicated by the Sinhala and Tamil swidden place-names analyzed by Michel Gelbert. The new swiddens were taken up in the wet zone, up-country, in areas of Anuradhapura, Madavacciya, Vavuniya, Padaviya, the Colombo-Puttalama-Kurunagala ‘C- oconut Triangle’, in areas from Trincomalee to Tirrukkovil and in the Pottuvila and Kataragama areas. According to this study, the establishment of swidden place-names should have started not before the 14th century and ended not later than the 18th, or early 19th century (Gelbert 1988:38-39). An important question: is this only a picture of expanding subsistence-swidden cultivation, or were the clearings also a means to open up the wet zone for economic expansion? Then we have the situation of yesterday and partly of today, of subsistence oriented swidden cultivation characte­ rized by household labour and consumption, combined with various additional food procurement activities. In the 17th century, Robert Knox mentions some of the dry grains on swidden land that we found during our field survey, and he describes cultivation practices that are very similar to what we find today. Michel Gelbert calculates from studies of aerial photog­ raphs that the swidden cultivators of today, producing solely for their own consumption, constitute no more than 2% of all swidden cultivators in Sri Lanka (Gelbert 1988:137). And finally today there is an increasing amount of cash­ crops grown on the swidden field. As in the case of wet-rice cultivation, subsistence-swidden cultivation is fast changing. More of the cultivator’s crops are grown for sale, and the traditional dry grains, such as kurakkan, are pushed back in favour of cash crops like tobacco and vegetables. In some cases the peasants in our study area declared that they have stopped growing kurakkan and similar grains altogether. This trend towards cash-cropping is not specific to our few studied villages. In 1983 the Department of Census and Statis­ tics pointed out in a report that swidden cultivation had ex­ panded throughout the Dry Zone, and swidden cultivators produced a growing amount of cash crops.13 Also, forest pr­ oducts in our study area, like honey, are now largely sold instead of being stored and consumed by the household. Land use and labour process in the study area An outline of the seasonal planning and associated ac­ tivities of subsistence-swidden cultivation will now be given, based on nine interviewed households, documentation of eq­ uipment and field-walking. However, the changing situation and the fact that the traditional, subsistence oriented swidden field also contained some cash-crops, as for example aba (mustard) must be kept in mind. The Maha season is the most important swidden season. One less important swidden season coincides with the Yala season in May-June. The sites of the Maha swidden field are selected in June-July. When asked what physical factors deter­ mined the given farmer’s selection of land for swidden cultiva­ tion, the answer was: depending on what was to be grown in the chena. Regarding dry grain such as kurakkan, our inform­ ants stated that dense jungle on a small hill was the best location. Flat lowland where water could collect during the rainy season was not desirable. The place should have lain fallow for at least five to eight years. A ten-year fallow is even better, but nowadays almost impossible to find. Michel Gelbert discusses the selection of chena land based on interviews carried out in three regions of Sri Lanka. We found that our informants looked in the first place for a su­ itable topographical setting, with the hydrological implications in focus. For Gelbert’s northernmost study area, which is north-west of our own, we find however that quality of soil and type of vegetation were the most important criteria for selection of a chena site (Gelbert 1988:80-83). The soil quality could, of course, also depend on the topographical setting. Often the new swidden site is selected while the male cultivator is doing other things in the jungle, like hunting with gun and dogs, or collecting honey. In our study area we found examples both of smaller swid­ den fields, cleared and cultivated by an individual household, and larger swiddens cleared by males from neighbouring households, or relatives, and subsequently divided into in­ dividual plots. Clearing starts in mid-July and the cut vegetation is then left to dry for about one month and fired for the first time around mid-August, by male household members (one or two from each household). Larger trees are left standing. After burning they generate fresh buds and leaves. From some sp­ ecies of trees these are plucked to be used as food, or as a beverage. Two or three days after the first firing, women (one or two from each household) pile up what was not burned at first and set fire to it. The fence is constructed during the first week of September, by male and female household members together. Each household is responsible for the fence bordering that part of the swidden field used by it. For clearing, the big, long- shafted billhook (loku katta) and the smaller, long-shafted bill­ hook (katta) are used. For fence construction, the hoe is used for digging and the axe and katta for cutting. Several techni­ ques are used in fence and hedge construction. If binding is necessary,‘wickers are used. Sowing and planting takes place from the first rains in mid-or-end October - or, in case swidden is combined with wet-rice cultivation, during the second week of November (after wet-field preparation is completed). Sometimes relatives exchange labour for sowing and planting. The top soil of the area to be sown/planted is loosened by a hoe. Males and females work together: men mainly hoeing, women mainly sowing. Figure 16:15 Chena field divided between different cultivators; seven types of seeds are simultaneosly broadcast. Diyakapilla village, late October. Photo: Eva Myrdal-Runebjer. tinues until horse gram is harvested at the end of March. Panicum millet is cut low, in bunches, with the ordinary sickle, as is swidden rice (which is grown in separate fields in other topographical settings) at the end of February, or early March. The other dry grains are cut ear by ear with a small sickle (kurakkan katta). Finger millet (kurakkan) is cut in mid­ February, sorghum (idal iringu) in late February or early March. Traditionally, a bowl-shaped basket (kude) made of kirival was used to carry the harvested ears. The one docu­ mented was made by the householder’s old father. A bigger variant was used in the kurakkan harvest that was studied. Harvesting with the kurakkan katta is women’s work. It is carried out by household members, sometimes with the help of near relatives who will, in turn, get help to harvest their own fields. Carrying the harvested produce from the swidden (sometimes after drying) is, in theory, male work. Usually only one harvest is taken from the swidden field. The fields are not weeded. Robert Knox witnessed the dry land cultivation of the 17th century. He mentions "Coracan’, "Tanna", "Moung", "Omb", "Minere", "Boumas" and "Tolla" growing on dry land (Knox 1981:112-113). The ear-by-ear cutting of certain grains by women harvesters was noticed by Knox: "The way of gather­ ing it (tana) when ripe, is that the Women (whose office it is) go and crop off the ears with their hands, and bring them home in a basket. They only take off the ears of Coracana also, but they being tough, are cut off with knives." (Knox 1981:112). Figure 16:16 Ear-by-ear harvesting of kurakkan. A1 1 /2 acre chenafield of mixed cropping. Ilukvava village close to Gallinda Mahavava, late February. Harvested ears are placed in a basket kept beside the harvester. Photo: Eva Myrdal-Runebjer. Davy says from his early 19th century horizon: "...reaping, or, when the straws are not saved, of gathering the heads of corn." (Davy 1983 (1821):201). Regarding the dispersed harvesting of swidden fields in the mid-19th century, Samuel Baker says: "Thus may be seen in a field of korrakan (a small grain), extensive coracana, Indian com, millet, and pumpkins, all growing together, and har­ vested as they respectively become ripe." (Baker 1983 (18- 55):35). Henry Parker describes both the implement and the cutting technique used at the end of the 19th century. He also notes women to be the harvesters in swidden cultivation: "The Si­ ckle, Dae-Kaetta, has two forms, a long-bladed one... and a diminutive one of similar shape which is only used for cutting off the heads of millet and other grains grown in the temporary clearings called Hena, or by Tamils Chena, this reaping being Figure 16:17 Basketful of kurakkan ears is carried to the field hut and emptied. It took 20 minutes for the harvester to fill the basket, which contained approximately 2kg of kurakkan ears. Ilukvava village, late February. Photo: Eva Myrdal-Runebjer. invariably performed among the Kandyans by the women alone." (Parker 1984 (1909):554) The ear-by-ear harvest makes it possible to make use of every single ear. During the kurakkan harvest we saw the women taking care to cut even the straw that had been tr­ ampled down. The dispersed harvesting season is noticed by Michel Gel- bert in the 1980’s. He states that harvesting is either done day by day, according to the needs of the family, or at one time. That this has implications for mobilization of labour is further noticed: "Grain has the disadvantage that if it is planted or sown simultaneously, the entire crop will mature and has to be harvested at the same time, thus requiring a big labour input within a short period." He further describes three harvesting techniques: hand-plucking of grain heads, hand-knife cutting of grain heads and sickle-cutting of stalks, observing all three techniques to be used for grains on swiddens in his three study areas. Tubers and roots, on the other hand, are dug out by crowbars, or sticks and hoes, if the soil is hard. Vegetables are cut, fruits and pulses are plucked (Gelbert 1988:159). This was stated to be the case in our study area as well. The swidden grains which are cut ear-by-ear are stored intact in an airy ‘storehouse’ built of palm leaves, on a wooden floor raised on wooden posts (atuva). They are threshed, ‘me­ al-wise’, in a wooden mortar (vamgediya) by the women of the household and ground on a rotary quern of stone (kurakkan gala). Vegetables are consumed directly after harvest, as are tu­ bers which are left in the ground until needed. Spices are dried and stored, if not sold. After the harvest, cows, buffaloes or goats are allowed to graze in the field. Compared to the wet-rice village, less open ground is usually available in the swidden village for animals to graze upon in the dry season. Fishing is carried out by men of the household, as in the mainly wet-rice village. On their way to the swidden field, or in the swidden itself, the women find a variety of wild edible roots, stalks, leaves, berries and fruits. Our informants stated that several wild plants will preferably grow in the chena field. Siriweera men­ tions at least one of these (tibbatu) as cultivated in the swidden field today and in ancient times.14 This might be an example of different levels of intensification in swidden cultivation, in different periods and areas. Approximately 70 different wild species, including four species of mushrooms that used to be gathered, were men­ tioned by our informants. Their local names were given in each case. However, several still remain to be botanically iden­ tified. They are collected for food, for making beverages and for medical purposes. Most of them are consumed on the spot (as berries) or within one or two days. Apart from those used for beverages or medical purposes, only a few species (‘vegetables’) are cut and dried and stored for about six or eight months. While men also pluck berries and fruits, the planned and everyday gathering is done by women. Figure 16:18 Discussing wild plants with an elderly woman from Ilukvava village. With increased cash-crop cultivation it is now mainly the older people who have knowledge of different wild plants, roots, fruits and mushrooms that can be used for food. Photo: Eva Myrdal-Runebjer. Another jungle related activity is collecting honey. The most important season is from May to September, which is the dry season. Beehives are found high up in trees and in cave sites in remote forest areas. Often three to four men (relatives or neighbours) go out together. They leave in the morning and come back before sunset. Trees are climbed with the aid of a foot-rope. Parker, quoting Nevill, mentions frail ladders made of cane being used in the late 19th century, to descend steep rocks, to reach the hives (Parker 1984 (1909):71). This was not mentioned by our informants. Clearing and cutting tools made of iron, such as axe blades, clearing knives with broad blades (vak pihiya), billhooks (katta) and small knives (ul pihiya, kopi pihiya) were used to hack branches and cut the honeycombs. For the latter purpose, whittled wooden sticks were also used. Earthen pots were used for long-term storage. All the other materials for honey collec­ tion are perishable: the big leaves folded around the honeycombs for transporting to the village, the coconut husk ‘sieve’ and the labu kataya (the hollow, dried bottle gourd) to keep the honey. Previously, honey was collected mainly for household consumption, but it is now collected mostly for sale, and the plastic sheets and glass bottles have made their entrance.There is no equipment of non-perishable material that is specific to honey collection. It could be investigated however whether honey could be traced chemically from storage jars. Swidden land from an archaeological perspective. What is interesting as regards dry zone cultivation is that, to my knowledge, we do not find ploughed dry land sown with dry grains-as seen, for example, in the interior of Tamil Nadu. Panabokke suggests this is determined by the texture of the soil, which according to him makes dry land plough agricul­ ture next to impossible in the Dry Zone of Sri Lanka (Panabokke, personal communication, October 1990). In Talkote however, we found that rice was sometimes sown on swidden fields in flat lowland, where water could collect. We were told that flat land minimized erosion. Flat fields were often turned into a kind of permanent ‘dry land field’ for paddy, enclosed by the usual paddy field bunds (see above, under wet-rice cultivation). The fields were manured by letting the cattle graze after harvesting. Paddy fields not linked to any tank were seen on flat land below small hills, during the field survey in the Kiri Oya valley, in 1988. Fields such as these can be ploughed after the rains have started, as in Tamil Nadu, but their water-logged state during the rainy season would probably rule out the growing of dry grains like kurakkan. The question whether dry grains found in archaeological material indicate swidden cul­ tivation on unploughed fields, needs further investigation. In a field survey situation however, we will find it difficult to determine a given settlement’s potential chena land, as swidden cultivation in this environment is very flexible re­ garding topographical requirements; and the distance from the homestead varies, depending on the availability of land and utilization of the surrounding land for other purposes. During our field survey we found most swidden fields to lie within 30-45 minutes’ walk from the village. Male ho­ usehold members lived on the chena to protect it from wild animals, not because it lay too far off. The rest of the family stayed in the village. This situation corresponds to what Gel- bert found in his northern study area (Gelbert 1988:90-93). This however is dependent on the overall land use situation, and we can’t count upon it as a static model for our study area. Is there, then, any possibility of detecting and dating pre­ modem chena cultivation apart from the indications found in the settlement site material? Any once inhabited area which is not covered by primary forest could have been used as swidden land. To test if datable layers of charcoal from swidden cultivation could be found, a small test-pit was dug in 1990. The test-pit was dug in traditional swidden land, within the traditional boundary of Ilukvava village. The area had been fired three years back and was now covered by bushes and small trees. We selected swidden land with a flat topography, so that not much erosion or deposits of soil would have af­ fected the soil profile. Figure 16:19 Test pit in fallow, flat chena field, in an area where swidden cultivation is traditionally conducted - to test archaeological visibility of chena cultivation. Ilukvava village, late August. Photo: Eva Myrdal-Runebjer. After digging 0.53m we still found the usual red-brown earth profile, but no layers of charcoal. Small particles of charcoal however, were seen dispersed in the soil all the way down. The conclusion was that where no soil deposits have taken place, the vegetational cover and the swidden technique (leav­ ing larger trees standing) don’t supply enough material to form a layer of charcoal, in the Lankan context. Unlike temperate regions such as Sweden, where the abandoned swidden is overgrown with a thick vegetational layer, the surface of dry zone swidden will remain exposed to forces of wind and water, dispersing the charcoal fragments that have formed. That the charcoal particles were dispersed throughout the pr­ ofile was tentatively interpreted as caused by biological forces (worms and roots). The technique of loosening the upper sur­ face with the hoe could also play a part, although it doesn’t explain the downward drift of the charcoal. Tank sediment was also suggested material for tracing the history of swidden cultivation on a larger scale. The idea was that erosion from the catchment area of a tank would form sedimentation layers in the tank, making it possible to identify and date distinct soot and charcoal layers which have formed. The study of tank sediment was taken up in 1991 and we found clearly distinguishable layers formed, some containing charcoal. The impression so far, however, is that this results from settlement activities above the tank.Cooperation with soil scientists of various disciplines is urgently needed to pursue this field of research. The selection of suitable test pit sites for sedimentation studies has to be further discussed: for instance, flat land below a hillock, where erosion deposits its layers above the swidden surface. Another possibility would be to study sedimentation layers in abandoned irrigation canals, for the presence of charcoal and catchment area vegetation residue, in the form of macro­ fossils, pollen and phytoliths. This will be discussed in future irrigation studies. Studies such as these would add information regarding the relative importance of swidden cultivation during the wet-rice based civilization and after its decline.15 Turning to what would indicate swidden cultivation on a settlement site, we have already mentioned the macro-fossils that might be found. Given poor preservation conditions, we should also consider using the indications provided by ar­ tefacts. Hoes, axes and bill-hooks are used also in wet-rice cultivation. Regarding harvesting implements however, we have found a specific implement, the kurakkan katta, which is used in the ear-by-ear harvesting of dry grains only. Whether this was in vogue also during pre-modern times is not certain. We found during the museum study that ar- chaeologically obtained harvesting implements were general­ ly smaller than most present day sickles, though not as small as the kurakkan katta, and without its distinct crescent shape. Figure 16:20 Kurakkan katta, diminutive sickle used for ear-by-ear harvesting of millet. Polonnaruva. Ethnographical collection, National Museum, Colombo. Photo: Eva Myrdal-Runebjer. This lack of evidence however, could stem from the fact that they were all obtained from urban centres and their surround­ ings, an environment that might not have favoured the ‘little of everything’ economy that is implied by the kurakkan katta (see below). Concluding discussion. To sum up, we find swidden cultivation in our study area of today/yesterday to be associated with other jungle oriented activities: trapping, hunting, foraging for wild plants and ho­ ney. None of these activities could be carried out as success­ fully in a predominantly wet-rice landscape. Traps to protect the swidden field, and using the swidden fields as ‘bait’ to trap the animals, have a distinct relation to swidden cultiva­ tion. This pattern is also presented in the topographical and anthropological literature of the 19th and 20th centuries (see under "Hunting” for further references). The utilization of forest products is further mentioned in mediaeval literature. The presentation of subsistence-swidden cultivation as form­ ing a technological complex connected to the other four ac­ tivities, is based on what was seen in the field - households growing mainly cash-crops on their swidden did not pay as much attention to foraging, for example. This seems easily explainable, since much of the collection of plants was linked to the women walking to and from the swidden field to harvest the various crops - that is, linked to polyculture. There was cooperation and collective work also among subsistence-swidden cultivators - swidden fields were cleared together by neighbours or relatives, traps were built in com­ mon, honey collected in the jungle in groups. Women some­ times assembled to go foraging for a special kind of wild plant. In contrast to wet-rice cultivators, among whom ownership, share-cropping or tenant status regarding a given piece of land determined with whom they would cooperate regarding timing of labour processes and sharing of produce, subsistences-wid- den work groups were the result of choice. What factors deter­ mined who chose to work with whom and why were not further investigated in this study. The other difference is also related to land: regarding wet­ rice, cultivation is undertaken in fields with a defined status of ownership or tenancy. We hear of sale, inheritance, share-crop­ ping etc. In subsistence-swidden, cultivation is carried out on government land which is considered no-man’s land by the cultivators. Subsistence-swidden cultivation with associated food pr­ ocurement techniques combines several activities on a small scale, "a little of everything" yearly cycle. The variety of crops grown is significant, as it permits a dispersed harvesting se­ ason which will, of course, also give a more secure return - it should be possible to harvest at least something even in a bad year. This has been noted by other field workers as well. Bryce Ryan quotes a villager from Bulupitiya, in Uva Province, saying: "...but with chena, if all else fails we will at least have yams." (Ryan, Arulpragasam, Bibile 1955:158). After clearing, burning, fencing and sowing, no work is carried out related to swidden crops apart from watching, which gives opportunities to men for trapping; and harvesting, which gives women the opportunity to collect wild plants. No labour has to be hired for any given stage of work. We should however note, as emphasized by several sc­ holars, the difference in maximum yield area-wise, as com­ pared to wet-rice cultivation.16 The dry crops from a swidden field will give any authority less to appropriate (or a cultivator to bring home) than a paddy field. As regards disposal of produce within the swidden village, we found that what the household produced was considered to belong to it (sharing, selling or consuming). In a joint under­ taking, the produce is shared among the participants. This seems to be in accordance with previous views. Siriweera states in his "Agriculture in Mediaeval Sri Lanka": "The per­ son who cleared the land and planted the crop got the pr­ oduce." (Siriweera 1978:44). Thus the land where swidden cultivation, hunting, trapping and foraging is conducted, is considered to be no-man’s land; and as the tools and devices used are owned by the user, or easily constructed by him alone, or in cooperation with others, there is no basis for any ‘share-cropping or tenant relations’ between subsistence-swidden cultivators. We should, however, note that many subsistence-swidden cultivators today, out of ‘subsistence necessity’, sell their la­ bour outside the village. This is so even regarding swidden cultivation itself. During the archaeological field survey, we met agricultural labourers from Udavalayagama working a swidden field for Pidurangala peasants, who paid them. This, according to Gelbert, seems to be a new but increas­ ingly common phenomenon: "Originally, wage labour was unknown in shifting cultivation. The extent of the plot which was to be cleared and cultivated depended on the available labour force in the cultivator’s family...For rich paddy cul­ tivators, mudalalis and other people, it has become profitable to employ labourers who cultivate cash crops in chenas for them. Therefore, in today’s chena cultivation the importance of wage labour is increasing." (Gelbert 1988:65). It is tempting, while studying the neat, swidden harvest- ing/hunting/foraging yearly cycle, to emphasize the possibility of self-sufficiency. And in a hunting/foraging community it is a possibility, as documented for the small group of "forest vaddas" in the 19th century. As cited above, Michel Gelbert estimated two per cent of swidden cultivators still to be exclusively subsistence-swidden cultivators. We must, however, consider the possibility that even long ago, from necessity, swidden cultivators might have sold or bartered their labour, or part of their produce. We should note that there are several forest products (knowledge of which is not easily acquired by an outsider) that easily enter a market and are sold or bartered - like honey and (nowadays) leopard skins. That this is not a recent phenomenon is indicated by M.B. Ariyapala, who states that honey is mentioned in the Puja- valiya as an article of trade, with other things like firewood, certain fruits and herbs. He further quotes the mediaeval text Saddharmalamkaraya, as saying that a man took herbs and fruits from the forest and sold them in the towns (Ariyapala 1968:334-335). Knox writes about the "Vaddahs" in the 17th century "Th­ ey kill Deer, and dry the Flesh over the fire, and the people of the Countrey come and buy it of them...Some of the tamer sort of these men are in a kind of Subjection to the King...they will acknowledge his Officers, and will bring to them Elephants- Teeth, and Honey, and Wax, and Deer Flesh: but the others in lieu thereof do give them near as much, in Arrows, Cloth &c. fearing lest they should otherwise appear no more." (Knox 1981:195). Davy mentions of the "village Weddahs" that "the bit of cloth they wear, and the iron heads of their arrows, they obtain by barter, receiving them in exchange for their dried venison, the skins of deer, or for honey and Wax." (Davy 1983 (1821):88). There is also a possibility that subsistence-swidden cultiva­ tion was combined with the cultivation of ‘cash crops’, as there is historical evidence, even before the decline of the wet-rice based civilization, that crops used for different cottage industries were grown on swiddens.17 Viewing subsistence-swidden/hunting/foraging solely fr­ om the point of view of returns, the important point to be made is that it can’t feed many non-productive people. Finally we observe that, for example, the practice of using the swidden field as bait for animals while at the same time protecting it from them, is nothing peculiar to Sri Lanka. Similar practices (and similar traps!) are reported in the 1940s from among the Lamet swidden peasants in present day north­ ern Laos, in south-east Asia; among the Finnish swidden rye cultivating settlers in 19th century Sweden, in north-west Eur­ ope; and among the Bembas of north-eastern Zambia, in sou­ th-central Africa, in the 1980s - to take some geographically and historically dispersed examples!18 So, the practice stems not from a culturally specific behavior, but from a given ma­ terial complication (animals eating what men cultivate) and given opportunities (what men cultivate lures animals into the man-made trap). Each of these swidden/hunting/foraging (Lamet and Bemba) or fishing (Finns) communities, of course, faced/face their specific, historically given situation of external relations. 1. Siriweera, W.I. 1986:1. 2. This has been noted in several recent studies. See for example Gelbert, M. 1988:151-154. See also Vite- bsky, P. 1984:72. 3. Lawrie, A.C. (1898) 1990:169-173. A similar situa­ tion is prevalent today in Dry Zone areas dependent on rainfed tanks. Regarding our study area, see Br- yde, A. 1986:91-93. 4. Wickremesekara, C. 1990:165. Talkote is also men­ tioned in Lawrie, A.C. 1990(1898): 173. 5. For a discussion of new settlements in such areas see Brow, J. 1978:159-194. For a discussion of the move from the traditional house-site area (gamgoda) see ibid:90-91. 6. This coincides with the definition of the basic work­ ing unit in agricultural production as defined by Sir­ iweera, W.I. 1978:46. 7. Regarding comtemporary times in our study area see especially Bryde, J.A. 1986. See also Brow, J. 1978 and Leach, E.R. 1961. For a summary of epigraphi­ cal and literary sources see especially Siriweera, W.I. 1978. 8. According to our informants the work related to pr­ eparing the threshing ground seems less elaborate in terms of labour organization than E.R. Leach des­ cribes in the North-Central Province in the 1950’s. See Leach 1961:271-285. This might however be a result of the method we were forced to rely on - that is, interviews - without the possibility of checking such information. 9. E.R. Leach gives a detailed description of how labo­ ur teams were formed in his study village in the North Central Province pointing to the fact that the mutual help system also was a means to establish and preserve power hierarchies in the village. See Leach 1961:280-283. Our ethno-archaeological study was not thorough enough to obtain information of this kind. It is however a most important aspect of the organization of labour. 10. See for example Gunasinghe 1990:117. 11. Gelbert also notes that few domestic animals are kept by swidden cultivating households in Sri Lanka: "But domesticated animals of a commerical value, poultry, pigs or cattle, are practically never kept." Gelbert 1988:151. 12.See for example Bray, F. 1989:132 and bibliography. 13. Department of Census and Statistics, Ministry of Pl­ an Implementation, 1983. Socio-Economic Indicato­ rs of Sri Lanka. Colombo:249-250. Cited from Bryde, J, A 1986:105. See also Gelbert, M. 1988:130-139, 221-223. 14. Siriweera 1978:43. The "straggler", thriving in hum­ an-made clearings but not actually planted there, not considered as a weed, but instead utilized by the cultivator is an interesting universal phenomenon, in my view related to small-scale subsistence oriented economies. The case of the stragglers that are not considered weeds is interesting also when thinking in terms of previous practices. In a Nordic context there is palaeobotanical evidence from the bronze and iron age for collection of seeds from many different wild plants that preferably grow on cultivated or fallow fields. See Helbaek, H. 1955:690-692. 15. For a discussion of the importance of swidden cul­ tivation during these periods, and its relation to the decline of the Rajarata civilization see for example Roberts, M. 1971:99-109. 16.See for example Gunawardana, R.A.L.H. 1971:17. 17. See the overview of historical records by Siriweera, W.1.1978:43 and Siriweera, W.1.1986 (1983):71-72. 18.See Izikowitz, K. 1951:182-183; Bromander, C.V. 1902: 270; Stromgaard, P. 1985:79.

EVA MYRDAL-RUNEBJER "Usually, facts are seen only in as much as there is place for them in the observer ’s mind - it is for this very simple reason that experienced observers see more things more quickly and more precisely than untrained ones!" (Francois Sigaut, Ecole des Hautes Etudes en Science, Paris).1 AN ETHNO-ARCHAEOLOGICAL PERSPECTIVE From where do we get our concepts to interpret the archa­ eological record as to function - material and social? The view on which this chapter is based is that we get it from our experience of the material world, from social practice, and from our human ability to identify, categorize and generalize our experiences. This doesn’t imply a narrow use of analogies based on similarities as to form, but a knowledge of the material re­ quirements of reality, a notion of what problems could be encountered in everyday life and how they may be solved. The term ethno-archaeology as used in this chapter refers to knowledge of present day material reality and current prac­ tice. This knowledge could be used to address the archaeologi­ cal record in Sigiriya from the point of view of concrete human experience, using data from the same area as the ar­ chaeological material is obtained. It is to be noted that other researchers have used the term to designate a different ap­ proach to archaeological material, focusing on other aspects of social life. It is important to underline that the concrete approach to the archaeological record, as argued here, has a long history, and that it is not possible to draw any humanly meaningful conclusions from archaeological material without relating to human experience of some kind, real or invented. The settlement archaeology research project addresses, a- mong other questions, the economic base - in technological, organizational, spatial and temporal terms - of premodem society in the Sigiriya region. This ethno-archaeological study was undertaken to gain knowledge of the study area, both in terms of the physical possibilities of basic human activities, the activities themselves in terms of technology, seasonal pla­ nning, labour investment and return, as well as their socio­ economic implications. Figure 15:1 Harvesting mustard in a poly-culture chena in Udavalayagama village. Photo: Eva Myrdal-Runebjer. With an ethno-archaeological perspective on field surveys, test excavations, archive and museum studies, and through our household-based field studies of contemporary villages, we wanted to get a notion of the spheres of activities that could be associated with a rural settlement, past and present. This in­ cluded the exercise of thinking ahead, from the present struc­ tures and activity areas to their ultimate destruction - that is, how they could be identified archaeologically; what structures, arte-and-ecofacts may survive and in what topographical set­ tings; what implications could be drawn as regards seasonal planning, labour investment, volume of production etc. This is not to say that the present socio-economic set-up is transferable to what can be archaeologically studied. The most basic of our hypotheses however was that the same laws of nature would apply; that is, the physical limits within which people could act would be possible to define in each case. The present study however, is to be seen as an initial survey of questions that could be addressed in an archaeologi­ cal settlement study, not as a report from a completed fiel­ dwork. As will become obvious from the following, neither our time in the field nor our personnel resources were adequate to answer all the relevant questions. As a complement to our concrete material approach, we had hoped to integrate a social anthropological study of land ownership and organization of labour, as seen from the village level. But without trained social anthropologists in the team, or time for field participating studies, it was impossible to achieve this. The team structure is discussed below. Here it should be added that the reason we could gain as much information as we did was that one of the team members was also a native of the area, living there himself and practising both irrigated and swidden cultivation, and therefore having a concrete practical experience of everyday rural life. After the following introductory discussion, the field studies will be presented in separate chapters. Documentation of rural material culture in Sri Lanka The documentation of rural material culture and technology is not well integrated in the Lankan academic sphere. The dif­ ference in field usage between the plough and the ard or the ard and the hoe is commonplace to any peasant, but it lies outside the academic discussion, much as is harvesting by ear and portion-wise pounding, in contrast to bunch-cutting, bundling and buffalo-trampled threshing of the entire harvest - technological complexes which we documented during past years’ fieldwork. The Anuradhapura Folk Museum, for ex­ ample, has an instructive exhibition of swidden cultivation equipment, but there is a lack of academic follow up in Sri Lanka of such initiatives. Food processing and storage techniques in the Lankan con­ text are other aspects of rural material culture that have been largely ignored by the humanities and social sciences. Food processing the world over is mostly the woman’s task, at least as regards foodstuff from the vegetable kingdom. Although it is essential to the community’s survival and obviously linked to the overall land use system, the seasonal and technological pattern of food procurement and storage, the time women could spend on other productive, or reproductive tasks, and to the socio-economic level of the individual household, few references to actual field documentation can be found. Apart from occasional notes by early travellers such as Robert Knox and John Davy, colonial administrators like Henry Parker and natural scientists-tumed-cultural historians such as Ananda Coomaraswamy and a few modern ethnog­ raphers, very little is published in this field of study (see under given headlines for references). The interesting Lankan initiative taken in the late 1970’s of holding a national symposium and later, a seminar, on Sri Lanka’s traditional rural culture, has not, as far as the present writer is aware, had any follow up during the 1980’s or 90’s. One important contribution to the knowledge of traditional domestic rural architecture today and the use of space within the dwelling, is made by Robert. D. and Bonnie G. Mac­ Dougall, in their study of selected households in a highland village (MacDougall 1977). Apart from documenting the stru­ cture itself, they have made a detailed record of the domestic articles and the activities in time and place of the rural hou­ sehold. Among previous ethno-archaeological studies in Sri La­ nka, Colin Kirk’s study of pottery production in Matale dis­ trict, in 1983, is probably the most extensive. Modern hand-made pottery was collected for this study and is now available at the Sigiriya Cultural Triangle office as reference material, together with drawings and photographs. The report has not been published. At present, Martha Prickett of IFS is organizing a nation­ wide study of clay sources, to be related to archaeological finds. This study is based on contacts with present day potters. One of the aims of the Department of Archaeology of the University of Peradeniya is to develop ethno-archaeological studies in general. As this is written, no publications are yet available.The few published ‘topographical’, ethnographical and ar­ chaeological studies related to rural material culture have, on the other hand, been appreciated by Lankan economic his­ torians pursuing a ‘new reading’ of the historical records, notably Leslie Gunawardana and W.I. Siriweera of the Un­ iversity of Peradeniya.2 FIELDS OF STUDY Five sub-projects were started, or continued in 1989, 1990, 1991 and 1992, in order to achieve a better understanding of rural archaeology, in addition to archaeological field explora­ tion and test excavation. An irrigation study and a botanical study were started in 1989 and continued in 1990, 1991 and 1992. The irrigation study is discussed in a separate chapter. The botanical study will be published in a separate volume. The study of museum collections of excavated iron tools and implements was started in the Anuradhapura Archaeologi­ cal Museum and National Museum, Colombo, in 1989. To a limited extent, the search for published reference material on Lankan food procurement activities continued during all four seasons. Household-based ethno-archaeological studies were started in 1989 and continued, with tentative explorations fur­ ther afield, in 1990. In 1991 and 1992 an effort was made to complete the limited studies that had been taken up. These are the studies discussed in this report. There are several important aspects of everyday rural life that have not been followed up in the fieldwork. A few of them are discussed below as possible fields of future study. Irrigation The irrigation study is reported on separately. The basic ques­ tions in focus were the investment and organization of labour, the potential of the given irrigation systems and their develop­ ment through time in these respects. This of course means that we, as archaeologists, have to approach various natural and applied sciences with specified questions. One aspect of the study was a field survey, documenting the physical properties of various irrigation facilities: delinea­ tion, length and height of tank bunds, present sluice technol­ ogy and remains of abandoned sluices, paddy field acreage in the command area of a tank, man made canals, or natural drainage from and to other tanks. Archaeological excavation techniques were used to docu­ ment tank bund constructions and obtain material for 14C dating. Another aspect of the study integrated with the ethno-ar­ chaeological village survey, focusing on labour investment and organization in the construction and maintenance of given irrigation facilities. This latter survey was continued by a separate unit between the 1989 and 1990 field campaigns. Figure 15:2 Ard-share with flanged upper end excavated from Anuradhapura. Provenance and dating not known. Anuradhapura Archaeological Museum. Photo: Eva Myrdal-Runebjer. Botanical survey The botanical survey is reported by T.R. Prematilleke in a separate volume. A list of cultivated and wild plants used in the households of the study area is included in this volume. The plants were listed as completely as possible in the ethno- archaeological field study. A study focusing on present and past uses of various plants in households of the area - where they are found, when and how they are harvested/gathered, prepared and stored - could help us to understand in what way given species found in an archaeological context could have been used.3 Knowledge of processing techniques, waste depositing and storage techni­ ques help us to identify the material archaeologically. This knowledge could, for example, give us important facts of the given community’s vulnerability in ‘bad years’.4 Study of museum collections In the Anuradhapura Archaeological Museum and the Na­ tional Museum in Colombo there are two major collections of excavated iron tools and implements The project obtained their kind permission to study both collections in 1989. The aim was (a) to investigate how well they were defined in terms of their use in time and space; (b) to document the implements through drawings, photographs and measurements, describe them as regards physical properties and systematize them in terms of function. Each implement was indexed and given a separate card, with a small photograph attached and with references to relevant literary sources. In this way we hoped to be able to build an easily available source of reference to be used by students as an introduction to the collections. Two days were spent in the Anuradhapura Archaeological Museum, documenting 50 items comprising all the agricul­ tural implements - bill-hooks, hoes, sickles and ard-shares; carpentry related tools such as axes (which could have been used for clearing also), adzes, hammers, nails and wedges; and the blacksmith’s small wedges. None of the implements could be said to be well defined in time of usage, but for most of them the provenance was given. The overall impression was that the artefacts were in a good state of preservation, though seemingly from early ex­ cavations. It wasn’t possible to obtain a chronology of the various items, as no datable stratigraphical references were given. Figure 15:3 Ard-share with diminutive flanges from present day Uva Province, Kalugahakadura. In form it resembles some of the "hoes” from the iron age/Early Historic site of Adichanallur, southern Tamil Nadu, India. Ethnographical collection of National Museum, Colombo. Photo: Eva Myrdal-Runebjer. Figure 15:4 View over Nagalavava cluster village. Wealth differentiation is seen from difference in building material and size of houses. Photo: Mats Mogren. Regarding the National Museum in Colombo, we expected at least 250 items, which were studied and partly published by R. Hadfield in 1912. Two and a half days were spent in the museum. Unfortunately only 45 items could be traced, so nothing definite can be said about the entire collection’s pot­ ential as a chronological reference for iron tools. Several of the items came from early excavations in Sig- iriya. Some of them were dated 5th century AD in the museum register. They were in a good state of preservation. Different types of ard-shares were noted, compared to the Anuradhapura collection. We intend to continue the study when the rest of the collection is available. Ethno-archaeological village survey In the ethno-archaeological village survey we addressed the questions of food-procurement techniques and activity areas, traditional rural housing and village layout, traditional meas­ urement systems, non-mechanized stone-cutting techniques and activity areas, and traditional medicine. Our interpretation focused mainly on the function of areas and artefacts, in terms of what, when, where, by whom and for whom. A series of questionnaires were designed to structure the interviews. From our ethno-archaeological perspective, we have foc­ used on the differences we documented (in land use, sea­ sonality, possibilities of complementary activities and related artefacts and structures etc) between irrigated wet-rice cultiva­ tion and subsistence-oriented swidden cultivation. We have defined these differences in terms of two technological com­ plexes. This might prove a useful starting point for future analyses of subsistence patterns of settlement sites. FUTURE STUDIES Pregnancy, childbirth and child care The basic prerequisite of the stratigraphies studied by ar­ chaeologists is, of course, the capability of past societies to raise children, feed them, socialize them and teach them the basic necessary skills to ensure the survival of yet another generation. This issue has various aspects. If we keep to the viewpoint of reproductive and productive activities, one important ques­ tion is how the cycle of pregnancy, childbirth and child care influences the life of women. Pregnancy and childbirth involve women’s physical par­ ticipation, and to fulfill these functions successfully certain activities have to be excluded. To a certain degree there are material constants involved in child care, such as breast-feed­ ing, mincing or chewing food for the child when it starts eating solids, protecting it from fire, deep water, wild beasts etc - but here the role of the mother will depend on her position in the prevailing society. We have heard of wet-nurses throughout history, for example. Figure 15:5 A young mother takes part in the kurakkan harvest. Her older children accompany her and take care of the 15-month-old baby. The work day is 4-5 hours. Two or three harvesters will complete the kurakkan harvest in this 1.5 acre field within two days. Our presence caused a bigger labour team to assemble, llukvava village, late February. Photo: Eva Myrdal-Runebjer. For future studies we wish to underline that: (a) there are material constants to be fulfilled to ensure the survival of children (b) these will require active involvement by the adult generation, but (c) women as mothers have a given material role in pregnancy and childbirth only. The work involved in ensuring the survival of children has been or­ ganized in various ways, depending on how society is organized and the social position of the parents. (d) whoever undertakes this task, her physical presence is necessary to a greater or lesser degree (dep­ ending on the age of the child). That is, it can’t be ‘technically developed’ to overcome the necessity for active human involvement and physical pre­sence. (e) the practical, social and intellectual training of chi­ ldren are aspects of child care that are socially determined and culturally defined. Household size and household wealth We did not study the relation between number of household members, residential area and storage facilities, to discuss archaeologically based inferences of household size and wealth.5 Note however the discussion regarding the unit of reproduction and area of rural houses below. Production for the household and specialized production We did not approach the question of how an ethno-ar­ chaeological study could help to distinguish specialized production from household production in the archaeological record.6 The issue comes into focus with the iron production study program. For pre-industrial iron production we will find no present day parallels, but there are blacksmiths in the area, and blacksmiths’ villages in districts further south, that could be studied for technology and debris accumulation. Communication and transport Questions of draught and pack animals used for transport, and the load-carrying capacity of animals and humans requires further study. The archaeological visibility of these activities also calls for more discussion. Formation of cultural layers The ethno-archaeological study can be developed to raise questions regarding the formation of cultural layers and the accumulation of archaeological material. What happens to the dwelling house floor, the compound and cattle pen when in use, and what factors act upon them after they have been abandoned?7 In future excavations of rural habitation sites, this ex­ perience could act as a cautionary tale for the excavator who is now dealing not only with stratigraphy to be identified in section drawings, but with underground surfaces, and fragile at that. And not least important, it could act as an idea-generating source when interpreting the material. This field of study seems to be under way in India, regard­ ing the palaeolithic and Chalcolithic cultures. Also, experimen­ tal excavations of newly abandoned sites have been conducted there. From a Lankan point of view, it is of course important to study the experiences of neighbouring countries like India,where some of the climatically or geologically given forces acting on the material might be common with the area under study in Sigiriya.Environmental history of the area The evolution of the landscape and environment at large is a related question. With studies such as these we enter a com­ plex field of soil science that has to be developed in coopera­ tion with geologists and paleobotanists.9 The approach to the material however, is similar to the ethno-archaeological.10 TEAM STRUCTURE The team structure was not the same during the four seasons. To be able to discuss achievements and setbacks the seasons will therefore be treated individually. In the first four days of the field season of 1989, the ethno-archaeological unit consisted of five or six people work­ ing eight hours a day. We were then reduced to three people working two hours in the afternoon for 11 days - that is, 54 hours, or about seven work days were spent in ethno-ar­ chaeological field studies in toto. Why was the field-unit reduced? Initially, the plan was to divide the team into smaller independent units that could work simultaneously with dif­ ferent sections of the household. It was very soon found that the idea of documenting rural life from this very detailed point of view was alien to the basic training the students got at the university. That is why independent sub-teams were out of the question. For the household to be visited by six people, of whom at least three had nothing to do but sit around, was of course disturbing, not to mention the waste of project resour­ces. Why then did we work only two hours in the afternoon? In this case, I wish to define the problem mainly in terms of language - that is, the writer’s lack of Sinhala and the students’ lack of English. In no undertakings involving interviews and documenta­ tion of human practices can it be satisfactory to communicate through an interpreter. But as it was the writer mainly who had to conduct the interviews, this turned out to be necessary. It was found that the only person who could help us to achieve at least something in the field, was the primary schoolteacher, a native of the village and attached to the project. He himself had practical knowledge, or at least hearsay knowledge of everything listed in the questionnaires, and therefore under­ stood the logic of the questions once the various terms had been translated back and forth. He was indispensable in our work, but could work with us only after school, from 1-3 pm. The security situation in the area and the aggressive elephants determined the end of our working day. In 1990 the core unit of two Lankan assistants and one Swedish consultant was increased by three Lankan members. We decided to extend our study to cover more questions than techniques related to food-procurement, although hunting, tra­ pping and the collection of wild plants and honey remained to be documented. It was decided that the team of 1989 should complete the food-procurement studies in the afternoons, wo­ rking from 1 till 3-4 pm. The half-day practice was again a result of the school teacher’s professional obligations and the Swedish consultant’s archaeological and irrigation field survey in the morning. The three new team members were to work in the addition­ al fields of study the whole day, helped in the morning by an assistant from the team of 1989. Traditional building techni­ ques and various areas of activity in the house, the courtyard, the traditional village layout and its topographical setting, con­ stituted one field of study, reported below by Damayanthi Gunawardana, Sudarshani Fernando and the writer. This study was continued by Sudarshani Fernando and Manjula R. Sir- isena in the field season 1992. They worked independently in the field, following the work plan from 1990.11 To be able to estimate yield and investment of labour as reported by the villagers, a study of traditional measurements of area, distance, volume and weight was undertaken by Pri- shanta Gunawardana and reported below. As part of the ethno- botanical program, a study of traditional medicine was started by Chandrika Jayasinghe, who also had the overall respon­ sibility of organizing these additional studies. The study of traditional medicine involved both field studies and reading manuscripts in different villages. A study of traditional stone­ cutting was started by Prishanta Gunawardana. These two lat­ ter studies are not completed. In 1990 we worked for seven weeks, but as four of the five Lankan team members had duties outside the project (three of them in Colombo) one to five days a week, the actual time spent in the field was considerably less. In 1991 the ethno-archaeological unit consisted of two members: the Swedish consultant and the assistant native of the area. Again, owing to other archaeological work and his own teaching obligations, the ethno-archaeological study was conducted for six weeks, between 4 and 6 pm on weekdays and in the field on a few Saturday mornings. The focus of study was trapping and fishing and traditional village boun­ daries. A field participating study of the aba and kurakkan harvest took place in February 1992. Figure 15:6 Documentation of labour processes. Sowing a chena field taken over by an individual cultivator; late October; Diyakapilla village. Photo: Eva Myrdal-Runebjer METHOD AND SOURCE-CRITICAL DISCUSSION The studies were built on three lines of investigation: labour processes, documentation of equipment related to the process, and documentation of its setting. The archaeological visibility of swidden cultivation was also discussed and a minor test excavation conducted. In all, we visited 28 households in relation to the food­ procurement and transport study, but 23 of these were only questioned regarding a few specific activities. Seven hou­ seholds were visited in the Nagalavava study of rural architec­ ture and village layout. In the study of agriculture and livestock farming, we had listed the various stages in the labour process beforehand, as we knew it from the literature and from practice. This was used as a ‘structuring’ background for the interviews. We al­ ways started by asking to be told about their work in their own words, starting with the beginning of the season; and we only interrupted in order to get the exact data of "where, by whom, for how long" etc, and if we found an anticipated stage miss­ ing, or if some not hitherto known stage of the work was mentioned. In this way the household could sometimes correct our questionnaire, for example, telling us that levelling of irrigated fields nowadays is done not once, but twice in the Sigiriya area. The limitations of the ethnographical analogy are obvious. What is materially given and what is historically specific in our ethnographical reference material has to be considered. An identification of ecofacts, such as paddy husks, wet-rice grains and kurakkan, combined with a knowledge of climate and natural surroundings, would provide a base for some con­ clusions regarding seasonal planning and land use. This is what most certainly is materially given. Knowing what kind of implements and other devices and structures were used would bring us a bit further in certainty regarding potential yield and seasonal planning, and a kno­ wledge of techniques used. Based on knowledge of yield pot­ ential in the given environmental setting, we could also discuss the maximum production capacity of a given area. This is materially given in the sense that we can test to what degree different techniques affect the seasonal cycle and potential yield. To infer past power relations on this base only, is not possible however. Viewing our work source critically, we can also see that we collected information at various levels of reliability. The capacity of the interviewer is, of course, always a factor, no matter how well selected the sources are, so the information obtained has to be judged by its internal consistency and in relation to what is known about the matter from other sources. The most reliable was the documentation of the actual implements and devices used for a labour process. The ans­ wers we got regarding when and by whom they were made, used and repaired, and for what purpose, we thought to be reliable, though they do not cover all possibilities, related as they were to a specific object. Next came our detailed questions about the labour process itself - what, where, when, with what and by whom. As long as their answers dealt with practices still in use, we could, by cross-checking information and viewing it in relation to the given whole, that is, the yearly cycle of work, constantly test the pattern emerging. When inconsistencies appeared, we cou­ ld go back and check again. Often references were made to past usage - for example, to former traditional varieties of paddy and their ripening time. This kind of information is specified in the field material as former practices, and could and should be checked with other sources. The most problematic and least reliable answers are those related to labour investment and return; that is, man-days per year, quantity of paddy harvested in a given area and number of animals caught in a given trap. The problem is not that it is difficult to obtain information regarding such questions, but for various reasons they involve a greater risk of subjective and misleading information. One obvious problem, when as­ king about time spent on a labour process, is that the informant usually has no reason to think in terms of hours when working for her/his own household, or pooling labour with relatives or neighbours. In the same manner, it is next to impossible for a villager to tell how much of a wild plant has been collected in one season, or how much time was spent on that. To a large extent, wild plants are gathered on the way to doing other things, along the path to the chena for example. These plants are neither weighed nor stored, but either immediately consumed (as with much of the berries) or eaten for dinner the same or the next day. Nevertheless, objectively, their collection represents both labour investment and nourishment, and we must find some way of even approximately measuring this. The answer is field participation studies. These could more easily be planned now that the preliminary fieldwork has given a frame of interrelated seasonal activities, to be studied in detail. Each household was documented on the age and sex of its members and their relationship to the householder. Also in­ dividual occupations. It was intended to document each hou­ sehold’s area of cultivated land, defined as regards type (irrigated paddy land, home garden, chena, highland for per­ manent cultivation, etc.) and nature of occupancy (jointly ow­ ned, singly owned, government owned, rented, leased or held on share-cropping basis). The size and age structure of the household and the land-holding pattern were included in the study, as these factors might directly influence food procure­ ment and house building activities and the kind of equipment used in the labour process. The various definitions of ownership, or tenancy, of cul­ tivated land, were taken from the socio-economic survey car­ ried out in the area by the Urban Development Authority in 1989. As mentioned above, we could not realize our intention of a parallel social anthropological study. Therefore only five of 28 visited households were documented regarding landh­ olding patterns. Most important for our study at this stage was the actual extent of the cultivated fields. There is one aspect of our study that constitutes a marked difference from sociological, or socio-economic surveys, apart from our emphasis on material culture and subsistence techni­ ques. Our study is ‘biased’ from the start, as we especially asked about practices now becoming obsolete, practices that are considered backward, searching for families that have not yet moved into the house of brick with corrugated sheet-metal roofs, families that still use the bullock cart instead of the motorcycle and tractor for transport. Most socio-economic surveys and many social anthro­ pological studies aim to explain the present situation to poli­ ticians and administrators, to guide their further actions. Our aim was different (though information regarding everyday life could, of course, be used in a variety of ways). We asked for the villagers’ participation in recording their own history. We sought their expert knowledge of rural practices, gained thr­ ough learning by doing. They have no reason in their everyday life to systematize it and analyze it diachronically. This must be the contribution of the academically trained student. Tog­ ether these two groups might be able to achieve what other­ wise they will not: an understanding of the base of the premodern society. As regards the structure of the questionnaires, we can con­ clude that they answered the actual stages in the labour process well, with minor adjustments carried out with the help of those we questioned. Neither the literature nor the academically trained team could have provided that information. This underlines the necessity of developing this area of research. We had, for example, been told that the people in Talkote kept female buffaloes (some in large herds) without ever using them for anything but, it was assumed, as a way of exhibiting the owner’s status. Coming to Talkote, we found that the female buffaloes as well as the cows were milked regularly, but that the dry sea­ son, culminating in August, was always a great strain on the animals. They often fell ill during this season and stopped producing milk, or did so in such small quantities that it had to be spared for the calves. When we visited the village, hardly any female buffalo or cow could be milked. They were said to be suffering from vasamgata and several newborn calves died, probably from undernourishment, during our stay. When it was possible to milk the animals, both buffalo and cow milk were highly esteemed and used for further processing in the household. A recent development however is that the milk now is mostly sold, instead of being consumed at home. Further, it had been assumed that the villagers wouldn’t eat beef. But we found that beef was greatly liked by them, but too expensive and couldn’t be afforded very often. The villagers don’t slaughter their animals themselves, but sell them to Mus­ lim traders and then buy the meat from the stalls, if they can afford it. Whether some kind of socially acknowledged restric­ tions circumscribe the slaughter of cows and buffaloes, or the reason for the practice of selling the living animal is mainly economical, is a field for future study. Hens and fish and wild animals may be killed by the householder, or some other male household member. Wild buffaloes sometimes die of exhaus-tion when caught, and their flesh is subsequently eaten.12 Lankan historians and archaeologists have looked into the history of beef eating and the practice of killing cows and buffaloes in the island. W.I. Siriweera, of the Department of History, University of Peradeniya, has made a survey of pub­ lished archaeological material and early literary sources con­ cerning the use of milk and meat from buffaloes and cows in pre-modem times. He concludes that there is evidence of the consumption of both. 13 Ponnampalam Ragupathy calls our attention to cattle bon­ es among the offerings found in the megalithic burials tenta­ tively dated between the 3rd century BCE and the dawn of the our era, in Anaikkottai, in the Jaffna peninsula. Bones of goats and sheep were also found (Raghupathy 1987:115-126,165). ACHIEVEMENTS AND SETBACKS After this discussion of intentions, we have to turn to a critical evaluation of implementation. In this context it is important to note that the settlement archaeology project has two aims. One is to produce new knowledge. The other is educational. In planning the ethno-archaeological study, both aspects had to be taken into consideration. We could discern two major interrelated problems during the field seasons. One, the lack of background knowledge of rural production processes and rural life in general, among a majority of the students; the second, the difficulties of com­ munication, due to language constraints. The more difficult problem was the lack of background knowledge of present day rural life and the lack of exposure to social anthropological source-critical discussion. To breed a source-critical approach to fieldwork and make the archa­ eological team familiar with the ideas behind the ethno-ar­ chaeological field study, was the main educational target during the fieldwork seasons. By dividing the team in 1990 and giving full responsibility to Lankan investigators, with the consultant in a true consult­ ing, but not field participating role, combined with regular discussions in the afternoons, we hoped to get away from the merely mechanical work of carrying out someone else’s ideas to carrying out one’s own, with a consciously thought-out aim. This meant, for example, that the Lankan team worked wit­ hout a structured questionnaire. The overall questions were discussed and decided upon together with the Lankan project leaders and the consultant. The fieldwork dealing with rural architecture and village layout was successfully completed in 1992, by two Lankan team members working independently. To put our newly obtained knowledge into practice, we arranged three evening seminars with the whole team, in 1990. One was concerned with traditional house building techniques and village layout. The team members discussed information obtained in the Nagalavava cluster village, in relation to what archaeologists had found in the large settlement in southern Talkote. Another dealt with food-procurement and storage techniques, viewed as a seasonal cycle and presented in terms of interrelated activities. This was discussed mainly in terms of archaeological visibility. Those two seminars were held almost entirely in Sinhala. The third, held partly in Sinhala and partly in English, was actually a presentation of the irrigation survey, but focused on what implications could be drawn, as regards function, from visible remains of the irrigation system. The discussion was, of course, based on what the ethno-archaeological team had documented.Our implementation however fell far short of our ambition. A more detailed documentation was started as regards imple­ ments, trapping equipment and techniques, fishing equipment, houses and courtyards, animal pens and tank bunds. Chena (swidden) fields and hunting grounds were discussed with the residents and visited by us Knowledge of the labour processes of swidden cultivation and wet-rice cultivation is based mainly on interviews. Fishing was discussed with villagers and noted in the sense that the tanks and streams had been delineated on the map in the field survey. Foraging was discussed too, but not studied in its natural setting, apart from the occasions when villagers sto­ pped to collect plants on our way to doing other things in the field. Looking back at the work carried out, it can be concluded that language constraints affected the quantity of work pos­ sible much more than the quality. The problem of not having enough time to spend in the field, with people constantly coming and going, is a practical problem, but again it affects the quantity of data gathered more than the quality. 1. Sigaut, F. 1988:24. 2. See for example Siriweera, W.I. 1982 and Siriweera, W.I. 1986. See also Gunawardana, R.A.L.H. 1983 and Gunawardana, R.A.L.H. 1984. 3. See also an interesting plant list included in Der- aniyagala’s thesis. Deraniyagala 1988:1412-1424. 4. See for example Messer, E. 1979. 5. See for example Kramer, C. 1979. 6. For a review and example of studies of specialization see for example Earle, T., Q. Costin and G. Russell 1986. It is to be noted that these authors have a different view that outlined above, as the control of production of objects of wealth (that is status-indicat­ ing objects) is regarded as the basis of control over the local production to the organization of agricul­ tural production, although the material he uses doesn’t cover the agricultural aspect. See M. Tosi 1984. 7. See for example M. B. Schiffer 1987 regarding the evaluation of accumulated archaeological material in relation to the layers and structures where found. See for example J. K. Stein 1987 regarding properties of the soil vital to the interpretation of cultural layers. 8. See K. Paddaya (1987:83) where "Ethnoarchaeolog- ical studies of the formation processes of present day settlements” (ibid:83) are asked for. Regarding ex­ perimental excavations of contemporary remains, see Bulletin of the Deccan College Research Institute, Pune. Regarding the use of ethnographical parallels for interpretation of archaeological material in an In­ dian context see Shinde, V. 1987 and Shinde, V. 1991 9. See Courty, M. A., P. Goldberg. R. Macphail 1990. 10. This "confrontation of the record with knowledge of present processes” is discussed in Gifford-Gonzalez, D. P. 1981. 11. Due to the sudden death of Damayanthi Gunawar­ dana during the 1991 field season the work had to be completed the following year. 12. Ambivalent views on the consumption of animal fle­ sh are, of course, observed all over the world. In an ethnographical record from the district of Varmland in Sweden 1918, for example, one informant is rep­ orted to have said "The meat of a horse was never made use of, and the poor would rather die of starva­ tion than eat horsemeat." (Keyland, N. Naringar 1: 307 in Nordiska Museet, Stockholm. Translation M- yrdal-Runebjer). The same informant also stated that previously the skinning of dead horses was held in such bad repute that no decent peasant would unter- take it. There is also an interesting account of how this last mentioned view began gradually changing in the early 20th century. Many Swedes today (includ­ ing the writer) will not talk or even think of themsel­ ves as eaters of horse meat, though most of us will, at times, ear "hamburger-meat", which is in fact horse meat processed beyond recognition. Is this hypocrisy a result of repression in the 10th-13th centuries by the Christian church, directed against the heathen practice of ritual slaughter of horses? Or is it simply that the horse in a large part of Sweden was neces­ sary for transport (and also became a close friend) until the first quarter of the 20th century? This ques­ tion has so far not been seriously studied. 13. According to Siriweera it is probably during a few hundred years from the end of the Anuradhapura period until the 14th century that the practice of not eating beef spread to the population at large. Os- teological material dated to sometime between 40- 0BCE-200CE suggests beef eating by people with access to the centre of Anuradhapura. From the 1st century CE people who were instrumental in the writing of what we consider literary sources of infor­ mation, argued against beef eating. During the late historical period there were people eating beef, but defined by high caste Sinhalese as low and unclean. See Ancient Ceylon. 2, December 1972 and Siriw­ eera, W.1.1982:5-16.

RAJ SOMADEVA INTRODUCTION Scattered in the 446.6 square miles which form the study area (Bandaranayake 1990:14) lie nearly 300 lithic records. The evidence from these inscriptions gives an insight into the development of the society which created them and so they cannot be ignored in a study of the Settlement Archaeology of the area. These inscriptions have served as vehicles of mass media at the time of their creation. Since these inscriptions containing information on edicts, ideas, prasastis, donations, etc. remain unchanged, they are a ready and reliable source of information for the present archaeologists to reconstruct the social history of a people of the past. The facts which these inscriptions reveal belong more to a literary than an archaeological context. So, on grasping the time and spatial context in which an inscription was made, the literary evidence it brings out gives a clue to understanding the material culture of its contemporary society and the social milieu that existed. All the inscriptions, which were found after extensive ex­ ploration, have been included in this study. There is no doubt there could be others which have not given up the secret of their location.1 This exploration was as wide as the limited resources of time and manpower could be stretched. The in­ scriptions found have been categorized according to chronol­ ogy and type (see fig. 13:2). The inscriptions of the area can be classified into two groups: those which express a complete idea, and those which are isolated symbols that cannot be interpreted. As both cate­ gories are the result of human activity and endeavour, they are discussed as a total collection and not as two distinct groups. The principal aim of this exploration was to record as large a collection of inscriptions as possible. Some of those col­ lected have already been read and published. Deciphering and publishing the new finds, identifying and mapping the loca­ tions and making an analytical study of all the epigraphical materials which were found within the study area followed. The inscriptions met the needs of a mass communication system, hence it was imperative that they should stand at places frequented by the people for whose benefit the informa­ tion (in the inscriptions) was meant, so that the purpose of the announcement was fulfilled. This does not follow that every such location was a settlement. These inscriptions may have stood at crossroads, fords, market places, entrepots or places where people gathered periodically, although they did not set­ tle in the immediate vicinity. A study of the location of the inscriptions from an archaeological point of view, shows clearly that they stood in close proximity to or at purana settlements, high roads and trade centres.2 The above principles can be applied very profitably to analyze the distribution pattern of the inscriptions in the reg­ ion. Although there is evidence of a connection between the geology of the region and this distribution pattern, it is also seen that there is a close correlation between the settlement pattern and the inscriptions themselves, quite apart from the geological factors. Distribution. There have been found only two inscriptions in the southern bank of the Kiri oya which falls within the easternmost corner of the region (see fig. 13:3). Of them, one dating to the 10th century AD is found close to the now abandoned Peikkulam tank; the other is of the 3rd century AD, located in the lower alluvial plain at Vavala (Manatunga 1990:82). This sector of the region has a topography which shows rocky outcrops that fall between 800 and 2200ft. contours. The landscape shows eroded rock remnants (Epitawatte 1990) which render the land unsuitable for Dry Zone swidden cultivation (Swan 1987:124-7). These geographical factors must have precluded settlement in this area, thus causing a very low population density, giving rise to no necessity for population movement, as the area is not very accessible and further travelling is impeded by the physical character of the land. In total con­ trast to the above, we see that the Sigiriya and Dambulla region has an abundance of inscriptions, dating to several phases of history; and so we can conjecture that this region must have been an area of continuous human occupation through many centuries. Archaeological sources other than these inscriptions give strong evidence that the Sigiriya-Dambulla region has been the site of intense cultural activity and was the settlement nuclei through long periods in history. Bandaranayake (1990: 21-32) also refers to the above factor. There are marked changes in the distribution pattern and density of inscriptions within each of the sub-phases of major divisions of history. Although the underlying reasons for this are unclear, it can be inferred that the expansion and contrac­ tion of regional land use and the fluctuation of regional oc­ cupational patterns must in no small way have influenced and pressurized them. In an overview of the total distribution within the region, it is possible to trace two patterns of locational setting: con­ centrations and isolated sites.3 Into the concentration pattern fall sites which were either centres of political importance or monastic centres, such as Dambulla, Sigiriya, Kaludiyapoku- na, Pidurangala, etc. Viyaulpota and Habarana stand as ex­ amples of isolated sites. Chronology Inscriptions dating between the 3rd century BC and the 18th century AD have been found distributed all over the island (EZ Vol. 1-6). Although there are variations in form and content, it is not difficult to trace the development and evolu­ tion of the inscriptions themselves, through these 21 cen­ turies. Epigraphists for the purpose of study have divided thebody of the island’s inscriptions according to their chronol­ ogy (Tennakoon 1957; Karunaratne 1984). To attempt to classify the inscriptions within a historical division of time, based on dynastic changes and the rise and fall of kingdoms, as is the system followed by the historians, is not theoretically sound or acceptable.4 On the contrary, dating the inscriptions must be founded on a close examination of their paleography, language and the evolution patterns that these fields have undergone. Definite explanations are difficult to find for either the longterm chan­ ges, or the sudden stylistic variations which occur in the script and usage. These cannot be relegated to a fixed time or space, but the cumulative effects of the elasticity, or rigidity, of the current foreign policies, invasions, migrations and immigra­ tions and the expansion and contraction of trade, must have played a subtle part in these changes. Epigraphists agree that the 3rd century BC to the 7th cen­ tury AD is the first major period of lithic record in Sri Lanka (Tennakoon 1957; Karunaratne 1984). Although there are ma­ ny paleographic changes within these ten centuries, the script still holds many of the basic characteristics of the original Brahmi letters. However for the purpose of the study, the thousand years between the 3rd century BC and the 7th century AD are divided into three phases: (1) 3rd century BC to 1st century AD - Early Brahmi phase; (2) 1st century AD to 5th century AD - Brahmi phase; (3) 5th century AD to 7th century AD - Later Brahmi phase. The 3rd century BC to 1st century AD is designated as the early Brahmi phase, as the oldest script known in the island dates to this period. The early Brahmi script, dating to the reign of the Mauryan emperor Asoka of India (268 BC) is used on the donatory inscriptions above the drip ledges of the rock shelters that were used as monastic dwellings, during this period.No considerable development or significant change in the script can be noticed during these 400 years.5 For this reason this period is considered here as the first or Early phase of the script. There is a clear development in the characters of the po- stlst century AD period. The extending of vertical strokes downward, a noticeable rounding leftwards, and the agglo­ meration of signs and marks without phonetic value, are char­ acteristics of the second, or Brahmi phase. Other scholars have made comparative studies of the Brahmi script of this phase and the contemporary scripts of Southern and Western India (Wickramesinghe 1904; Karunaratna 1984).6 In this study, the era is referred to as the Brahmi phase, on consideration of the evolution of the script and the fact that it retained its basic original forms all through its evolution. Further, this period is considered separately, since this phase marks the beginning of an unbroken evolution of the script, which continued into later phases. The script of the 5th - 7th centuries shows the beginning of the transition from Brahmi to Sinhala. Some of the letters are more rounded, and a slight evolution of the original Brahmi, which makes its roots almost untraceable, is a significant fact in the study of the characters. But some characters still con­ tinue to hold similarities to their root. For these reasons this timespan is referred to here as the Later Brahmi period. During the period between 7th century AD and 12th cen­ tury AD we find a very marked and rapid evolution of the script taking place. Since the first two centuries show only very slow changes, and the following two centuries evidence a rapid rate of change, these two sub-phases are considered as separate phases. Taken collectively, these five centuries can be considered the evolutionary stage from Brahmi to Sinhala, hence the period is referred to as the ‘Brahmi-Sinhala Transi­ tional Period’. The final stage of transformation, or the Proto­ Sinhala phase, occurs during the 12th and 15th centuries. Although the characters are not those of today’s Sinhala, they are remarkably akin to the script in use today. The post-15th century saw the completion of the transformation into Modem Sinhala. The chronology is based on the principles of division des­ cribed above (see fig. 14:2). A chart showing the density and dispersal of inscriptions within the region is shown (see fig. 14:4). Source Material Most of the inscriptions within this area have been read and published previously by eminent scholars. A large number of Early Brahmi inscriptions at Sigiriya, Dambulla, Ambulam- be, Kandalama and Digampataha have been published in ASCAR (1911-12) and the corpus of the Brahmi inscriptions of Ceylon by Paranavitana (1970). Several inscriptions were recently uncovered by the Cultural Triangle project team at Dambulla. (DE. 1988) Ranawella (1984) has documented so­ me inscriptions from the environs of Sigiriya. Very recently a hitherto unknown inscription was discovered at Unalugala by the exploration team of the current SARCP project.7 Of the six lithic records attributable between the 1st and 2nd cen­ turies AD that have been found in the study area, only two have been published. They are the Habarana rock inscription (Muller 1883) and the Dambulla rock inscription (Parana­ vitana 1983). The Nirauva, Beliyakanda and Kaluarachchig- ama rock inscriptions (Somadeva 1990) have also been recently deciphered. The slab inscription at Pidurangala, whi­ ch can be ascribed to the 6th or 7th century AD, and a pillar inscription from Egodavava, in the vicinity of Sigiriya, which can be ascribed to the 10th century AD, are being edited by the author. Only one inscription dated between the 10th and 12th cen­ turies AD has been found in this area. This is credited to be the work of King Nissankamalla (1187-1196 AD) (Wickramesin- ghe 1904). No inscriptions belonging to the period between the 12th and 15th centuries have been recorded. The only inscription that falls into the Modern period is that found at the Dambulla cave complex, said to be the work of Sri Vira Parak- rama Narendrasimha (1707 AD).8 The Early Brahmi Period Although there is no final consensus on the earliest date for the Brahmi inscriptions found in the rest of the island, it is possible to ascribe the earliest date for those within the pre­ sent study area definitely to the first half of the 2nd century BC. The basis for this is the Early Brahmi cave inscription at Dambulla, which says: "Devanapiya maharajhasa Gamini Tisasa mahalene agata anagata catu disa sagasa dine” ("The great cave of the great king Gamini Tissa, the friend of the gods (is given) to the sangha of the four quarters present and absent”). Paranavitana (1970:Lii) accounts Devanapiya Gamini referred to in this inscription to be Saddhatissa. In yet another inscription in the vicinity that reads: "Damarakita - teraha lene agata anagata cutu disa sagasa dine Gamini Abayasa rajhiyahi karite." - "The cave of the elder Dhamma- rakkhita (is given) to the sangha of the four quarters present and absent (the cave) fashioned in the reign of Gamini Ab- haya" - we find a Gamini Abay a referred to. This individual could well be Saddhatissa’s predecessor, his brother Dutu- gemunu (161-137 BC) (Paranavitana 1970:Lii).9 As with all the inscriptions of the period spread over the island, those of this area are also associated with rock shelters (Paranavitana 1970:xv). Hence the location of rock shelters has been one of the most salient factors that determined the distribution and dispersal patterns of inscriptions belonging to this period. Although every rock shelter does not bear an inscription (Adikari this volume) those that do are found at well drained locations free from water-logging and damp, but with easy access to a perennial water supply. The agglomeration of in­ scriptions of the Early Brahmi period, in the Mirisgoni Oya basin, Sigiri Oya basin and Dambulla Oya basin, highlights this basis for choice of shelters for occupation. The brevity of the texts is a common feature of inscriptions belonging to the period. Within the study area, as elsewhere, the inscriptions only testify to the donation of the cave. The donors’ and the recipients, who are invariably referred to as "agatha anagata chatudisa sagasa dine” - which translates: "to the members of the sangha of the four directions present and absent" - are not identified by name. Although this is the general rule, there are a few which omit one or the other of the points mentioned above (IC No. 859/861 and 865). Although these inscriptions are brief, the texts tell us directly or indirect­ ly of the political, social and economic conditions of the times. The naming of kings and provincial rulers in these don­ atory texts is a pointer to the fact that the one time rulers of the area had associated themselves zealously with the affairs of religion and supported the sangha in the spread of Buddhism.The Early Brahmi inscriptions of the study area show that the only figure of note in the contemporary political arena was Saddhatissa (137-119 BC). Further, the Ambulambe cave in­ scription (ASCAR 1911-12:121) speaks of a pachina raja whom the historians have recognized as a provincial ruler. According to Paranavitana’s interpretation, this inscription (IC lxiii) pachina raja (Skt: prachina raja) denotes hereditary or dynastic kingship.10 If Paranavitana’s interpretation (which he justifies by presenting the argument that the early Chola kings were referred to by the title muttaiyar, which by connotation is linguistically equivalent to pachina raja) is acceptable, it can be inferred that before political authority was centralized under a single canopy of kingship, there must have existed a plurality of authority under several independent rulers. Gunawardana (1985) has expressed these same views and also drawn attention to the fact that by an analysis of the Ambulambe inscriptions it is possible to trace the genealogy of a pachina raja. Such genealogies can be read in some of the inscriptions particularly of southern Sri Lanka.11 These in­ scriptions refer not only to a king-maharajah12 and provincial rulers, but also to a title called parumaka (Skt: pramukhaP). Many scholars have put forward various interpretations and theories around this term, but their central idea is that pa­ rumaka referred to an elite strata in the administrative system (Ellavala 1969; Paranavitana 1980). Fourteen parumaka are referred to in the inscriptions within the study area. The pa­ rumaka of the Kandalama cave inscription (IC no:860) has also been a totabojhaka (Skt: tritha bhojaka) that is to say, one who enjoyed the revenue from a ferry.14 It is remarkable that most of the references to parumaka are found in the inscrip­ tions of the Dambulla monastic cave complex. The gamika (Skt: gramika) - village elder or leader - and the gahapati (Skt: gruhapati) - householder - are named in several of them, thus showing that the concept of identity of family and village had been established in society by this time. This is further strengthened by the fact that one gamika and six gahapati are mentioned in the inscriptions in this area. Often included in these donatory inscriptions are terms of kinship, such as bariya (Skt: bharya) - wife - "Tisa ayaha bariya upasika rakiya lene” - ("The cave of female lay devotee Raki, wife of prince Tissa") (IC no.832), jita (Skt: duhitru) - daughter - "gamika Vahaba jhitaya parumaka Pushadeva jhaya(ya) lene upashika Sivaya lene” - ("The cave of lay devotee Siva, wife of pramukha Pussadeva, daughter of gramika Vasabha") (IC no.853), puta (Skt: putra) - son - "pachina raj ha puta Tisayaha lene” ("The cave of prince Tissa, son of the pachina king.") (IC no.831). The names of five gam - villages - also occur in the writing. (Seneviratne 1990). So we can infer that they must have been situated in close proximity. In his study of the Brahmi inscription, Paranavitana (1970) has commented on the different vocations and trades practised in Sri Lanka of the Early Historic period. The present research also picked out two such instances where vocations are men­ tioned. They are the dunuvaya (Skt: danugraha) - archer - of the Dambulla cave inscription (IC no.846) and ananika (Skt: arannyanika) - an officer in charge of irrigation works men­ tioned in the same inscription. One inscription refers to a bamana (Skt: brahmana) and several researchers show that Brahmanism had taken hold in society before the acceptance of Buddhism as a popular state religion.15 The etymology of personal names such as Nagadata (Skt: Nagadatta) (IC no.859), Tisaguta (Skt: Tisyagupta)(IC no- .852), Pusadeva (P: Pussadeva) could be evidence of the exist­ ence of non-Buddhist beliefs.16 If the name Naguli which appears in one of the Sigiriya cave inscriptions (IC no.869) is a derivation from the Skt: langula - ‘plough’- then it is not wrong to suppose that this must have been a society where agriculture was of great importance. These secular inscriptions are also overshadowed by the inscriptions referring to the clerical segment in contemporary society. According to tradition, Buddhist adherents fell into four categories: Bikshu - monks, Bikshuni - nuns, Upasaka - male lay devotee, Upasika - female lay devotee. The words tera (Skt: thera), bata (Skt: badantaf sadivihariya (P: Sad- dhiviharika^) are used to refer to bikshu or monk. The monk written of in the Kandalama cave inscription (IC no.860)) is given the added adjectival description of majimabanaka (Skt: majjhima bhanakaf The Bhanakax^ sect were members of the Buddhist clergy who maintained an oral tradition of the pre­ cepts and teachings of the Buddha. According to Adhikaram (1946) majjhimabhanaka was a sect that kept up the oral tradition of the majjhima nikaya of the Sutrapitaka.19 Within these inscriptions, which span a period of three centuries, is reflected the centripetal movement from dispersed pockets of political power towards a central authority. This seems to indicate that in the Early Historic and Middle His­ toric periods there was greater institutional and elite interest and administrative interaction in these areas; and therefore probably a greater demographic concentration. Brahmi period The content and form of the inscriptions of the 1st and 2nd centuries AD mark them out as a group distinct from the Early Brahmi records. The past tradition of cave inscriptions now gives way to rock inscriptions (IC vol.ii). It is difficult to determine reasons for this change. However, the abandoning of the caves by the monks, in preference of complexes built on the architectural patterns of the lay dwellings, must in some oblique way have influenced the change (Somadeva 1993). All the inscriptions of this period record donations to monasteries. The authorship, date and purpose of donation is very clearly indicated in these commemorative edicts. The inscriptions are invariably headed by a stanza of blessing.20 The authors always belonged to the elite of society, viz: kings, ministers and commanders of the army, etc 21 These inscriptions in the study area bring into focus the work of four kings who are the best known in the political administration of the island. The first of these is Vahaba, better known as Vasabha (67-111 AD) (Mv xxxv:59) who is also mentioned in the Kaluaracchigama slab inscription (Somadeva 1990). This founder of the Lambakarna dynasty22 and patron of agriculture, is eulogized as one of the greatest ad­ ministrators of the land. The second is Gajabahu 1 (114-136 AD) referred to in the Nirauva edict as Gamini Abhaya. Ac­ cording to the Mahavamsa (Mv xxxv:115) and the Rajavaliya (ed. 1976) of a later date, this great warrior king had for the first known time led a military expedition to South India. According to the Mahavamsa, King Mahasena (274-301 AD) the great champion of agriculture and irrigation, who in later times was deified as a rural god named Minneri deviyo by his grateful peasant subjects, was succeeded by Sri Meg- havarnabhaya (301-328 AD) who carried on the great work of his predecessor and holds a place among the great rulers of the island. The Vavala rock inscription (Manatunga 1990) is cre­ dited to mark the work of Sri Meghavarnabhaya, notably the historic event of bringing the Sacred Tooth relic of the Buddha to Sri Lanka (Mv xxxvii). The Tonigala inscription too, was meant to honour the work of this great king (EZ Vol.iii:172). The fourth great king is Kutakannabhaya, whose work is best known through the Dambulla rock inscription (Paran- avitana (ed.) 1983). He is referred to a nara isera (Skt: nara isvara), to mean ‘a mortal as powerful as God Ishvara’. These inscriptions reiterate what the chronicles reveal - that during this period there was a tendency towards centralization of political power in the island.23 What is clearly brought into focus through inscriptions of this period is the contemporary advance, with technological innovations in the field of agri­ cultural irrigation. The expansion of wet cultivation is marked, as several tanks are named in the inscriptions. The underlying causes for the vast development of irrigation, as compared with the earlier period, are clearly stated by some scholars (Gunawardana 1978). Evidence is found that, with the expansion of tank-irrigated agriculture, a simultaneous rise of revenue to the king’s coffers also took place. The institution of a dakapati or jakabaka (Skt: Udaka prapti) - water tax 24- from the users of the tanks as revenue, is also chronicled in the inscriptions of this period. Two other taxes were collected during this period: the mataramajibaka (Skt: matrika matsaya bhaga) - on fishing in waterways - and the bojiyapati (Skt: bhojaka prapti) - a levy on lands under cultivation. These are referred to in inscriptions in other areas of the island, but those of the study area are silent on this subject. Vapi hamika, or private owners of tanks, are mentioned in inscriptions outside the study region, but not spoken of in the inscriptions of the area under present research. But the inscriptions of this area describe the sale and purchase of tanks by private individuals. The Nirauva rock inscription speaks of a king’s son who purchased a tank for the price of 100,000 kahavanu and donated it to a monastery (Somadeva 1990). These inscriptions are also pointers to the beginning of the accumulation of wealth by monasteries.25 In timefc previous to the study period, the inscriptions enumerate cave shelters as the foremost donations of worldly possessions made to the monks; but this differs greatly during this period, when tanks and revenue from tanks were the gifts which kings bestowed on the monasteries. The uses to which the accruing revenue could be put were specifically stipulated in the donatory in­ scription, possibly to prevent the misuse or misappropriation of funds thus collected. The Nirauva inscription states that the donation was made for the use of the bikkus of the Dakshina Naga Pabbata Vih- ara. But we are unable to ascertain the nature of the donation, as one section of the inscription has been eroded. Many of the inscriptions outside the study area refer to various forms of Buddhist architecture, but the two inscriptions within the area only refer to the pabbata vihara complex. The introduction of new elements into the context of the inscriptions is a reflection of the rapid social and economic changes which had taken place in contemporary society. The reconciliation of the political disputes that marked the years immediately previous to the study period, which had seen the rise of the Lambakarna dynasty as the rulers of the island, the political stability that followed their ascendancy, the introduc­ tion of new technology in irrigation and the expansion of agriculture in its wake, must be accounted as factors of change which brought about a resurgence among the people. Water, the perennial problem of the Dry Zone, was solved to a great extent by the network of irrigation reservoirs. This caused considerable agricultural growth, surplus production and augmenting of revenue, which made donations of largesse to the monasteries a possibility. This stability that we interpret through the inscriptions must have been sustained by two principal factors: the personal magnetism and personality of Vasaba, Gajabahu I and Mahasen, scions of the Lambakarna dynasty, and their individual systems of wielding power, the icons of which they stamped on the society they ruled. The Later Brahmi Period The second Ijalf of the 5th century AD marks the beginning of the Later Brahmi Period, which extends until the mid-7th century AD. Historians refer to this as a period of political instability. The remarkable feature of this period is the dearth of inscriptions throughout the island. There is no definite reason that can be pinpointed as its cause, but one can infer that the political instability of the times must have indirectly affected the situation.27 There is a sameness in the content of the inscriptions of the Later Brahmi period. As all the edicts bear the word vaharala, they are referred to as the vaharala inscriptions. Much discussion has arisen and contrary ideas put forward on the purpose for which the edicts were written and their substance. The most valuable interpretations have been made by Paranavitana (1955). If his inferences are correct and acceptable, these edicts concern the manumission of monastic slaves. There are six vaharala edicts in the study area. One outstanding feature of the Later Brahmi inscriptions is that their authors have been commoners. This is evident by their content and format. The lettering of these edicts is not systematic, nor stylistically in­ scribed. The scribe seems to have neither known the finer points of the script nor the proportion and size of the letters. Tennekoon (1957) commenting says that these inscriptions are not of royal origin, their format is weak, the stone has not been adequately dressed and he hazards the inference that they must have been worked by incompetent and untalented artisans. The Transitional Period The inscriptions dated between the mid-7th century AD and end of the 13th century AD are categorized here as belonging to the Mediaeval period of inscriptions. They differ greatly in substance and format from those of both the Early Brahmi and Later Brahmi periods. The inscriptions of this period were usually incised on stone pillars about 6ft. in height. The pillars were generally four-sided, but octagonal shapes are not un­ common. 29 Many of these pillars carry lettering on four orthree faces. In most pillars the letters are incised within horizontal guidelines, but there are rare instances when the letters are incised in vertical columns, without any guidelines being marked. The Viyaulpota pillar stands unique among the inscriptions and still remains a conundrum to the epigraphist (EZ Vol.iv:176). Some of these pillars have ornamental capi­ tals in the shape of a punkalasa, which is the traditional sym­ bol for fertility and prosperity.30 The inscriptions found of this period can be discussed under two headings: the greater number dealing with royal immunity grants, and others that are legislative edicts. Both these categories are found within the study area. The word ‘attani’31 occurs in many of the pillars, so they are referred to as attani inscriptions. The texts of the attani inscriptions have been very carefully organized. The king on whose authority the donation was made is identified, the date of gifting, in whose presence the donation was witnessed,33 the beneficiary institution and the rules and conditions governing the donation are all recorded. All the edicts give a list of categories of people, materials and articles that are taboo within the donated property. Every attani edict concludes with a curse such as ‘may those who contravene this legislation be bom as dogs or crows’. In traditional Buddhist literature the dog and crow are thought of as animals who lead a very difficult existence. In no instance is the donor king identified by his personal name, but always by his regal title.34 The comprehensive list of titles of the officers that has come to light, through reading these inscriptions, shows that there existed clear cut legislative, executive and judicial administrative elements in the contem­ porary state organization.35 Land has been the most common donation and Buddhist monasteries, padhanaghara36 piriven37 nunneries and hospitals have been the beneficiaries. A donation to a single individual is rare. At the conclusion of the attani edict there is placed one or more stylized symbols of the sun, moon, etc. The symbol of a man with a plate in his hand, which is seen in the inscription at Polonnaruva is interpreted as representative of the preta called Matanga39 (EZ Vol.v:pl. 9). The above is a general description of the form and content of the attani inscriptions between the 7th and 10th centuries in the island. Six kings are identified through the inscriptions in the region. The majority of them belong to the Manavamma dy­ nasty, mentioned in the Culavamsa (Cu xlvii). The first of these is identified as Sena II (853-887 AD) by historians and as the Culavamsa tells us, he invaded the Pandyan provinces of South India and protected the island from "all calamities and dangers". His chief queen Sangha is said to be the builder of the Sangasenapabbata vihara that belongs to the Abhayagiri monastery. King Sena II is named as the author of the Kimbis- sa and Viyaulpota pillar inscriptions. Chronologically, Udaya II (887-898 AD) to whom is credited the Peikkulama pillar inscriptions, succeeded Sena II. The pillar inscription at Iripinniyava and Rambava, in the Anuradhapura District, have been traced to this king (EZ V- ol.i:163). The Culavamsa records that this king was so mag­ nanimous that he gifted 13 lakhs of kahavanu for social welfare. The largest number of inscriptions are credited to Sena IPs son, Kasyapa IV (898-914 AD). These include three in the Mapagala-Ramakale area and the pillar inscription presently in the Sigiriya archaeological museum, the provenance of which is not known. The chronicles speak of the generosity of Kasyapa IV. Among the many religious edifices he constru­ cted are the multi-storeyed building at the Abhayagiri mon­ astery, which is identified as Kasubradmahapa in the Jetavana inscription by Mahinda IV (956-972 AD); and the Kassap- asena monastery gifted to the Sagalika sect, referred to in the Bilibava pillar inscription as Kasubsensenevirad pirivena (EZ Vol.ii:38). A fragmented pillar inscription at Pidurangala speaks of Uda Mahapa as its author. Ranawella (1984) identifies him as Dappula IV (924-935 AD) of whom the Culavamsa says that he repelled a Chola invasion. Another personality we meet in the inscriptions is Sena III, who is said to have made many donations towards public welfare. The last of the kings mentioned in the inscriptions of the period is Kirti Sri Nissankamalla (1187-1196 AD) of the Kal- inga dynasty, known as the author of the Dambulla rock in­ scription, also mentioned in the graffiti on the Mirror Wall at Sigiriya, and whom the Culavamsa describes as the only king to gift Thulabhara. Other than these kings, nine royal personages are identifi­ able. Some of them are officers who were present at the dona­ tions. The others come into the category of persons prohibited from entering the premises donated. Another group of seven persons is named as being connected to the state because of their occupations. Many facts about the agricultural practices of the day can be gathered by sifting the information in these inscriptions. For example, two types of agriculture are discussed: wet rice cul­ tivation and shifting cultivation, referred to as chena today. Officers who were associated with wet rice cultivation are mentioned as Velbandi and Velkami. One of the Mapagala- Ramakale pillar inscriptions by Kasyapa IV refers to dolos sena (twelve chena). Not only paddy (Oryza sataiva) but Undu (Phaseolus nungo) was cultivated in the region at this time. Twelve great tanks are mentioned in an inscription at Ra- makale. It could be that these tanks were situated within the ‘Sigiri Bim’ referred to in the Viyaulpota pillar inscription (EZ Vol.iv:176). Two monasteries are named in several inscriptions: the Mahanapau Vehera, identified as Mahanagapabbata Vihara (ASCAR 1910-11) and the Saguna panhala of the Viyaulpota pillar inscription, as beneficiaries of largesse from Kasyapa IV. Legislative enactments pertaining to the monasteries also appear in the edicts. They lay down rules for the day to day conduct of the clerics and the manner which the wealth and property of the institutions should be managed. One clear fact that emerges is the expansion and amassing of wealth by the monastic establishment during this period. Post-fired graffiti Some inscriptions that are definitely not commemoration ta­ blets or edicts, are found in association with the ruined stupas in the study area. They do not have a complete text, nor are they similar to the inscriptions so far discussed, as they are post-fired graffiti. About 200 of these have been collected from Dambulla, Manikdena and Kaludiyapokuna. As much importance had not been attached to these post-fired graffiti, there is no record of the sites from which they were originally collected. Officers in whose keeping these artefacts lie today, say that to the best of their knowledge they had been collected from within the domes of ruined stupas. Such bricks have also been collected from areas outside the study region. It is remarkable that not a single complete brick with graf­ fiti has been found, so it is difficult to infer what purpose they must have served. The highest number of letters found so far on a single specimen is six. Although the letters vary in size, they are very clearly incised, in considerable depth. Pale- ographically, the script can be dated to the 7th and 8th cen­ turies AD. Content-wise, the post-fired graffiti can be discussed under three headings. Those which - 1. have a script that belongs to the general evolutionary pattern 2. show only geometric symbols 3. contain both script and symbols Because the finds from the first category do not have enough lettering, their content cannot be clearly determined. However, by connecting the few letters that are decipherable, the idea contained can be very loosely translated, as such words can be distinguished under two headings: those which are numerals expressed in words and parts of certain words. It is difficult to ascribe a reason for depositing within a stupa a brick which contains a reference to a numeral. According to Aryasingha (1983) by the 3rd century BC numerals were in use in inscriptions, so it is puzzling why numbers were spelt out in letters and the numerals not used in the 8th and 9th centuries. It would have been easier to inscribe numerical symbols rather than several letters on the brick, so it is of interest why the latter was resorted to rather than the simple numerical symbol. The preliminary problem to be solved is whether these bricks were inscribed by the masons with some symbol relat­ ing to their construction work; or whether they were symbolic of the Buddha; or whether they were offerings by people to gain merit for themselves, for a better life in their next birth. Those bricks which contained more than two letters were carefully examined and words such as vihara, vihira and oth­ ers of the combination of the sound vi/ha/ra were deciphered. The word vihatra was discernible on one fragment. On closer examination it seemed as though Sanskrit had been written in the Sinhala script, which was in a transitionary phase at that time. Ninety-one copper tablets inscribed in this same script, found at the Indikatusaya, at Mihintale, have been documented as having been excavated from within the relic chamber. Paranavitana (1931:199) has ascribed these inscriptions to st­ anzas from the Pamcavisati sahasrika pragna paramita, the well known Mahayana sutra. The term vihara occurs in tablets nos. 33, 36 and 62 of this collection. The fragments of bricks from the present study area do not give sufficient data to determine to which literary tradition they belong. The only clue we can follow is that these bricks, like the tablets at the Indikatusaya, contain a stanza in Sanskrit written in the Sinhala script of that time. This is further rein­ forced by the statement in the Culavamsa which says that Sigiriya was presented to the Mahayana bikkus of the Ab- hayagiriya monastery, after Kasyapa’s death. So, if this region came under the Mahayana sect, it is inevitable that the in­ fluence of Mahayanism would have penetrated the monas­ teries within and in close proximity to its centre. We see that the fragments of Bodhisatva figures found scattered at various places, especially at Kaludiyapokuna and Manikdena, show Mahayanist iconography. An inscription from Kaludiyapo­ kuna refers to Mete budun, the Maitriya Bodhisatva. Since all these give evidence of the fact that Mahayanism flourished in this period, we can assume that these bricks may have been votive tablets inscribed in the manner of the Mahayanist find­ ings at Mihintale. This hypothesis needs further investigation and study before any final conclusions are arrived at about the significance and place these bricks held within the religious ceremonials attending the building of a Buddhist stupa. The geometrical symbols that form the third type of bricks vary from simple straight lines to curves and angular forms. It is difficult to set up a theory on which to base a study of these symbols, as they do not give rise to any deductive inferences. A cursory overview of these symbols shows a close similarity to the Brahmi script of ten centuries previous. It is possible to think that these graffiti are only a deliberate attempt at copying the Brahmi symbols, with no knowledge of their symbolism or phonetic value. The crux of the problem is - what did the scribbler want to say? What was his message? The term ‘ma­ sons marks’ could well be applied to these, as is done in the case of isolated symbols of historic times. But this does not answer the riddle, for there are hundreds of symbols in a single location which differ one from the other, no two being alike. Yet another possibility is that they are votive tablets and that the markings held some significance for the devotee who in­ scribed it. This raises the question: did the inscribing of sym­ bols, or free etching on clay tablets, have a place in the rituals of religion at that time? If these marks were accepted symbols they would have been limited in number and repeated over and over again; but since this is not so, we ask - did random etching have a significant place in a ritual? Protohistoric symbols Several isolated graffiti symbols belonging to the megalithic period have been found here. Although many potsherds bear­ ing graffiti symbols have been found elsewhere (Seneviratne 1984:297) the first graffiti symbols carved on stone have occurred in this study area. Many interpretations and theories regarding megalithic g- raffiti symbols found in South Asia have been put forward. Paleography Although it would seem, at first glance, that a study of the paleography of inscriptions found in a very small area, within an island of comparatively small extent, would be a waste of time and give no conclusions of importance, this study has been able to assess how far the literacy and technology of the capital centre was of influence at the periphery of the power zone. Several regional variations of script have been picked out and this opened new avenues for further study in the field. These variations are at best corruptions of the original script. The cause for this probably lies in the ignorance of the in­ scriber who would not have clearly understood the shape and form and phonetic value of each symbol. The texts clearly show editing and the inclusion of omitted letters is noticeable in some instances. Summary As referred to above, the distribution patterns of the inscrip­ tions show marked changes from period to period. These changes are notably in the fluctuation of density. It is incor­ rect to dismiss these fluctuations as a natural process for which no explanation can be given; because a comparative study of epigraphic dating with contemporaneous events and the peaks and depressions of the micro-socio-economics of each period is mirrored in these ostensive stone sentinels. For example, fig. 13:5-13:8 show changes in the distribu­ tion of the inscriptions of 13 centuries - from the 3rd century BC to the 10th century AD. Two possible reasons for this are (1) the centralization of power; (2) the territoriality of the power centres. All through the annals of Sri Lanka’s history there has been an interdependence between the secular and the eccle­ siastical powers. As Rahula says (Rahula 1956) "the king protected the Sangha and the Sangha protected the king." That this interdependence for the welfare of each tended to give greater power to the monasteries, is evident in the Anurad- hapura period. The protection and social welfare of the sub­ jects, economic production and its distribution, were the responsibility of the secular partner. The ordering of the worldly lives of the subjects, in terms of their future spiritual well-being and education were the concern of the monastery. This mutual division of responsibilities characterized the ethos of the system, where an interdependence, often of uneasy suspicion but surface friendship, prevailed. On the basis of power thus generated by their partnership, the monasteries functioned not only at the urban political centres, but in far flung rural settlements too. A monastery was not necessarily the geographical centre of a settlement, but would remain outside the settlements, sometimes several miles away, and still serve as the religious centre for several settle­ ments. Fig. 13:6 depicts the distribution and density of inscrip­ tions in the period between the 1st and 4th centuries AD. A comparison between fig. 13:5 and fig. 13:6 shows the dif­ ference of distribution patterns. These changes could have been the result of expanding cultivable land and proliferation of settlements. New innovations in the socio-political sphere too, can be seen during this period. The rise of the Lambakama dynasty is a significant factor to which these changes can be traced. Historical and archaeological sources record the reigns of the Lambakama kings as being a period of intense agricultural production. It could be inferred that many large and small tanks must have been built in pursuance of this agricultural policy. Tanks are mentioned for the first time in the inscrip­ tions of this period. These irrigation and agricultural activities brought in their train a new system of taxation. A feature of the time was the extension of power of the clerics, by being the indirect taxing agent when the rulers made the monasteries the direct ben­ eficiaries and collectors of taxes from the tanks presented to them. This meant a concrete manifestation of the power the monasteries wielded. It also shows a diversification and bre­ aking up of the till then traditional concept of the king being Vastusvamy, or the ruler, symbolically the undisputed ben­ evolent despot. The aramaya now became the non-labouring source con­ sumer, whose main concern was developing production res­ ources through the settlements. With this development there came a spatial expansion of settlements, and the monasteries began to move out of the isolated hermitages in rock outcrops into more secular surroundings. At the same time the number of new monasteries also increased. The tanks and lands men­ tioned in the donatory inscriptions of the period may have been markers of the territorial limits of the monastery. That there are no donatory inscriptions dating from the 5th to 7th centuries AD is remarkable. This could reflect a break­ ing down of the relationship between the state, the monasteries and the common people. One of the earliest events on record to bring about this instability may have been the strained relationship between Kasyapa and the Mahavihara, and later, the aligning of Moggallana with the Mahayana sect. There was a resurgence of political administration during the 7th to 10th centuries AD. This marks the second phase ofthe rise of monastic power, which is reflected in the inscrip­ tions of the era. These inscriptions clarify the limits of the territories of the monasteries mentioned in them. One med­ ieval Buddhist record describing the erection of a pillar in­ scription, states the king caused a drum to be beaten at the aramaya, and a pillar was placed at the point where the drum­ beat was inaudible, and announced that the territory up to this point (from the aramaya) is owned by our temple". In this research of the inscriptions of the region, an attempt has been made to study in depth as much of the content as possible and give as brief a summary as possible. Often ar­ chaeological interpretations and historical points of view have been applied to interpret and give context to the texts of the inscriptions and place them in their chronological location in the cultural development of the island. A very clear fact that presents itself is that the distribution and density of these in­ scriptions changed, qualitatively and quantitatively, from time to time. A simplistic approach of trying to find a single reason for these changes will not give any satisfactory answer. The prob­ lem itself is complex and only a multi-dimensional approach will resolve it. The shifting of centres of government and the stability and instability of the state have directly and indirectly affected the content, form and distribution of the inscriptions. The best illustration of this is the prolific number of inscriptions during the 1st century AD and the 3rd century AD, when the Lam­ bakama dynasty was at its zenith; and in the period from the 7th century AD to the end of the 10th century AD, when the Manavamma dynasty held sway. These are periods which historians identify as periods of prosperity and political sta­ bility. In contrast, history shows the 5th century /J) - 7th century AD as a period of economic decay and political in­ stability. Inscriptions are few during this period. In the post- 14th century a change in the distribution pattern is marked by a spread into the Wet Zone and the central highlands. This reflects the geographical shift of the centre of power. A third factor: (the economic system) the rise of the Gross National Product and the capacity of the state to absorb the variations of production and distribution by subsidizing, through the dis­ tribution of largesse from the bounty of state land and support­ ing agricultural production, is seen in the texts, which speak of the gifting of tanks and land. Thus the economic factor also plays a part not to be ignored. The personality of the rulers, the forcefulness of their po­ wer, though not a compulsory element, played a very decisive role in the distribution pattern of these dumb, but very telling sentinels of time, which speak of the state and the elements that fashioned its form and character.1. Approximately a quarter of the study area is under thick forest. According to the villagers, a large num­ ber of ruins are present in these forests. It is possible that inscriptions can be found among them. 2. Codrington, in his article (1930) on the editing of several inscriptions belonging to the 12th century AD in the Southern Province, states that these must have been located along an ancient roadway. Sim­ ilarly Paranavitana (ed. 1963:16) states that an in­ scription found at Soraboravava, in the Uva Province had been placed at a market site. Thapar (ed.)1990:l- 5 discussing the Mauryan inscriptions of India, states that they were located at points where people often met or frequently passed. 3. By concentration is meant the occurrence of at least three or more inscriptions belonging to a single period, eg: Dambulla, Pidurangala and Sigiriya. Ex­ amples of isolation are Magallena, Viyaulpota and Habarana. 4. Earlier classifications have been based on language and paleography. For the purpose of this article, cl­ assification is on the basis of a deeper paleographical study. 5. Minor changes in the form and shape of characters are seen in the inscriptions of the 3rd century BC - 1st century AD. These cannot be accepted as a major development of the script, but rather as accidents, due to ignorance of the phonetic values of each char­ acter, or unskilled workmanship. 6. There are marked changes in the character of the script from the 1st century AD to the first half of the 5th century AD. Often these changes have brought about a similarity to the script of the Brahmi inscrip­ tions of South India. 7. Personal communication with Mats Mogren. 8. Vira Parakrama Narendrasimha (1707 AD) is one of the last of the royal line that ruled the Kandyan King­ dom (Abayasinghe, Dewaraga and Somaratne 197- 7:65). 9. Several kings of the Early Historic period have used the title Gamini Abhaya, so it is difficult to deter­ mine which particular king is referred to here. The only evidence by which we recognize an individual mentioned is by a date, or genealogical fact noted in an inscription. W.Prachina has several meanings in Sanskrit, "the ea­ st", "ancient", are the most common. If this meaningis accepted, pachina raja could be "the king from the east", or an "ancient king". Paranavitana (1970 ) has pointed out that the former translation is inapt. 11. Paranavitana shows that the genealogy of the dasabe rajun (the ten brother kings) mentioned in the Ma­ havamsa, is established by the inscriptions. Several inscriptions on this subject have been found in the south and south-west of the island (see. Motteyakallu cave inscription, ASCAR 1950:28pp). 12. The term maharaja has been used in inscriptions to convey an explicit meaning. It signifies the consec­ rated authority in the kingdom. This term is used only for the legitimate head of the state. A yuvaraja (sub-king) or a regional lord was never referred to by this term. Therefore this differs from the maharaja concept of the Indian sub-continent. 13. The literary meaning of pramukha (Skt.) is ‘forem­ ost’. The Tamil equivalent is parumukan. AA.Bhojaka (Skt.) means ‘consumer, the one who enjoys the benefits, the collector of taxes, the administrative head of a region’. 15. The Mahavamsa tika describes sotthisala as preach­ ing halls used by Brahmins. The Mahavamsa refers to a Brahmin called Pandula as Pandukabhaya’s tea­ cher. The Oruvala sannasa (15th century) mentions several Brahmins. (EZ Vol.iii:51). 16. The personal names mentioned in the inscriptions show that belief in astrology was common at this time. Names such as Tissagupta mean ‘protected by the Tissa nakata'. A nakata is an astrologically aus­ picious time. Pussadeva means a personalization of the Pussa nakata. This practice of incorporating astr­ ological terms into personal names shows that astrol­ ogy held an important place in people’s lives and beliefs. 17. According to the Buddhist Vmaya, a novice had to live under the guidance of his teacher (acharya) and preceptor/mentor (upadyaya). Other novices are cal­ led Saddhiviharika. 18. According to the Mahavamsa, a written text of the oral tradition of the Buddhist canon was compiled during the reign of King Valagamba (89-77 BC). The bikkus who carried on the oral tradition are ref­ erred to as bhanaka. In the vedic tradition uchcha- rana (pronunciation), and smaranaya (memorising) were the methods of learning.19. The Buddhist canon comprises two sections: the Vinaya pitakaya and Sutra pitakaya. The Sutra pitaka contains the doctrinal teachings and the Vin­ aya pitaka the disciplinary rules for the sangha (pr­ iesthood). Later, the Abhidharma pitaka was compiled. The Sutra pitaka is sub-divided into the Diga nikaya, Majjima nikaya, Anguttara nikaya, and Kuddaka nikaya. Those bikkus who carried on the oral tradition of the Majjima nikaya were called the Majjima bhanaka school. 20. The use of an auspicious word at the beginning of an inscription came into practice in the beginning of the 1st century A.D. Siddam and Svasti (Skt.) meaning ‘May you be blessed’ were the most common words used. Often only Si was used as an abbreviation. 21. Senevitissa in the Beliyakanda slab inscription is one such individual. 22. According to the Mahavamsa, the period from Va- saba to Kitsiri Mevan belonged to the Lambakarna dynasty. This was the second dynasty to rule Sri Lanka. 23. Several historians have shown that by the 3rd cen­ tury BC political power had been decentralized. The terms parumaka, gamika, pachina raja, dasa bat- hikana, etc. that occur in the Early Bhrami inscrip­ tions refer to individuals who administered the decentralized power centres. 24. Prapti and bhaga (Skt.) connote ‘a share’. Generaly this term in Sri Lanka’s inscriptions refers to ‘taxes’. 25. Gunawardana (1979) has shown the tendancy to secularization of the monasteries in the post-1st cen­ tury BC. 26. In the inscriptions of the 1st century AD to the 3rd century AD the monasteries are usually referred to as pabbata vihara. This is the Pali derivation from par- vata vihara (Skt.). This literally means a ‘vihara in the rock’. The Mahaparanirvana sutra refers to a place situated in a rock outcrop close to Rajagaha city, in India, as pabbata vihara. According to Bas- nayake (1983) g/r/ vihara also refers to the pabbata vihara. Bandaranayake (1974) has put forward ano­ ther definition of the pabbata vihara. Tl. Between the fall of Kasyapa and the consecration of Manavamma as king, a period of political instability prevailed. 28. During the past three decades there has been much controversy among scholars over the meaning of the 28. During the past three decades there has been much controversy among scholars over the meaning of the word vaharala. The most acceptable so far has been put forward by Paranavitana (1955). 29. The naming of the text and its layout were important considerations in shaping and preparing the stone for engraving. For example, see Ranava pillar inscrip­ tion. 30. The term punkalasa is a derivation from (Skt) pur- naghata. It means a filled pot. It is possible to in- tepret it as a pot filled with water, milk or grain. Today, a coconut (Cocus nulifera) flower placed in a new clay pot signifies the punkalasa. A 15 century poem describes a punkalasa as being a symbol of prosperity. " sadava sudu vali pivituru bandava ran dada visituru situva dora dora rambaturu tabava pun kumbu visituru " "(the city is decorated) strewn with clean white sand (roadways), hung are beautiful gold flags, banana trees planted at every doorway and placed are or­ namental pots to bring prosperity”, (punkumbu - pu­ nkalas) (Vagiranyana (ed. )1947:77) 31. The term attani occurs only in inscriptions after the 8th century. That is a derivation from (Skt) asthana which means the "Royal Court" (Wickramesinghe 1904:161, note 8). 32. The lunar calendar has been followed for dating the­ se inscriptions. Days are given with reference to the waxing or waning moon. 33. The inscriptions give evidence that several royal of­ ficers were present when a donation was made. The officers named as having .been present are Maykap- par - bodyguard (to the king) Mahale - chief secr­ etary, Kudasala - an officer of the treasury. 34. It is clear that after the 3rd century BC, kings used titular names to refer to themselves. This custom was strongly adhered to in the period between the 8th and 10th centuries AD. At this time the kings used two alternate titles, namely Sirisangabo (Skt) Sri San-ghabodhi and Abasalamevan (Skt) Abhaya Shilameghavarna. Some even used both titles, at one and the same time. 35. The pillar inscriptions of the 8th century AD to 10th century AD have named several royal officials (Ra- nawella 1984:56). 36. In the post-1st century AD period in Sri Lanka, a new architectural feature, called the padhanaghara,was introduced. It was planned to be used by the meditating monks (Bandaranayake 1974). 37. The modern connotation of pirivena is an education­ al institute for Buddhist monks. But originally it ref­ erred to a place occupied by more than one monk. 38. The Sun and Moon are two important symbolic mot­ ifs in Lankan art. They usually symbolize perpetuity. The kings of Sri Lanka often attempted to claim de­ scent from the Sun and Moon. (Surya vamsodbutha or Chandra vamsodbutha). 39. According to popular Buddhist belief, people who lead immoral lives are born into the preta world. It is a state in which the spirit lives in perpetual hunger and misery. REFERENCE Abayasingha, T., L. Dewaraja and G. P. V. Somaratne. 1977. Udarata Rajadhaniya (Sinhala). Colombo: Lake House Investments Co. Adikaram, E.W. 1946. Early history of Buddhism in Ceylon: or State of Buddhism in Ceylon as revealed by the Pali commentaries of the 5th century AD. Migoda. Aryasinghe, A. 1983. A short study of Brahmi Numerals in Sri Lanka. Sanskrit. Volume 17, No.3. Colombo: Ministry of Planning and Implementation. ASCAR 1910-11, 1911-12. 1950. Archaeological Survey of Ceylon; Annual Reports. Colombo: Department of Ar­ chaeology. Bandaranayake, S. 1984 (ed.). Unesco Sri Lanka Project of the Cultural Triangle, Sigiriya Project. First Archaeo­ logical Excavation and Research Report (January - Sep­ tember 1982). Colombo: Ministry of Cultural Affairs. Bandaranayake, S. 1990. Approaches to the Settlement Ar­ chaeology of the Sigiriya-Dambulla Region. The Settle­ Commemoration Volume. Studies in South Asian lecture. J.E Van Lohuizen Leew (ed.). Leiden: E.J. Brill. Gunawardena, R.A.L.H. 1979. Robe and Plough, Monas­ ticism and Economic Interest in Early Medieval Sri La­ nka. Arizona: The University of Arizona Press. Gunawardana, R.A.L.H. 1985 Prelude to the state. The Sri Lanka Journal of the Humanities. Vol. viii, No. 1-2. Hettiarachchi, S.B. 1988. Social and Cultural History of An­ cient Sri Lanka. Sri Lanka Series No.9. Delhi: Sri Satguru Publications. Karunaratna, S. 1984 (ed.). Epigraphia Zeylanica, being lit­ hic and other inscriptions of Ceylon. Vol. VI. Colombo: Department of Archaeology. Law, B.C. 1987 (ed.). On the Chronicles of Ceylon. Delhi: Sri Satguru Publications. Liyanagamage, A. 1987 (ed.). Anuradhapura Yugay a (Sin­ hala). Colombo: Sri Lanka National Library Services Bo­ ard. Manatunga, A. 1990. The Kiri-Oya Valley. The Settlement Archaeology of the Sigiriya-Dambulla Region. Bandar­ anayake, S., M. Mogren and S. Epitatwatte (eds.). Colom­ bo: PGLAR. Muller, E. 1883 (ed.). Ancient Inscriptions in Ceylon. Lon­ don: Trubner and Co. Paranavitana, S. 1980 (ed.). Inscriptions of Ceylon, Later Brahmi Inscriptions. Moratuva: Department of Archaeol­ ogy- Paranavitana, S. 1970 (ed.). Inscriptions of Ceylon, Early Brahmi Inscriptions. Colombo: Department of Archaeol­ ogy. Paranavitana, S. 1955. Interpretation of Vaharala. Epigraphia Zeylanica. Vol.5, part I. Colombo: Government Press. Paranavitana, S. 1920. Pre-Buddhist Religious Beliefs in Ce­ ylon. Journal of the Royal Asiatic Society, Ceylon Branch. Vol.xxi. Rahula, W. 1989 (ed.). Lakdiva Budu Samaye Itihasaya. (Si­ nhala). Colombo: Gunasena and Co. Ranawella, S. 1984. Epigraphy of the Sigiriya Region. First Archaeological Excavations and Research Report, UN- ESCO-Sri Lanka Cultural Triangle, Sigiriya Project. B- andaranayake, S. (ed.). Colombo: Ministry of Cultural Affairs. Ranawella, S. 1984-85. Ektana: The Supreme Council in the Early Mediaeval Sinhalese Administration. Kalyani. Vo- l.III and IV:56-65. Seneviratne, S. 1990. The locational significance of early Iron Age sites in intermediary transition eco-system: a prelimi­ nary survey study of the upper Kala Oya region, North Central Sri Lanka. The Settlement Archaeology of the Sigiriya-Dambulla Region. Bandaranayake, S., M. Mog- ren and S. Epitawatte (eds.). Colombo: PGIAR. Seneviratne, S. 1984. The Archaeology of the Megalithic Black-and-Red ware complex in Sri Lanka. Ancient Ce­ ylon. 5231-299. Sircar, D.C. 1965. Indian Epigraphy. Delhi. Somadeva, R. 1990 The Beliyakanda, Kaluarachchigama and Neerauva stone inscriptions. The Settlement Archaeology of the Sigiriya-Dambulla Region. Bandaranayake, S., M. Mogren and S. Epitawatte (eds.). Colombo: PGIAR. Swan, B. 1987. Sri Lankan Mosaic, Environment, Man, Con­ tinuity and Change. Colombo: Marga Institute Publica­ tion. Thapar, R. 1990 (ed.). Asoka and the decline of the Mauryas. Delhi: Oxford University Press. Tennakoon, V. 1957. Perani Lankava ha shila lipi. (Sinhala). Colombo: Gunasena. Vagiragnana, Rev. H. 1947 (ed.). Guttila Kavya Deepaniya. Colombo: W.E. Bastian and Co. Wijeratna, D.J. 1952. Interpretation of Vaharala. University of Ceylon Review. 10 (1): 103-117. Appendix: A catalogue of inscriptions Inscription (S-l) Cave no:l 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:872 SER. pg 201.no: 1 ArD.Re,no: 1593 Sigiriya Cave Inscription (S-2) Cave no: B 12 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:868 SER.pg 201.no:2 ArD.Re,no: 1585 ASCAR. 1911-12,pg 106 Sigiriya Cave Inscription (S-3) Cave no: B6 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:867 SER.pg 202.no:3 ArD.Re, no: 1587 ASCAR. 1911-12. pg 106 Sigiriya Cave Inscription (S-4) Cave no: B9 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:869 SER.pg.no:202.no:4 ArD.Re,no: 1589 ASCAR. 1911-12 pg 106-7 Sigiriya Cave Inscription (S-5) Cave no: B20 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:870 SER.pg.no:202,no:5 Ard.Re,no: 1591 ASCAR. 1911-12 pg 106-7 Sigiriya Cave Inscription (S-6) Cave no: B21 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:871 SER.pg202.no: 6 ArD.Re,no:R 1592 ASCAR. 1911-12 pg 106-7 Pidurangala Cave Inscription (P-1) Pidurangala monastery complex 3rd Cent: BC - 1st Cent: AD Cave donation SER. pg207. no: 15 Pidurangala Cave Inscription (P-2) Pidurangala monastery complex 3rd Cent: BC - 1st Cent: AD Cave donation IC. no:873 CJSG. Vol.2 pg 753.no: 753 SER. pg208. no: 16 ArD. Re.no: R753 Pidurangala Cave Inscription (P-3) Pidurangala monastery complex 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:874 SER.pg208.no: 18 Enderagala Cave Inscription (E-l) Enderagala cave complex 3rd Cent: BC - 1st Cent: AD Cave donation IC.no: 862 ArD.Re. no: R 2425 Enderagala Cave Inscription (E-2) Enderagala cave complex 3rd Cent: BC - 1st Cent: AD Cave donation IC. no:861 ASCAR. 1911-12 pg 122 ArD.Re.no: R 2425 Enderagala Cave Inscription (E-3) Enderagala cave complex 3rd Cent: BC - 1st Cent: AD Cave donation IC.no:863 SER. pg 210. no:24 ArD.Re. no:2425 Enderagala Cave Inscription (E-4) Enderagala cave complex 3rd Cent: BC - 1st Cent: AD Cave donation IC. 864 SER.pg.210.no:25 Enderagala Cave Inscription (E-5) Enderagala cave complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:865 SER. pg 210. no:26 ArD. Re.no:R 2429 Vavala Cave Inscription (V-l) Vavala ancient monastery 3rd Cent: BC - 1st Cent: AD A cave donation SER. pg 210. no:28 Vavala Cave Inscription (V-2) Vavala ancient monastery 3rd Cent: BC - 1st Cent: AD A cave donation SER. pg 211.no:29 Kaludiyapokuna Cave Inscription (K-l) Kaludiyapokuna archaeological site 3rd Cent: BC - 1st Cent: AD A cave donation SER. pg 211. no:30 Dambulla Cave Inscription (D-l) Cave no.3, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC.no:3 DER. pg 90, no:l Dambulla Cave Inscription (D-2) Cave no: 25, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC.no:836 DER.pg 90,no:2 Dambulla Cave Inscription (D-3) Cave no:7, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC.no:837 DER.pg 91, no:3 Dambulla Cave Inscription (D-4) Cave no: 58, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:838 DER.pg 91,no.4 Dambulla Cave Inscription (D-5) Cave no: 24, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no: 839 DER.pg 91.no:5 Dambulla Cave Inscription (D-6) Cave no; 10, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD May be a cave donation IC. no:840 DER.pg 91. no:6 Dambulla Cave Inscription (D-7) Cave no: 31, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC no: 843 DER.pg 92. no:9 Dambulla Cave Inscription (D-8) Cave no: 21, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC no:844 DER. pg 92. no: 10 Dambulla Cave Inscription (D-9) Cave no: 1, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC no: 845 DER. pg 93. no: 11 Dambulla Cave Inscription (D-10) Cave no:6, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no: 846 DER. pg 93. no:12 Dambulla Cave Inscription (D-ll) Cave no:20 Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC.no: 847 DER. pg 93.no: 13 Dambulla Cave Inscription (D-12) Cave no: 19. Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:848 DER.pg 94. no: 14 Dambulla Cave Inscription (D-13) Cave no: 26. Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no: 849 DER. pg 94. no: 15 Dambulla Cave Inscription (D-14) Cave no: 26, Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC A cave donation IC. no:850 DER. pg 94. no: 16 Dambulla Cave Inscription (D-15) Cave no: 15, Dambulla monastery complex 3rd Cent: BC- 1st Cent: BC A cave donation IC. no:851 DER. pg 94. no: 17 Dambulla Cave Inscription (D-16) Cave no: 22, Dambulla monastery complex 1st Cent: BC A cave donation IC. no:852 DER. pg 95.no: 18 Dambulla Cave Inscription (D-17) Cave no:23, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:853 DER. pg 94 no: 19 Dambulla Cave Inscription (D-18) Cave no: 56, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:854 DER. pg 94 no: 20 Dambulla Cave Inscription (D-19) Cave no:74, Dambulla monastery complex 3rd Cent: BC - 1st Cent: AD A cave donation IC. no:855 DER. pg 95 no: 21 Dambulla Cave Inscription (D-20) Dambulla monastery complex 3rd Cent: BC- 1st Cent: BC A cave donation IC. no: 856 DER. pg 96 no: 22 Dambulla Cave Inscription (D-21) Cave no: 74, Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC A cave donation IC. no:857 DER. pg 96. no:23 Dambulla Cave Inscription (D-22) Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC May be a cave donation IC. no: 859 DER. pg 96. no:24 Dambulla Cave Inscription (D-23) Cave no: 12, Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC May be a cave donation DER. pg 96. no:25 Dambulla Cave Inscription (D-24) Cave no:4, Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC May be a cave donation DER. pg. 96 no: 26 Dambulla Cave Inscription (D-25) Cave no: 58a, Dambulla monastery complex 3rd Cent: BC - 1st Cent: BC A cave donation DER. pg 97. no:27 Magallena Cave Inscription (M-l) Magallena Viharaya, Digampataha 3rd Cent: BC - 1st Cent: AD May be a cave donation IC. no: 866 SER. pg 216. no: 40 Potana Cave Inscription (P-1) Potanegama cave site, Kimbissa 3rd Cent: BC - 1st Cent: BC A cave donation EN. no: 13 ArD. Re. no: R 2895 SER. pg 216. no:38 Dambulla Rock Inscription (D-26) Upper terrace, bedrock on the West side 1st Cent: BC - 2nd Cent: AD Most part of text is illegible DER. pg 97 no: 28 Dambulla Rock Inscription (D-27) Bedrock, south of the Dambulla rock 1st Cent: AD - 3rd Cent: AD Most part of text is illegible DER. pg. 97. no: 29 Dambulla Rock Inscription (D-28) Bed rock, front of the main caves 2nd Cent: AD - 3rd Cent: AD Most part of text is illegible DER. pg 97. no: 30 Dambulla Cave Inscription (D-29) A cave situated south-west of the Dambulla rock 1st Cent: AD - 2nd Cent: AD A record of the construction work of stone steps DER. pg 98. no:31 Dambulla Cave Inscription (D-30) A cave situated south-west of the Dambulla rock 1st Cent: AD - 2nd Cent: AD A record of the construction work of stone steps DER. pg 98. no: 32 Dambulla Cave Inscription (D-31) Roof of the cave no:l May be some kind of donation 18th Cent: AD DER. pg 99. no: 33 Dambulla Rock Inscription (D-32) Upper terrace, Dambulla monastery complex 1st Cent: BC Record of construction work of stone steps. IC.vol 11, no:4 AC.pg 98-99 ArD. Re.no: 1166 Sigiriya Cave Inscription (S-7) Cave no: BIO 1st Cent: AD - 2nd Cent: AD May be a cave donation ASCAR. 1911-12 pg 106-7 SER.pg202. no:7 Sigiriya Cave Inscription (S-8) Cave no: B16 1st Cent: AD - 2nd Cent: AD Cave donation IC.no: 1186 SER. pg 203. no:8 ASCAR. 1911-12. pg 107 Pidurangala Cave Inscription (P-4) Pidurangala monastery complex 1st Cent: AD - 2nd Cent: AD A cave donation SER. pg 203.no: 19 Dambulla Cave Inscription (D-33) Cave no: 27, Dambulla monastery complex 1st Cent: AD - 2nd Cent: AD A cave donation IC. no:841 DER.pg.92.no:7 Dambulla Cave Inscription (D-34) Cave no:9, Dambulla monastery complex 1st Cent: AD- 2nd Cent: AD A cave donation IC.no: 842 DER. pg 92. no:8 Nirauva Rock Inscription (N-l) lkm west of Ibbankatuva cemetery 1st Cent: AD A donation made for 2 Buddhist monasteries SASDR. pg 195 Tolombugolla Slab Inscription (T-l) Tolombugolla, 3km south of the Ibbankatuva cemetery 1st Cent: AD - 2nd Cent: AD May be a record of the construction of stupa Not previously published Beliyakanda Rock Inscription (B-l) Beliyakanda, 2 km south from the Ibbankatuva cemetery 1st Cent: AD A donation made for Buddhist monastery called Giritissapabbata Vihara SASDR. pg 196 Kaluarachchigama Rock Inscription (Ka-1) Kaluarachchigama, lkm from west of Talakiriyagama 1st Cent: AD A donation for above monastery SASDR. pg 196 Enderagala Cave Inscription (E-6) Enderagala cave complex 1st Cent: AD - 2nd Cent: AD A cave donation IC. no: 1185 SER. pg 210. no:27 Vavala Rock Inscription (V-3) Vavala ancient monastery 3rd Cent: AD Not previously published Sigiriya Rock Inscription (S-9) South of Water Gardens 5th Cent: AD - 7th Cent: AD Most part of text is illegible SER. pg 203. no:9 Pidurangala Rock Inscription (P-5) Pidurangala upper terrace 5th Cent: AD - 7th Cent: AD A record of the donation made for the monastery EN. pg 2 ArD. Re. no: R 759 SER. pg 208. no: 20 Kaludiyapokuna Slab Inscription (K-2) Kaludiyapokuna monastery complex (K-3) 5th Cent: AD - 7th Cent: AD Record of the freeing of monastery slaves SER. pg 212. no:32 Kaludiyapokuna Guardstone Inscription (K-4) Kaludiyapokuna monastery complex 5th Cent: AD - 7th Cent: AD Record of the freeing of monastery slaves SER.pg 213. no: 33 Kaludiyapokuna Slab Inscription (K-5) Kaludiyapokuna monastery complex 5th Cent: AD - 7th Cent: AD Record of the freeing of monastery slaves SER. pg 213. no:34 Kaludiyapokuna Stone Steps Inscription (K-6) Kaludiyapokuna monastery complex, flight of steps at the southern vahalkada 5th Cent: AD - 7th Cent: AD SER. pg 213.no: 35 Kaludiyapokuna Slab Inscription (K-7) Kaludiyapokuna monastery complex 5th Cent: AD - 7th Cent: AD ArD Re.no: R431 SER. pg 213. no:36 Pidurangala Slab Inscription (P-6) Pidurangala monastery complex 5th Cent: AD - 7th Cent: AD A record of the freeing of monastery slaves Not previously published Mapagala Pillar Inscription (Mp-1) Mapagala-Ramakele Area 898 AD - 914 AD A record of the grant issued by the King Kasyapa IV ASCAR 1911-12 EZ. Vol: V pg 345-355 ArD. Re. no: R 2698 SER. pg 203. no: 10 Mapagala Pillar Inscription (Mp-2) Mapagala-Ramakele area 898 AD - 914 AD A record of the grant issued by the King Kasyapa IV ArD. Re. no: R1599 SER. pg 204-205. no: 11 Mapagala Pillar Inscription (Mp-3) Mapagala- Ramakale area 898 AD-914 AD A record of the grant issued by the King Kasyapa IV ArD. Re. no: 2698 SER. pg 206.no: 12 Sigiriya Pillar Inscription (S-10) Location not known 9th - 10th Cent: AD May be record of a grant ArD.Re.no:R2696 SER.pg 206. no: 13 Polattava Pillar Inscription (Po-1) Between Polattava and Mapagala 10th Cent: AD May be a royal enactment ArD. Re. no: R429 SER. pg 207. no: 14 Pidurangala Pillar Inscription (P-7) Pidurangala monastery complex 924-935 AD A record of the grant issued by the King Dappula IV ArD. Re. no: R760 SER. pg 209. no:21 Kaludiyapokuna Wall Inscription (K-8) Kaludiyapokuna monastery complex 898 - 914 AD A record of the grant issued by the King Kasyapa IV EZ. Vol.III, pg 253-260 ArD.Re. no: R429 SER. pg 211. no: 31 Kaludiyapokuna Slab Inscription (K-9) Kaludiyapokuna monastery complex 938 - 946 AD A royal enactmentEZ Vol. Ill pg: 260-269 ArD.Re.no: R 430 SER. pg 214.no:37 Kimbissa Pillar Inscription (Ki-1) Exact location not known 853 - 887 AD A record of the grant issued by the King Sena II ArD. Re. no: R 2894 SER. pg 216 no: 39 Peyikulama Pillar Inscription Peyikulama, south-east of the Sigiriya rock 887- 898 AD A record of the grant issued by the King Udaya II ASCAR 1956 pg 12. no: 27 Ard. Re. no R2352 SER. pg 217. no: 41 Viyaulpota Pillar Inscription Viyaulpota, west of the Sigiriya rock 853-887 AD A record of the grant issued by the King Sena II EZ. Vol:IV pg 176-180 SER. pg 217. no: 42 Dambulla Fragmentary Pillar Inscription Dambulla Somavati Dagaba premises 9th - 10th Cent: AD May be record of a grant Not previously published Taranavava Fragmentary Pillar Inscription Taranavava. ASCAR - Archaeological Survey of Ceylon, Annual Report. AC - Ancient Ceylon. ArD — Archaeological Department Inscription Register. CJSG - Ceylon Journal of Science. Section G. Cu - Culavamsa. DER - Dambulla Project: First Archaeological Ex­ cavation Report. EN Epigraphical Notes. EZ - Epigraphia Zeylanica. IC - Inscriptions of Ceylon. Vol.l. Mv - Mahavamsa. SASDRSettlement Archaeology of the Sigiriya- Dambulla Region. SER - Sigiriya Project: First Archaeological Ex cavation and Research Report.

EVA MYRDAL-RUNEBJER INTRODUCTION The present study is an archaeological case study of an aban­ doned pre-modem irrigation system in Sigiriya, which is in the northern sector of the Dry Zone of Sri Lanka. The full field report on this work is to be presented in a separate volume. What appears here is only an outline of the study and a discussion of its preliminary results. Previous discussions regarding pre modem irrigation tech­ nology and wet-rice cultivation in Sri Lanka have focused on dating and technological development and decline in an over­ all Dry Zone perspective. Apart from the structures visible on surface - most often documented by dedicated engineers and surveyors - the principal basis for such studies has largely been inscriptions and literary sources. In contrast, the focus in the present study is on a given system of moderate size, with the possibility of relating the data to rural and urban settlements and elite secular and religious structures. The source materials here are the material remains themselves, and the methods are archaeological - that is, field surveys and test excavations to obtain a frame for dating and information about constructional features. Written sources regarding the construction of this specific system have so far not been found. A few inscriptions can be used to infer a terminus ante quern for part of the system. Moreover - as no historian will deny - in any situation written source material will only provide us with part of the informa­ tion we seek when we discuss dating, technology and produc­ tivity potential. The material remains are manifest functional history, the interpretation of which, however, is unfortunately not always obvious, as will be clear from the discussion that follows. The field survey of part of the system the Vavalavava canal commenced in 1988, although no results furthering a function­ al or chronological understanding of the Vavalavava canal were reached at that time (Mogren 1990:59). In 1989 two minor trenches were taken up in the tank bund itself, address­ ing the question of chronological relationship between a cyclopean wall connecting to the tank bund and the stone­ pitching of the bund (Gunasiri, PGLAR archive report 1989). The canal survey was taken up again during a two week Figure 12:1 Research area ofVavalavava - Sigiri Mahavava irrigation system survey. part-time survey in 1990, by the present writer, two assistants and one to two labourers. It was continued during a full SARCP field season in 1991 by a single Lankan archaeologist, one assistant, the Swedish consultant and from two to ten labourers. In 1992 the field team consisted of eight Lankan archaeologists, the Swedish consultant and up to eight lab­ ourers. OBJECTIVES Population, productivity and the organization of labour The key words in this study are human labour and productive potential. The study tries to define the archaeological remains of this irrigation system in terms of function, technique of construction and dating, potential of production and labour investment. Based on such data analysed in relation to the settlement pattern, it is possible to address three important questions regarding the premodem society of the area during given periods: (1) the maximum size of the population that could be supported by the system - an upper limit of the system’s production capacity. (2) the control and appropriation of agricultural produce (the level of which is basically testified to by all forms of non-agricultural construction, in the present area and within the time-frame covered by the study, from peasant houses to the monasteries and iron production sites). (3) the control and organization of labour. The first of these questions will be addressed within the frame of the study discussed below. The second, which re­ quires a more in-depth study of rural settlement sites, along with the related third question, will only be discussed in the coming report as an outline of a possible field strategy and methodology. HYPOTHESES AND METHODS Questions at issue It is a widely recognized fact that a timely supply of water is the most important yield increasing factor in wet-rice cultivation in semi-dry tropical regions like the Dry Zone of Sri Lanka and southern and eastern Tamil Nadu. Artificial irriga­ tion is a means of establishing control over the water supply. At the same time, while irrigation is the key factor, it is definitely not the only yield-increasing factor. It allows for a variety of productivity-increasing practices to be developed, as shown for example by David Ludden regarding Tamil Nadu and Francesca Bray regarding East and South East Asia (Ludden 1985; Bray 1986). The difference between the small scale irrigation system based on local precipitation on the one hand and the macro irrigation system, with feeder canals enlarging the catchment area on the other, is to be noted however. They have, as pointed out by many scholars and obvious to every peasant, different implications regarding possible return and labour in­ vestment and the organization of labour. Broadly speaking we could distinguish the following pr­ emodern agricultural possibilites (with different supportive capacities) in the study area (see Myrdal this volume):(1) Polyculture swidden cultivation combined with other forest-based food procurement activities such as gathering wild plants, honey collection, hunting and trapping. (2) Paddy cultivation based on village tanks which, owing to the irregular supply of water, often has to be combined with swidden cultivation and related practices. This combination still forms the most important subsistence pattern for the pe­ asants in the study area. (3) Paddy cultivation based on a macro system of irriga­ tion, with a larger and more secure water supply which would make paddy cultivation the most important food procurement technique. Owing to the man-created, wet-rice agricultural landscape, several forest-bound food procurement practices would then be ruled out. To understand the material pre-requisites of irrigated ag­ riculture under small scale systems, a series of interviews have been carried out with elderly male villagers (mostly former vel vidanes) in eight villages in the Sigiri Oya basin. These have been followed up in Talkote village by field surveys of village tanks, small few-family tanks (vav kotu) and micro tanks (te­ rmed amuna by the villagers ) in the paddy fields. The ethno-archaeological study in Talkote village also pr­ ovided information regarding paddy cultivation and main­ tenance work. MACRO-IRRIGATION SYSTEMS IN THE SIGIRIYA AREA Macro-tank: Sigiri Mahavava In the Sigiriya area, the only premodem irrigation system so far discovered which has the potential of increasing the supp­ ly of water above the localised village tank level, is the Sigiri Mahavava itself and the three canal systems we now know of, that might have been functioning as supply canals to the tank. As Assistant Superintendent of the topographical survey of the Sigiriya area in 1898, D. Blair was able to establish the outline of the Sigiri Mahavava bund as it is known even on the present topographical map of the area. This was an improve­ ment, as compared to Blakesley’s survey of the tank bund in 1874, when the bund was traced down to the pathway between Palutava and Polattava only, and the presence of the southern half of the tank was unknown (Blair 1898). Supply canals to the east of Sigiri Mahavava: eastern and western canal systems Commenting on the large size of the tank, Blair suggested that for the Sigiri Mahavava to have functioned properly it would have needed a connection to a wider catchment area than its natural topography affords. He went on to suggest a connection to the Kiri Oya basin, east of the tank. H.C.P. Bell also mentions a canal from Vavalavava as the connection between the Kiri Oya basin and Sigiri Mahavava, and cites the oral tradition in the area that this canal was constructed by King Mahasen (276-303 CE) (Bell, Fernando, Moysey 1914). The southern extension of an abandoned canal from Va­ valavava, built over Kiri Oya, was mapped by surveyors be­ tween 1897 and 1902, and in the 40’s the northern extension of the eastern canal bund and the western, partly parallel canal bund, was depicted in the revised map. In 1990 Chandana Ellepola presented the idea that water was brought from Va­ valavava to Sigiri Mahavava through a connection between the eastern and western canals (Ellepola 1990:178). If water was brought from the Vavalavava canal to Sigiri Mahavava, this is a trans-basin canal which connected the upper Kiri Oya catchment with the Sigiri Oya basin, and which may thereby have fed not only the vava but directly irrigated the land between the canal and the Kiri Oya itself. Looking at this closely, we notice that there are, in fact, two systems: one, the eastern canal system, which starts at Vavalavava and continues north, according to the topographi­ cal map, ending about 800m east of the northern most part of Sigiri Mahavava, on top of a low ridge that constitutes water­ shed between the canal and the vava\ the other, the western canal system, which starts midway of the eastern canal, runs parallel to it for a short stretch and holds a seasonal stream on its western side. The stream ultimately falls into Sigiri Ma­ havava. The preliminary field survey in 1990 and 1992, the recent contour mapping by the Survey Department (ground truth not established), and the l/2m interval contour mapping carried out for PGIAR by a licensed surveyor and leveller in 1992, have not given any evidence that water from the Kiri Oya actually passed into the western system, or that the two canals were connected (see fig. 12:2 and 12:3). The canal bed surface at the end of the western bund is situated about 1.5m above the eastern canal bed, and no indication of a southern continuation of the western bund has so far been found. The present canal bed surface of the western bund is about 1.2m over the height above mean sea level of the documented ancient sluice in Vavalavava. The present western canal bed surface is about 2.6m above the canal bed base of the trench taken up 1.5km to the south in the eastern canal.1 The western, shorter canal might therefore have had two purposes. One to collect run-off water from the slightly ea­ stwards slope and further north, hold the water of the stream, ultimately draining into Sigiri Mahavava; and/or to protect the eastern bund from destruction by run-off water. The eastern canal from Vavalavava might also have had a double purpose: one (which might have been its initial func­ tion) to irrigate the land between the canal and Kiri Oya. This is an area where it has now been discovered, large scale and technically advanced iron production had been developed du­ ring the 2nd century BCE and the 3rd century CE (see Fo- renius and Solangaarachchi this volume); the other, to supply the Sigiri Mahavava with water, i.e. the water that was left could have drained into the northern most part of the Sigiri Mahavava. Whether the eastern canal really joins the northern most part of the tank however is not yet clear. Judged by present day topography it could have continued north, through a present system of minor tanks, and at the end joined Sigiri Oya just south of Hirivadunnavava. The field survey of the northern part of the eastern canal has not been completed at the time of writing. The trenching of the eastern canal in 1992, in three places, gave evidence of an initial heavy sedimentation of the canal bed. The sediment indicates erosion from the bund and natural slope above the canal by rain, rather than sediment transported and settled in running water. Based on these facts, we have to ask whether the eastern canal was ever put into use or, alternatively, if it had functioned for a very short period only. Soil sample analyses for diatoms in the canal sediment, and char­ coal dating now under way will hopefully give a more com­ plete answer than we could gain from ocular observation only. Ancient sluice at Vavalavava Preparing the restoration of Vavalavava, the Irrigation De­ partment documented an ancient sluice at the site where they constructed the new one.2 The modem sluice was constructed 10 feet above the ancient one. Comparing the Irrigation De­ partment plan with the topographical map, it was indicated that this ancient sluice was identical to the sluice to which the abandoned eastern canal was connected. As the northern part of the canal bund is restored, the close command area is much changed by activities of new settlers. We were not able to trace the beginning of the canal on the ground either in 1988,1990, or 1992. In 1992 we lost track of the canal about 1.1km north-west of the Vavalavava bund. The contour mapping recently earned out by the Survey Department was not able to depict the beginning of the canal bund either. This lack of evidence is a problem since there is a reference by H.L. Moysey, in his field survey of 1884, to two sluices from the then totally abandoned and jungle covered tank. He states that he was able to visit only one of them, but does not indicate anything about where (on the tank bund) the sluices are situated (Bell, Fernando, Moysey 1914). As the location of the ancient sluice documented by the Irrigation Department and the beginning of the canal on the topographical map con- cides, we have decided to base our calculations on the efficien­ cy of the system on the height above mean sea level ascribed to the ancient sluice by the Irrigation Department plan. It should be kept in mind, though, that ground truth has not been established by the project itself. Southern supply canal One hundred and eighteen years after Blakesley’s initial sur­ vey of the Sigiri Mahavava bund, SARCP field team mem­ bers, guided by peasants living in the area, found that Sigiri Mahavava was connected by a canal bund continuing from the southern end of the tank to the Vegollavava and from there by a still existing waterway, termed ‘ancient irrigation channels’ in the topographical map, probably to the Mah- agona Oya. This canal could then have provided the vava with an additional supply of water without a canal connection to the Kiri Oya basin (see fig. 12:3). post quern for the bund of the eastern canal, south of the possible connection and the western bund. The eastern canal itself is of course one project, in the sense that it must have been built from the south northwards, always in relation to the known level of the outlet and water­ head of Vavalavava. Whether this took place, gradually or as one single effort, could be tentatively tested by obtaining a terminus post quern of the bund along the canal. Three probable stages could be hypothesized: (1) from Vavalavava up to a seasonal stream that joins from the west, now continuing through a breach to the Alakolavava tanks in the iron production area. Here the canal bed is slightly bowl-shaped, forming a local watershed to the north in the canal itself. This was probably to protect the bund from the inflow of water from the west, and to trap silt car­ ried by this water.3 It might also have been possible to end the canal bund here, regulating the water outlet from Vavalavava in relation to the watersupply from the natural stream. The present breach might have been the site of a sluice. (2) from there up to a hypothesized connection to the western canal leading to Sigiri Oya, (3) from the hypothesized canal connection northwa­ rds. Regarding the Sigiri Mahavava itself, we will test the as­ sumption that it was built in stages. Minor streams coming up to the abandoned bund, two minor tanks which today function independently along restored parts of the ancient bund, signs of local gathering of water behind the abandoned bund, and cultural remains visible in the bundfill are cases in point. Five different areas of possible independant functioning have been identified, together with one stretch of the bund - the midway narrow passage where Blakesley stopped his sur­ vey, terming it a canal - that might specifically belong to the unifying project. These initial observations will be followed up in the field by mapping, documentation of visible construc­ tional features and, if time allows, trenching of bunds to obtain a terminus post quern. Study area After the preliminary field survey, the study area was delineated to comprise Sigiri Mahavava, the possible irriga- tion/feeder supply canals east and south of the vava, the storage tank supplying the canal - that is Vavalavava, and a feeder tank ‘competing’ with the eastern canal and maybe with Sigiri Mahavava for water from the same catchment - that is Peikkulam tank. This tank mainly feeds the other side of the east-west watershed. Water from this tank ultimately flows into the Elahera canal, just south of Konduruvava (see fig. 12:1). What the physical properties and possible datings tell us of maximum water-holding capacities in given periods and the maximum command area, could be termed the potential Va­ valavava and Mahagona Oya canal/Sigiri Mahavava function­ al history - that is, the physical limits of its yield increasing potential in given periods, and its labour investment implica­ tions. Further studies But what did this mean at the village level? How was the potential realized? The manifest functional history of the vava is of course found in the command area. To address these questions, a settlement study area could be delineated com­ prising the direct command area of the tank in the south and the Sigiri Oya basin up to, but not including Hirivadunnavava in the north, and the area between the Vavalavava canal and Kiri Oya (see fig 12:1). This however is not included in the present study. Regarding the additional water supply to the tank, there are seven possible connections that could have functioned con­ temporaneously: - a connection only to the Sigiri Oya basin - or to the Sigiri Oya and Kiri Oya basin - or Sigiri Oya and Mahagona Oya, - or Kiri Oya - or Mahagona Oya - or Kiri Oya and Mahagona Oya - or all three at the same time. We also do not know if all, or some of the various pos­ sibilities were realized at different times. At least eight important questions have to be answered: (1) Was there any flow of water from the eastern, Vav­ alavava canal to the western, Sigiri Oya canal? (2) Where did the eastern canal end? (3) Did the eastern canal have the double purpose of irrigation and feeder supply? (4) What is the time-frame for the construction of the three canals? (5) What is the time-frame for the usage period of the canals? (6) What is the time-frame for the construction of the tank? (7) What is the possible volume of water that could be transported through the system, considering the va­ rious possibilities? (8) How many man-days went into its construction and maintenance? Another important question is to find out which areas could benefit from the water of the large tank and the canals. In this connection, several statements by villagers from neigh­ bouring villages regarding water courses from the tank, should be followed up by field surveys. FIELD WORK SEASONS 1989-92 Village based systems In 1989 commenced the study of present day village-based systems. The interview part of that study is now complete. The most important conclusion to be drawn from these inter­ views, regarding maintenance and water management, is wh­ at has often been pointed out in literature - namely, the possibility of maintenance and management of rain-fed village irrigation to be carried out within the limit of the village itself. An expression of this locally managed system was a former vel vidane’s statement that "the tanks have always belonged to us" (that is, the cultivating peasants in the vil- lage). In one case it was described how a vava (Nagalavava Mahavava) was built on the villagers’ (‘the ancestors’) initia­ tive, but by specialized labourers from outside the village. They were called "otteruwo", and were said to have come from India. According to the tradition, they were paid in kura- kkan (PGIAR archive report Herath, Yasapala 1989/1990). The field survey in 1989-90 corroborated this assumption of locally organized labour and management. Several aban­ doned vav kotu were mapped. Some of them had a known history of a "few-families-tank" in a cultural landscape con­ stantly changing regarding subsequent restoration, abandon­ ment and construction of small scale irrigation facilities. Our attention was also drawn to what the villagers termed amuna. In this case the term designates a minor embankment within the paddy field, building a waterhead to irrigate paddy fields on a slightly higher level than the fields between the tank bund and the amuna (but naturally below the level of the tank). This way of enlarging the cultivable area is another example of a locally based irrigation technique, though of course we know nothing of its time of origin. It is a labour-saving technique, as water would otherwise have to be lifted up to the fields. In 1990 we further documented an abandoned and miscal­ culated minor canal project, initiated by the villagers in the 1950s. The intention was to connect the biggest rain-fed tank of the village with a seasonal stream further south, in a then jungle covered area. Having dug approximately 250m starting from the tank, the participants found that the water drained the wrong way, and the project was abandoned. No levelling instruments were used. According to one participant they simply dug, under the leadership of the village headman. The project was undertaken on the assumption that the land further south was on a higher level. This is true, though the difference is slight. Such a trial-and-error technique couldn’t have been used when major projects were successfully undertaken 4 Eastern irrigation and feeder supply canals The eastern canal is built mainly along the gentle eastern slope on the other side of the watershed that constitutes the boundary of the Sigiri Mahavava’s direct catchment area. The canal would not only have brought water from the Va- valavava, it would also have gathered run-off water through its entire length. It consists of an earthen bund mainly built along slightly inclining ground. For most of its length there is thus but one constructed side; the bund material taken from what then becomes the canal bed. There are, however, st­ retches where a double bund could be discerned. About 2km from the ancient documented sluice, it further crosses a low sandy/gravelly ridge. Here the canal bed has been dug more than 3m into the ridge. Water will be transported and col­ lected on the upper side, in this case the western side, of the bund. The bund along most of the surveyed canal was eroded and only between 0.3-lm high. A bund height of up to 2m was documented however. The canal bed at its widest was 15m. The canal now carries water for short stretches during the rainy season, but breaches in the bund have changed the direc­ tion of drainage along the bund in several cases. No other constructional features such as stone lining or sluices were noticed during the survey. Vavalavava is now restored and as discussed above, we have not been able to trace the beginnings of the canal. Instead, we started from the crossroad near Alakolavava, close to where the western canal also begins. From here we could follow the bund about 2km to the south, where it disappeared a short stretch after the major breach (see below). The field situation here is difficult owing to newly undertaken dry land cultivation, where the ground has been partly levelled by bu­ lldozers. Spot-height levels were taken along the present canal bed (see fig. 12:2). Late in the field-work season in 1990, we found the con­ tinuation of the bund further south, but lacked the time to continue the spot-height survey. In 1992 we continued the ground survey down to ca 1.1km from the Vavalavava bund, where the trace was completely lost. North of the crossroad, we could follow the bund only a few hundred meters, halted by the presence of elephants. Pr­ essed for time, we started northwards instead, from where the canal crosses the road from Polattava to Peikkulam tank. The bund was found to have been rebuilt into a tank bund (Pahalamakarayavalavava) north of the road. Soon after the (now abandoned) tank bund ended, we lost track of the bund in the jungle. Seasonal streams within the area were found to drain north-west. According to the topographical map, the bund ends about 800m east of the vava. Because of dense scrub jungle in this elephant infested area, it was not possible, during this preliminary survey, to find the northern most part of the canal bund, to check on a probable connection to the tank or to a possible outer moat of the Sigiriya complex. Villagers who had been doing swidden cultivation in the area north of the road, said the bund was visible for short stretches, but much dilapitated. Our hypothesis, that this canal was used directly for irriga­ tion purposes, is built on the following observations: At many places it will not be possible to use the Kiri Oya itself for directly irrigating fields to the west, as the bed is below the adjoining land. That arable land exists below the canal is seen from the small abandoned vavas in the area, draining towards Kiri Oya, and also from the fact that part of the canal bund had been rebuilt into a separate tank (Pahalamakarayavalavava). We further suggest that the canal was intended for irriga­ tion also along its southern course. As mentioned above, there is between the canal and Kiri Oya an area that must have been of great economic importance, at least during the centuries around the beginning of our era. We also suggested a site for an outlet to irrigate this area, a bowl-shaped watershed, in the canal itself (see above). Judging by the lay of the land outside the canal bund, the watershed did not result from erosion caused by the major breach. The bund seems to have been deliberately constructed ‘down hili’, where a stream joins from the west. As mentioned above, one functional explanation could be that it was de­ signed to meet the inflow of water, and to act as a silt trap for the stream from the west. The bund must once have been considerably higher, for the water to ‘climb’ the watershed northwards. We will not be able to prove the irrigation hypothesis: we can only show it to be probable. A more thorough settlement site survey in the command area of the canal would strengthen the hypothesis, but knowing the thorny, elephant infested, dense scrub jungle terrain and the available resources, this is not suggested at present. A way of finding out if the canal was ever in use along its northern-most stretch and to obtain a terminus post quern for the construction, would be to trench the bund north of the point where the western and eastern canals might be connected. Indications of a waterflow in the eastern canal, north of this possible connection, would indicate that the eastern canal has been functioning. Other parts of the bund that are important to trench and date, are south and north of the major watershed in the canal, as discussed above. In 1992 trenches were taken up at all three locations.5 The analyses of sedimentation phases in the canal bed however are not yet completed, hence the excavation campaigns in 1992 are to be discussed in the final report. There are, however, indications of a short time of use and neglected maintenance. Western irrigation and feeder supply canal The western canal bund is much shorter. Judging from the topographical one-inch sheet, it is about 1km. It begins in slightly high ground, east of a small seasonal stream, and turns east, close to the eastern canal bund. No signs of a continuation further south were seen during the preliminary survey. It ends about 1.2km south of the Polattava village vavas. Here a seasonal stream continues, ultimately draining into the ancient Sigiri Mahavava. We could follow the bund through the scrub jungle only for 520m before a herd of elephants approached us. It was clearly visible where we could follow it, and though it is abandoned and densly overgrown, we saw only one minor breach. From the lay of the land it could be inferred that it collects water coming from the west, and it keeps it on that side. There were no signs of water passing east to west. Its main purpose must have been to hold the water from the stream on its western side, so that it could be brought to the vavas, and maybe more importantly, to protect the eastern bund during the rainy season.6 The important question is whether water was allowed to pass from the eastern canal to this one. Unfortunately, the field situation at this crucial point is very difficult. The Vavalavava canal bund is cut away first by the newly built road to Va­ valavava, directly after that by the Lenava-Alakolavava road. It is faintly visible continuing through a newly abandoned home garden. From here however it is clearly visible, running parallel to the west of the road until it turns northeastwards into the jungle. The contour survey conducted in 1992 gave no indication that the western canal was connected to the eastern one, as discussed above. Sigiri Mahavava bund From the preliminary survey of the tank bund, started in 1990, we have decided to divide the further study of the vava into five parts. One concerns the northern-most part of the vava closest to the Sigiriya complex itself.7 As proved by the present situa­ tion, this part of the tank can function as a rain-fed tank (Sigirivava) independently of the abandoned parts to the south. Here, trenching of the tank bund, the tank bed and part of the outermost rampart in the vava bed, was commenced and completed in 1991. The major questions were of the relation, regarding chronology and function, between the tank bund and the Sigiriya complex. The Sigirivava bund has recently been restored, but in 1991 there remained unrestored parts to the north and south. No traces of protective pitching of stones have been found along this northern-most part of the ancient Sigiri Mahavava bund. It was observed that the direct command area falls partly within the area of the outer rampart of the Sigiriya complex, west of the tank.Trenching the bund at a breach in the northern-most part, it was further found that the bund has been built on cultural layers containing bricks and tiles. On top of these there are thin layers of pure sand, indicating a period of total abandonment. Stones probably fallen from a demolished wall on the hillside were resting on the lowest sandy layer, further st­ rengthening the abandonment theory. They had later been built into the bund. Charcoal samples were taken from the cultural layer to obtain a terminus post quem for the construction of the bund. Inside the tank there is a l-1.5m high earthen bund. It continues also outside the tank bed, running approximately east-west for about 2km. In the east it joins high ground, a low ridge running north-south. To the west it doesn’t come up to the present tank bund, but ends about 150m east of it. It runs almost parallel to the Eastern Precinct rampart. It was assumed that this bund at one time constituted the outermost rampart of the Sigiriya complex (Karunaratne this volume). One possibility that was followed up in the field was that the Vavalavava canal, which according to the topographical map ends below the bund on the same ridge, might have supplied water to an outer moat held by the outer rampart. Along the ridge running north-south, there were no signs of any watercourse or sluice construction that indicated a water supply. A possible connection must therefore be situated fur­ ther west. No conclusive evidence could be found, neither of a joining canal bund nor sluice. There is a breach in the rampart, but here the ground slopes steeply to the south, and that is why it seems unlikely that water could have been brought in at this point. From what was found regarding the outermost rampart in the vava we now suggest that the canal water drained directly into the vava; if there was indeed a connection between the canal and the vava. Following the straight line of the rampart westwards, we end up about 40m north of what is left of the higher outer rampart, in the south-western comer of the complex. This part of the rampart was trenched as a part of the Cultural Triangle Project. Comparing what was revealed of the constructional features in the trench there, and in the trench of the low bund inside the vava, we suggest that they represent two building phases. This assumption is strengthened by the fact that they wouldn’t meet if the direction of the nearly 2km long straight bund is followed westwards. The inner moat at the southern side of the Sigiriya complex is built in sections, probably due to the ground sloping westwards. The outer moat at the south-western comer might also have been such a section. The bund inside the vava was built on a former stream bed at its western end, where our trench was taken up. In streambed sedimentation layers below the bund, Black-and-Red Wa­ re potsherds and iron slag were found. Red Ware from a variety of vessel forms was found in the bundfill itself, giving the lower time limit to the construction of the bund if a more precise Sri Lankan pottery chronology could be built up. Just to the north of the bund, decaying limestone bedrock reaches up to 0.5m below the present surface. Only 0.2m of the sedimentation layers above are from a ‘post-bund’ situa­ tion, as seen from the profile below the bund construction. There are no signs of a dug out moat bed. The grain size indicates sedimentation in still, or slowly flowing water. The tank bed has approximately the same height above mean sea level on both sides of the bund. This situation makes the assumption that the bund retained an outer moat (as seen in the south-western corner of the complex) on its northern side a bit unlikely. Neither are there any signs of a canal bed. There is a 160m wide slowly rising tank bed to the north of the bund. A test pit in front of the end of the bund yielded no traces of a bund construction. Either the bund was never extended to the west, or was completely removed when the tank bund was built. To obtain a terminus post quem, soil samples were taken for charcoal from the layer immediately below the bund con­ struction. Soil samples were also taken for particle size an­ alyses, and for macro fossil, pollen and diatom analyses from tank sedimentation layers, and from the stream bed layers below the bund. Charcoal samples were also taken from these layers. Preliminary conclusions - Sigiri Mahavava There were activities in the area before the building of the outer rampart, as indicated by Black-and-Red Ware and iron slag (which wouldn’t have been moved a very long distance from the iron production site itself) found below the bund. The outer ramparts of the Sigiriya complex seem to have been built in several stages. The use of the land west of the tank for agricultural pur­ poses seems to have been taken up after the decay of the outer rampart, as the paddy fields under the tank fall within the limit of the outer rampart to the west. The tank bund is also built after two phases of elite related structures (bricks and tiles) had been existing at the southern side of the Sigiriya rock and after the following abandonment and decay (sand layers and fallen stone wall). South of Mapagala, the bund has a different character. The inner side of the bund is pitched with stones, and there are remains of a probable sluice construction (a nicely cut, long stone slab) in a breach in the bund. It is to this part of the vava that the Sigiri Oya flows. EVA MYRDAL-RUNEBJER Two test pits were taken up in 1989 to date the stone pitching in relation to the Mapagala cyclopean wall, and to obtain a date after which the bund was built (PGIAR archive report, Gunasiri 1989). It was found that the Mapagala wall predates the tank bund and the stone pitching. From the ex­ cavation of the Mapagala wall we know that the wall was built after 306-517 CE but before 432-550 CE (Kumaradasa this volume). The former date also gives an absolute terminus post quern for the building of the tank bund south of Mapagala. Below the bund there are several cultural layers. Black Ware and a still undated Roman coin indicates a settlement with outside connections, from a period before the present tank bund was built. As Blakesley pointed out, there are other indications that the Mapagala wall might predate the maximum extension of the tank. The wall runs along the eastern side of Mapagala as well, though this side would have faced the tank when it was fully developed (Blakesley 1876:59). The Sigiri Oya might have been embanked further south­ east, before the building of the present bund. The features of the latter (stone pitching and possible sluice, with nicely cut stone lining) however indicate a tank ‘above’ village based irrigation. The stone pitching might be a later addition. It is not found anywhere else along the bund. It might be an unfinished project to protect the bank, started when the volume of water was enlarged (the maximum extension of the tank bund and the completion of, or planning for the eastern and/or southern supply canals). As indicated by the cyclopean wall, labour not related to the basic production and on a larger scale than would be expected of a subsistence agricultural village, had already been undertaken before the stone pitching at least was completed. Further south, at Kayamvala, there is now a village tank made up of a reconstructed part of the Sigiri Mahavava bund. As seen from the present situation, it can function indepe­ ndently of the major tank. It is now fed partly by a small seasonal stream. There is much cultural material in the present bund fill, indicating a settlement at least before the reconstruction (whi­ ch is modern). To find out the relationship between the or­ iginal bund (at the site that has proved to function independently) and signs of settlements, we will try to locate an abandoned part of the bund close to the reconstructed sec­ tion and take up a test pit. South of Kayamvala, the possible water-filled area nar­ rows (as understood by the cartographers following the high water contour of the tank). Finally, it has more the charac­ teristics of a canal than a tank. This is also seen in the field. Blakesley suggested that this was a canal with the purpose of bringing water into the tank to the north (Blakesley 1876:56). We now know that the bund continues south, with a possibility of holding approximately the same amount of water as the tank below Mapagala. From Kayamvalavava, Sigiri Mahavava has at present a low earthen bund. No protective stone pitching is visible, and no signs of sluices were seen during the preliminary field survey. The tank bund continues into a canal bund down to Vagollavava as was documented in the field season 1992. No major stream is known to be embanked by the vava bund south of the canal joining the northern and southern part. There is one seasonal stream (now draining to Kaluagalavava) that might have been embanked before the larger bund was built. To conclude: the following hypotheses regarding the Sigiri Mahavava are to be tested in the field and further discussed: That there existed smaller tanks, for example embanking Sigiri Oya, Kayamvala-ala and the stream to the south before the unification of the bund into the Sigiri Mahavava. The case of Sigirivava is to be further studied. Our hypo­ thesis is that the outer rampart predates the vava. There are indications from the pre-bund layers north of the bund that there is no tank sediment prior to the bund construction. ACKNOWLEDGEMENTS Much material evidence relating to the premodem Vava- lavava and Sigiri Mahavava irrigation systems has vanished through modernisation. Without the possibility of consulting maps and archive material related to previous documentation of these systems, there would have been a much less substan­ tial base for the present study. These background studies were made possible by the kind assistance of the department heads and staff of the Survey Department, Colombo and the Irriga­ tion Department, Colombo, and the irrigation engineer and staff of the Dambulla office of the Irrigation Department. To all these institutions we convey our deepest thanks. A special word of thanks also to the head and staff of the cartographic unit of the Postgraduate Institute of Archaeology, Colombo, who supported the field team by obtaining indispen­ sable aerial photographs and maps to be used in the field. 1. Technical excavation report. SARCP: Vavalavava c- anal 1992. Eva Myrdal-Runebjer. Archive report, P- GIAR. 2. Irrigation department. Matale subdivision. Vavala­ vava. Tank contour and dataplan. 17.12.69. 3. See for example D. Blair, cited in Brohier (1934) 1979:7 regarding the layout of canal bunds to meet adjoining streams. 4. See Leach 1961:224, 314-315 regarding an indiv­ idually undertaken field canal project that failed due to a similar lack of proper levelling. 5. Technical excavation report. Vavalavava canal. 1992. Eva Myrdal-Runebjer. 6. I am indepted to Martha Prickett for the latter sug­ gestion. 7. Technical excavation report. Vavalavava canal. 1992. Eva Myrdal-Runebjer. REFERENCES: Bandaranayake, S., M. Mogren and S Epitawatte (ed.) 1990. The Settlement Archaeology of the Sigiriya-Dambulla Re­ gion. Colombo: PGIAR. Bell, H.C.P., W.M. Fernando and H.L. Moysey. 1914. Ap­ pendix - Nuvaragala Kanda. ASCAR 1914 Blair, D. 1898. Report of the Topographical Survey and An­ cient Irrigation works in the Tamankaduwa District. As­ sistant Superintendent. Annexure B Administrative Report 1898. Colombo: Ceylon Government Press: B 12-B 18. Blakesley, T.H. 1876. On the ruins of Sigiri in Ceylon. Jour­ nal of the Royal Asiatic Society Ceylon Branch 8. Bray, F. (1986) 1989. The Rice Economies. Technology and Development in Asian Societies. Oxford: Basil Blackwell. Brohier, R.L. (1934) 1979. Ancient Irrigation Works in Cey­ lon. Part 1-3. Colombo: Ministry of Mahaweli Develop­ ment. Ellepola, C. 1990. Conjectured hydraulics of Sigiriya. Ancient Ceylon. 11: 169-227. Gunasiri, R. Sigiri Mahavava. Unpublished Excavation Rep­ ort. PGIAR Archive. Cat. No. 1989. Gunawardana, R.A.L.H. 1971. Hydraulic Society in Early Medieval Ceylon. Past and Present 53: 3-27. Herath, K. and A. Yasapala. Unpublished Irrigation Study Interviews. PGIAR Archive. Cat. No. 1989/1990. Leach, E.R. 1959. Hydraulic Society in Ceylon. Past and Present 15: 2-26. Leach, E.R. 1961. Pul Eliya - A village in Ceylon. Cam­ bridge: Cambridge University Press. Ludden, D. 1985. Peasant History in South India. Princeton: Princenton University Press. Mogren, M. 1990. Project Strategies: Methodology and Per­ spectives. The Settlement Archaeology of the Sigiriya- Dambulla Region. Bandaranayake, S., M. Mogren and S. Epitawatte (eds.). Colombo: PGIAR. Myrdal-Runebjer, E. Vavalavava Canal. Unpublished Tech­ nical Excavation Report. PGIAR Archive. Cat. No. 1992. Maps: Polonnaruva sheet G/12. Pt.board G21/1A; G21 3A. Survey Department. Colombo. Tank contour and data plan. 17.12.69. Irrigation Depart­ ment. Matale sub division. Archived at the Dambulla office of the Irrigation Department.

Studies of Sigiriya have in the main concentrated on the western sector of the royal city complex. The Water Gardens, summer palaces, Royal Audience Hall, the fresco pocket, the Mirror Wall and the palace on the summit have always held the interest of both tourist and archaeologist alike. It was towards the end of the 19th century, however, that some attention was first paid to the Eastern Precinct, a long time previous to the present survey. Of those who started their work on the eastern sector, Blakesley was the first to document his work. Although ex­ ploration of the eastern sector was not his sole objective, Blakesley seems to have roamed widely over the area, as seen from his notes. He has presented facts about the moats, ram­ parts and other structures built around the rock and even about the Sigiri Mahavava. Though Blakesley has not presented any systematic analysis of how the Eastern Precinct had been used, the systematic and descriptive plan he has presented shows that he had a broad understanding of the eastern area. His plan is an important source for the further study of Sigiriya (Blakes­ ley 1876). Extensive archaeological research into the Sigiriya com­ plex was carried out during the last decade of the 19th century, under H.C.P. Bell. Particularly as a result of surveys and ex­ cavations Bell made after 1894, he was able to focus attention on the entire Sigiriya complex in a clear and broad perspective. He mentions in his report how he cleared the jungle which concealed the buildings and ramparts of the western sector. He did not stop at that, but also cleared the jungle in the eastern sector, up to the footpath leading to Pidurangala village. (Bell 1896:9) H.C.P. Bell’s most important and splendid achievement in all his explorations at Sigiriya is his comprehensive plan of the Sigiriya complex, in which he has meticulously plotted not only the features of the western sector, but all the features of the Eastern Precinct (Bell 1907). He assumed that the main entrance to the ‘city’ area of the Sigiriya complex is from the west (Bell 1899:7). After Bell and up to the 1980’s, the attention of researchers was mainly focused on the northern, southern and western sectors. The investigations begun under the Cultural Triangle from 1982 onwards, have provided a basis for the study of the Eastern Precinct and its functional activities as part of a total concept of urban form and function (Bandaranayake 1984). On the basis of data gathered from these 1982 and post-1982 explorations, the view that the Eastern Precinct at Sigiriya could have been an inner city area closely associated with the royal palace complex, was reiterated and began to be tested (Bandaranayake 1984). To confirm or reject this view, it be­ came necessary to carry out extensive surveys and ar­ chaeological excavations in this area. Several preliminary surveys were made in the early 1980’s, but a .concerted pro­ gram was began in 1988, with the present archaeological re­ search activities under SARCP (Solangaarachchi 1990:107). Figure 11:1 Pillar structure in Eastern Precinct during excavation. Seen from south. Photo: Priyantha Karunaratne. This paper presents the results of surveys carried out in the eastern sector of the Sigiriya urban complex. A number of scholars have researched the urban planning system of Sigiriya and expressed various views, but very little attention has been focused on the Eastern Precinct and the eastern sector. Studies of the ruined buildings and the garden complex are scanty, making it very difficult to develop a comprehensive interpreta­ tion of the urban plan at Sigiriya (Bandaranayake 1990:41-56). Although the architectural and garden precincts pertaining to the residences and activities of the king, the royal family and the court are clearly identifiable and have been extensively investigated, our knowledge regarding the settlements of the courtiers and officials who maintained the kingdom and of the other social groups who constituted the wider urban popula­ tion, is almost non-existent. This would be a very important aspect in the understanding of the social organisation and set­ tlement pattern of Sigiriya and the Sigiriya region, hence the necessity to investigate the function served by the eastern sector in the formation of the Sigiriya complex. It was to explore such problems that the present study was launched, with the following objectives in mind: 1. to carry out a broad survey of the Eastern Precinct at Sigiriya and its environs^ in order to document its extent and its inherent features. 2. to study the function of the Eastern Precinct within the total spatial organisation of the Sigiriya complex. 3. to examine the interpretations made in respect of the Eastern Precinct and to understand its settlement pattern and social character arising from such interpretations. Most of the data required to fulfill these objectives was obtained from field research. In addition to this research and as a part of this study, data gathered from the surveys and excavations carried out by other units of the ‘Sigiriya-Dambulla Settlement Archaeology Research Project’ was used to find at least tentative answers to some of the problems highlighted above. A comparative study of the opinions expressed by early research workers on Sigiriya has also been attempted. And the importance of the Eastern Precinct in the context of the Sigiriya complex has been noted by reference to books con­ taining the writings of ancient authorities such as Kautilya, Manu and Yagnavalkya, who gave theoretical guidance in the construction of ancient royal complexes (Begde 1982:3-32). A considerable number of persons participated in field surveys and excavations. The fieldwork was carried out in 1988,1989 and 1990, under the SARCP program. Present research These studies were carried out in three stages, in 1988, 1989 and 1990. The Eastern Precinct was divided into three zones : SEP 1, SEP 2, SEP 3 (see fig. 11:2). The first season was from September to October 1988. Owing to the situation that prevailed in the country during this period, this work could not be completed. However, research and surveying of the zone marked SEP 1, located close to Sigiriya rock, was completed (Solangaarachchi 1990:106- 112). Extensive research activities in the eastern sector were commenced in the second field season, from September to October 1989. Research work in the regions SEP 2 and SEP 3, which could not be completed in 1989, was accomplished in 1990. This also included a contour survey of the Northern Gat­ eway, a plan of the pillared building immediately to the north of the gate, and excavating four 2x2m test pits at selected locations in SEP 1 and SEP 2. Further, during the field research period of 1990, research activities were launched in the region of Sigirivava and in the area of the rampart running through the Mahavava and the stretch of land lying between the eastern entrance and Pi- durangala rock. Generally speaking, as a result of such research earned out during these three years, we had a clearer understanding of the eastern sector of Sigiriya and its environs. Exploration As explained above, research was carried out in several stages. The area was divided into several sections, each sec­ tion being subjected to intensive research. The principal tech­ nique followed was field-walking. Aerial photographs, one-inch maps and plans drawn by earlier research workers were studied before exploration activity was undertaken. Field research included collecting surface samples, par­ ticularly pottery remains and other debris, since such finds help us to understand ancient settlement patterns. Plotting rough sketch plans, based on the data collected from field research, was an important aspect of the documentation. We can consider this work a preliminary attempt to lay out the main strategy of our research. SEP 1 Although the Sigiriya Eastern Precinct (SEP) was divided into three main field areas in 1988, SEP 1 was again sub­ divided into eight parts and named SEP1A-H (Solangaarach- chi 1990:106-112). For the purpose of the present study, a new division was found necessary. The main field areas, SEP 1, SEP 2 and SEP 3 were left unchanged, but the eight sub-divisions, SEP 1A-H, here stand for separate units, differing in extent from those in the 1988 field project, (see fig. 11:2) Site A This zone includes the Lion’s Paw platform and stone terraces situated at different elevations below it (see fig. 11:2). The 1988 research report records the terraced walls on the eastern slope of the Lion’s Paw platform (Solangaarachchi 1990). Interpreting the function of these walls, which are scar­ cely visible on the surface and badly eroded at places, was one aspect of this study. It is presumed that as on the western slope of the Lion’s Paw platform, garden features would have been added to the northern and eastern sides, to maintain the aes­ thetic balance and symmetry of the landscape and building complex. The remains of terrace walls on the eastern slope indicate that terraced gardens, similar to those restored on the western slope, had existed on the eastern side too. Today the only available approach to the Lion’s Paw plat­ form is along the Mirror Wall staircase. However, the struc­ tural remains that have been discovered help us to conjecture that there once existed other means of access from the north and east, to the Lion’s Paw platform. Site B The area with large terraces made of gneiss rocks lying on the southern slope of Sigiriya rock has been marked Zone ‘B’. Here, as below the Lion’s Paw platform in the south, three terraces have been constructed, dropping in elevation in the direction of Sigiri Mahavava. A section of these terraced walls has been so constructed that the boulders lying on the Eastern Precinct have been incorporated in them. These terraced walls are in a better state of preservation than those below the Lion’s Paw platform. Although less importance has been attached to the terraces of the southern slope than to those of the Lion’s Paw platform of the royal complex, it is clear that these ter­ races must have been a part of a well-planned royal garden. Site C The region between the areas marked A and B, sloping to the east of Sigiriya rock, has been marked ‘C’. The section of the rampart in the Eastern Precinct that curves towards Sigiriya rock is the only part of the original construction that remains. Debris fallen from the buildings that were once situated on the summit of the rock are found in plenty in this area. Site D This is the area that was marked SEP 1G and SEP 1H in the 1988 campaign (Solangaarachchi 1990). The area sur­ rounded by the curved rampart, called ‘the Middle Rampart’ in that report, was marked ‘D’ in the present research. The two gaps in the curved rampart, or at least one of them, may have given access to the Lion’s Paw terrace. When we consider the construction and layout of ramparts in the eastern approach, a resemblance is shown to the rajan- gana rampart in the western sector. However, the boulders within the middle rampart of SEP 1 (Solangaarachchi 1990:106) are not marked by cut marks for brick work, for buildings, as found on the boulders of the western slope. SEP 2 Site E This is the area through which the gateway leading north from the eastern approach was located (see fig. 11:2). It can be assumed that as much, or more importance was attached to this entrance as given to the now restored northern gateway of the Western Precinct, facing the present road leading to Piduran- gala Vihara from the Sigiriya complex. To the left of the northern gateway of the Eastern Precinct are the stone pillars of a square building constructed across the moat. This has been included in Bell’s plan of 1907. According to his report, a seven-hooded cobra sculpture was discovered near this build­ ing (Bell 1896:9). Architecturally, this building has a remark­ able singularity compared to other buildings in Sigiriya, in that it is the only building with stone (gneiss) pillars in the whole of the (Sigiriya) complex. What function this building served isuncertain. The cobra stone could lead to the conclusion that this building may have been used for some purpose connected with water, or served a security purpose. Since a large amount of potsherds from different types of vessels were scattered around the building (container vessels, a few lids, spouts etc.) and a few pieces of terracotta frag­ ments, which could not be identified, were also found, this building may also have functioned as a venue for religious or cultural rituals. Site F Central building in the Eastern Precinct A remarkable feature in the Eastern Precinct is a long rock with square post-holes. The post-holes form a rectangle ap­ proximately 80x10m, which is evidence of a pavilion having stood on this rock. This building, built on a rock surface, must have had timber pillars to support the roof. According to the positioning of the post holes, it may be presumed that this building was rectangular in shape, placed in a north-south direction, facing the east. The roads leading from the northern, southern and eastern gateways, which seem to have been the main entrances to the Eastern Precinct, all seem to have met in proximity to this building, as can be seen in the conjectured plan of the road­ ways (see fig. 11:4). Site G During explorations, the question was raised whether the present Sigiriya-Habarana road passed through an earlier gap in the southern rampart, or had an opening been cut to allow access? Blakesley (1876) in his plan of Sigiriya, depicts a gap in the southern rampart through which a footpath to Pidurangala village is shown. Bell (1907) has made a very comprehensive plan of the Sigiriya complex. Here too, this footpath to Pid­ urangala is shown. Neither Blakesley nor Bell identifies the break in the rampart through which the footpath runs as a gateway. When the current road map (Survey Department) was su­ perimposed on Bell’s map, the present Sigiriya-Habarana ma­ in road coincided with the Sigiriya-Pidurangala footpath. S crutiny of the Sigiriya complex plan shows that the gap for the road lies immediately opposite the northern gate of the Eastern Precinct. So it could be assumed that the roads passed through an existing gap in the rampart. Hence there is a great pos­ sibility that this gap would have been the southern gate of the Eastern Precinct. SEP 3 The flat land surrounded by ramparts lying to the east of the Sigiriya Habarana road is the site marked SEP 3. This area is not heavily forested now, having been subject to chena cultivation a few decades ago. Therefore unobst­ ructed observation of the ground is possible through the sh­ rubs. There is no evidence of any settlements or buildings on the surface. By digging a few test pits one should be able to have a clear understanding of the area. The main and only gateway (J) to the Sigiriya complex from the east is located here, exactly at the centre of the Eastern Precinct. SiteH Site ‘H’ is the area situated outside the Eastern Precinct, between the northern rampart of this precinct and the outer rampart of the Sigiriya complex (see fig. 11:2). The above mentioned pillared building is located next to this area. In our survey we discovered potsherds strewn on the surface, at va­ rious points. The road leading from the northern gate of the Eastern Precinct towards Pidurangala would have run through this area. Although no evidence of any structure could be found, this survey revealed definite evidence of settlements in this area. Site I A large area to the east of the gravel road, leading to Pidurangala from the northern gate of the western approach to the Sigiriya complex, lying between the outer rampart of Sig­ iriya and Pidurangala rock, has been marked site T’ (see fig. 11:2). At this site there is evidence of the northern outer rampart that starts at Vilpitaravava. This rampart runs east-west, past the Apsara camp site and merges into the ground and disap­ pears here. It could be construed that this outer rampart hadnot been completed and so what is seen at the site is the uncompleted eastern end of the wall. In the region outside the outer rampart, potsherds and iron slag deposits were found at several places, extending to Pi- durangala. This shows that at some point of time there had been a large settlement in this region. Considering this loca­ tion and its connection to the Sigiriya complex, one can at­ tribute this settlement to have been contemporaneous with the Kasyapa period. Sigiri Mahavava region Sigiri Mahavava, situated to the south of the eastern approach to the complex, is the main feature of the irrigation system of the ‘Sigiri Bim’. The 5km long bund gives an indication of the capacity of the reservoir and the well-planned irrigation system that supported the agriculture-based economy of Si­ giriya. According to Myrdal, it may have been three tanks of smaller extent which were later connected (Myrdal: ’’The Archaeology of Irrigated Agriculture", this volume). This tank also fed the surrounding moats and ornamental ponds of the pleasure gardens (Ellepola 1990:205; Myrdal this volume). The rampart that runs through the Mahavava presents a conundrum. It runs east-west for about 2km through the vava and suddenly takes a 90° turn to the north, and extends for nearly a kilometre. This raises many questions. What purpose did the rampart originally serve? Why was it located within the Mahavava? Which was built first, the vava or the rampart, or were they part of some more complex planning? Myrdal points out that the canal bund which runs from Vavala to­ wards the southern rampart of Sigiriya may have supplied water to the outer moat (Myrdal this volume). It must be noted that this rampart is of like proportions to now existing outer ramparts and further, that this wall too disappears into the ground, like the incomplete outer rampart (see under site G above). When the lines of these two ramparts are superimposed on a plan of the complex, it is possible to trace a point where the two incomplete walls would have met, as the nothem outer rampart meets the western outer rampart at Vilpitara (see fig. 11:2). It must be noted that the conjec­ tured extension of the rampart within the vava, towards the west, would not connect to the rampart in the south-western section of the complex. The conjectured extension would end up about 40m north of the south-western corner (see Myrdal this volume). Excavations After the Eastern Precinct of Sigiriya was subjected to exten­ sive research, a better understanding of the structural layout of the region became possible; and a rough idea of how the Eastern Precinct and its environs had been utilised took sh­ ape. To understand the function of the Eastern Precinct, survey activities alone were not enough; therefore test pits, 2x2m in size, were dug at Aligala, meda pavura (central rampart), the ‘Mica’ and Apsara sites situated within the area surveyed (see fig. 11:4). Aligala excavations This site is located in the area marked SEP 1. A few frag­ ments of pottery that could be dated to the Kasyapa era were found in the first and second strata. Strata beyond these ye- ilded evidence of both prehistoric and protohistoric habita­ tion. Although this site stands within the citadel, the curious fact emerges that the prehistoric strata has been left undis­ turbed through the Kasyapan phase (see Karunaratne and Adikari this volume). Mica site excavations This excavation site was located in the SEP 2 area. Accord­ ing to the evidence gathered from excavations (see fig. 11:4) there has been no settlement in this area. Cultural evidence was found only from the third stratum and that too, may have been deposited by water drawing down from higher lands (Karunaratne unpubl.). Excavation of the central rampart This excavation site is located in the steep slope below the Lion’s Paw terrace, close to the Aligala site. All strata of deposits have been formed as a result of redeposition by erosion. Cultural artefacts that were found (pottery, pieces of lime mortar, pieces of brick) seem to have slipped down from the direction of the Lion’s Paw terrace. Hence the site will not be discussed further (Sirisena unpubl. b). Apsara excavation This site is located in the area marked H (see fig. 11:2). Important evidence of settlement belonging to the Kasyapan era was uncovered in this region. Of the seven contexts, evidence of settlement was found in contexts 1-3. All belonged to the same era. First stratum yielded 10kg of tile pieces, while the second stratum yielded 5 iron nails and 80kg of tile pieces. From the third stratum we found an iron nail, a copper coin (which can be presumed to be a Roman coin) and a piece of resin-coated buff ware which, perhaps, would have belonged to an imported wine jar (see fig. 11:6) (Sirisena unpubl. a). It is of interest that settlement evidence was found at the Apsara site, which is outside the Eastern Precinct, but no traces of buildings belonging to the Kasyapa era were found in any of these three pits in the Eastern Precinct. Although this could actually be a special feature, it does not mean that there were no buildings in the Eastern Precinct. Considering the extensive area of this region, data collected from three test pits of 2x2m size would be inadequate to determine the whole of the Eastern Precinct. However, using the data collected from the exploration and excavation and comparing it with other areas in Sigiriya, we may be able to express some opinion about the Eastern Precinct. What was the utility of the Eastern Precinct and the area adjoining it, as parts of the Sigiriya complex ? During the explorations of the Eastern Precinct, the hypo­ thesis was put forward that the Eastern Precinct was probably a garden and not a settlement area. It was also thought that the area between the outer rampart and the boundary of the Eastern Pre­ cinct was the inner city, occupied by the nobility and other social elite. But the evidence from the excavations at the Eastern Pre­ cinct was insufficient and inadequate to confirm these hypoth­ eses. Yet again, no evidence was found to confirm the concept of earlier scholars that the eastern precinct was the inner city.Going by the evidence discovered from explorations and considering the internal plan of the eastern approach, one could deduce that the eastern approach would have been used as the main entrance of the city of Sigiriya.The post-holes that mark an elongated pavilion, on a rock located in the middle of the SEP 2 area - centre of three gateways into the Eastern Precinct - resemble a garden building rather than part of a built-up urban area. It would not be wrong to assume that the Eastern Precinct too, was ado­ rned with beautiful ornamental gardens. It is clear that the roadway up to the rajangana rampart (citadel wall) across the Water Garden, was built right through the centre of the Western Precinct. It can be assumed likewise that the roadway starting from the eastern gate of the Eastern Precinct bisected the precinct. The Culavamsa (Ch.39:v.9) comments that the entrances to the city of Sigiriya had been adorned with parks. Bell and Paranavitana and Bandaranayake (1984:5) identify the Western Precinct as a royal park. Taking these records and investigated evidence into ac­ count, we see in the Eastern Precinct the same motifs of a rampart to wall off the precinct, a road bisecting the area (see above), with a pavilion which spans this centre roadway, and significant signs of yet uninvestigated ramparts; and we can assume that the Eastern Precinct too must have been an or­ namental royal garden. Inner city The inner city of a royal complex should consist of roadways, markets, conference halls and other buildings used for com­ mon purposes, like religious centres etc. (Begde 1982). Con­ sidering the extent of the Sigiriya royal complex, its inner city appears to have been spread over a considerable area. The construction of ramparts outside the Eastern Precincts could have been an attempt to expand the city complex. Thus it can be inferred that the structural plan of the inner city of the Sigiriya complex, or rather the inner city as conjectured by the writer, was located outside the Eastern Precinct, but within the two outer ramparts. The presence of a large amount of potsherds that could be dated back to 5th century AD, from the surface of site H (see fig. 11:2) points to the existence of a contemporary settlement. Apsara excavations confirm masonry of buildings in this area (Sirisena 1989 a). The discovery of fragments from imported wine jars, buff coloured on the outside and coated with a thick black resin inside, points to this site being an area of elite habitation. The Smutisastra, Nithisastra, Silpa Sastra and Purana record that nobles and the elite resided outside the royal complex, but within the inner city. The remains of a stone-pillared building near the northern gate of the Eastern Precinct and the potsherds found close to it, imply that this was an upper class residential area. According to Kautilya, the residences of the royal teach­ ers, ministers and clergy and the place where religious fes­ tivals were held, had to be situated north-east of the royal palace (Begde 1982:15). Considering the facts, it is justifiable to assume that the inner city of the Sigiriya complex was located not within the Eastern Precinct of Sigiriya, but outside the precinct, surrounded by the outer ramparts. Sigiriya complex in comparison with city plan layout in Indian literature In studying the theoretical concepts of ancient Indian city architecture, it is an interesting question whether Sigiriya architects also used such classical works. This will also help us to solve the basic problem in our research. We find a number of Indian classical works by such em­ inent scholars as Manu, Yagnavalkya and Kautilya, dealing with royal and religious architecture. In those works, ancient royal complexes, which included palaces and fortresses, were categorised according to their security plans. These have been termed giridurga (protected by rock), pankadurga (protected by mud), jaladurga (protected by water), wanadurga (pro­ tected by jungle), prakaradurga (protected by rampats) etc. In this manner the necessary structural features have been des­ cribed in detail. These works must have influenced the ar­ chitecture of India and Sri Lanka in no small way. The Sigiriya complex was adorned with ponds, moats, fountains and tanks and has well-planned micro-hydraulic sy­ stems (Ellepola 1990:169-227). According to Vastu Sastra by Visvakarma, "summer palaces should be built in such parks, with halls cooled by ponds around such palaces" (Begde 19- 82). Halls cooled by ponds and summer palaces have been discovered within the Western Precinct of Sigiriya (Bell 18- 99:8; Bandaranayake 1990:47). When we collate the data relating to various architectural features discovered in surveys of the Eastern Precinct, it can be presumed that several such buildings must have existed in this Eastern Precinct. The central point of the Sigiriya complex is the royal palace. This is in keeping with Visvakarma architectural con­ cepts, according to which the royal palace has to be built at the centre of the city and other buildings located around it in a concentric pattern. According to Kautilya, the residences of the royal teachers and distinguished persons of the court sho­ uld be housed in a north-easterly direction from the palace. Evidence from the Eastern Precinct shows this area to have been occupied by an elite group (see above). In Indian architecture, the royal gates adorned with lion figures were called lion gates (sinha-dvara) (Bedge 1982). The Culavamsa description, says that Sigiriya had several staircases adorned with lion figures (Ch.39:v.3). Probably the term ‘Sigiriya’ originated from the ‘Lion Gates’. According to recent research on the park gates in Sigiriya’s Western Prec­ inct, opinions have been expressed about the lion figures ador­ ning them (Gunawardana unpubl.). If there had been such lion gates they would have been placed not only in the western sector and the Lion’s Paw yard, but also at the entrances in the Eastern Precinct. Taking these facts into consideration, there is no doubt that the Sigiriya architects were familiar with the Indian classical sources, but in formulating their plans and designs it is clear that the local architects did not slavishly follow the rules and concepts of the Indian masters, but drawing inspiration from them, used their own knowledge and talent to preserve the identity of the indigenous architecture. A study of the architecture of the Boulder Garden shows interesting parallels and correspondences with Chinese archi­ tectural rules and concepts of integrating the natural features of the landscape into the final layout (Bandaranayake, per­ sonal comments). Thus, the master architects of Sigiriya were familiar not only with the architecture of the Indian sub-con­ tinent, but knowledgeable about other sources - and so must have had an all-round education that helped them to master their art. The concept of the outer city Bandaranayake in his model of the EMHP city has defined an outer and suburban area (Bandaranayake 1991). From the explorations done outside the ramparts, it was evident that the outer region was inhabited. Large quantities of potsherds were discovered on the surface. These sherds provide evi­ dence that people who rendered services to the Sigiriya com­ inhabited by many. It is probable that people belonging to the urban social strata settled in areas close to the royal complex, outside the rampart (Mogren :"The Archaeology of Talkote" this volume). This would be the conceptual ‘outer city’, or suburban area, which was not secured by ramparts. Conclusion Data collected from the surveys and research carried out in the Sigiriya Eastern Precincts have been reviewed above. Here we discuss the concepts formed as a result of those reviews. Speaking of the layout of the ancient capital of Anuradhapura, Bandaranayake (1991) conjectures that the royal court lived within the citadel, or rajangana wall. Be­ yond this wall was the inner city, bounded by the inner city wall and the rajangana. The area outside the inner city wall was designated the ‘outer city’. A rampart enclosed this outer city area. In discussing EMHP urban planning, Bandaran­ ayake has referred to the "concentric organisation of social space". A deviation from this is seen by the present writer in regard to the Sigiriya complex. According to views based on SARCP research regarding the composition of a royal complex, there has been an inner city between the inner moat and the outer rampart. Outside the outer rampart lay the outer city precincts (see fig. 11:7). The main deviation appearing in the new structural sketch (presented through current research associated with the Si­ giriya complex) is that the outer city of the Sigiriya complex is the area outside the ramparts of the royal complex (see fig. 11:7). This seems to have been the beginning of a new innova­ tion in the context of spatial allocations of a royal complex. Thus, this new spatial arrangement had been: the royal palace on the summit of the rock, a forecourt or park that until now had been identified as the inner city, the Boulder Garden and the Water Garden on the west, all surrounded by a rampart. The hypothesized inner city falls between the two ramparts side of the complex - this area up to now had been identified as the outer city. It is now clear that the outer city lies outside the outer ramparts (see fig. 11:7). At Sigiriya, the small tanks lying in the periphery of the citadel, outside the proposed outer city, would have been in operation during the 5th century AD. Thus, it is clear that the density of population gradually diminished in proceeding fr­ om the outer city region, largely giving place to rural settle­ ments. This is a general conclusion that could be arrived at, based on contemporary research of the Sigiriya royal complex. In order to assess the success of the research so far carried out, it would be necessary to subject the Eastern Precinct and the outer city, or satellite suburban settlement of Sigiriya, to fur­ ther archaeological excavation. REFERENCES Bandaranayake, S. 1984. (ed.). Sigiriya Project, First Ar­ chaeological Excavation and Research Report. Colombo: Central Cultural Fund, Ministry of Cultural Affairs. Bandaranayake, S. 1990. Sri Lanka’s contribution to the stu­ dy of Asian Garden History: the Royal Gardens at Sig­ iriya and Anuradhapura. Ancient Ceylon. 10: 41-56. Bandaranayake, S. 1991. The Morphology and Spatial Or­ ganisation of the Premodem City in Asia: the ‘First’ and ‘Second’ Urbanization in Sri Lanka. Paper presented at the 12th Conference of the Association of the Historians of Asia. Hong Kong, June 1991. Begde, P.V. 1982. Forts and Palaces of India. New Delhi: Sagar Bell, H.C.P. 1885, 1896, 1897, 1898, 1899, 1901, 1906, 1907, 1908, 1909. Archaeological Survey of Ceylon An­ nual Report (ASCAR). Blakesley, T.H. 1876. On the Ruins of Sigiriya in Ceylon. Journal of the Royal Asiatic Society, Ceylon Branch. 8. Vol.I and II. Culavamsa: Being the more recent part of the Mahavamsa. Geiger, W. (ed.) 1925,1927. London: Pali Text Society. Ellepola, C. 1990. Conjectured Hydraulics of Sigiriya. An­ cient Ceylon. 77,vol.5: 169-227. Gunawardena, P. Sigiriye Doratu (Sinhala). Unpublished B.- A. dissertation, Department of Archaeology. University of Kelaniya Archives. Cat. No. 25. Karunaratne, P. Preliminary Report on Mica Site Excavation, 1989. Unpublished Excavation Report. PG1AR Archives. Cat. No. 1989/1-10. Sirisena, M. (a). Preliminary Report on Apsara Excavation. Unpublished Excavation Report. PGIAR Archive. Cat. No.1989/1-10. Sirisena, M. (b). Preliminary Report on Middle Rampart Ex­ cavation. Unpublished Excavation Report. PGIAR Archi­ ves. Cat. No. 1989/1-10. Solangaarachchi, R. 1990. Preliminary Explorations in the Eastern Precinct. The Settlement Archaeology of the Sigi­ riya - Dambulla Region. Bandaranayake, S., M. Mogren and S. Epitawatte (eds.). Colombo: PGIAR.

W.A. KUMARADASA Figure 9:1 Great retaining wall around Mapagala on the western side of the rock. Seen from the north. Photo: Mats Mogren INTRODUCTION About 400 metres south of the Sigiriya rock are two medium sized hillocks, covering an extent of about 17 acres, which today are known as ’’Mapagala”. The site covers 250,000m2, inclusive of adjacent areas around the rocks. It is delineated by the Sigiriya complex on the north, the Sigiri Mahavava on the east, Petangala, the rest house and Kayamvala village on the west and the Sigiriya Oya on the south.The twin rocks which are called Mapagala North and So­ uth, are about 11 acres and 5 acres in extent respectively. The North rock differs from the South rock in that its summit abounds with natural boulders, whereas the South rock is devoid of such boulders. The surface of the northern summit is a flat area where brickbats and fragments of a small, dil­ apidated stone wall are visible. There is also a natural pond with a gradient from north to south. No other structures are visible today (Bell 1908, mentions the remains of a brick structure), but potsherds, brickbats and pieces of tiles are fo­ und. Pieces of quartz are found in plenty in the soil layers upon the summit. Potsherds (Black-and-Red Ware), small quantities of iron slag and rare flakes of chert are also to be seen. On the South rock, which is covered with thorny jungle, there are large quantities of brickbats. The ruin of a structure of unidentified function has some bricks still in situ and the remains of a dilapidated retaining wall are also to be seen. There are signs of a bridge-like road spanning the pass con­ necting the two rocks (Bell 1908). This road had been con­ structed with two lines of stones arranged on either side of an existing stone slab. Traces of retaining walls, which have been subject to erosion, may have been used as flights of steps leading towards the west. Although clear signs of such a flight of steps leading east are not visible, traces of a path are discer­ nable. Traces of some sort of constructions are to be found tow­ ards the east of the North rock, the west of the South rock, and among the boulders. Some steps subjected to erosion, which can be considered the ancient access road to the North rock, are found opposite the southern face of the Sigiriya rock. There are six small, natural caves in the rocks. Bell refers only to one cave on the eastern front. Although dripledges have been made in some caves, they look very rough com­ pared to those in Sigiriya. No inscriptions are found. Potsherds, brickbats and other cultural finds are visible. Fragments of Black-and-Red Ware bowls were found in some of the caves. In cave No. 2 there were remains of a wall, built of small pieces of stone, plastered with clay mortar. It seems these caves had been used on several occasions. The construction that makes Mapagala unique is the ter­ race wall built around its base. For the main wall, which is up to 4m high and about 2m wide, undressed stones of a size about 2.5x2.5m have been used. The construction, made by connecting natural boulders, has added strength to the wall (Chandralatha unpubl.). The stone blocks used in the construc­ tion of this wall are much larger than those used in the walls of Sigiriya. Even though the wall may look primitive, having been built of undressed stone blocks, it is undisputable evidence that the builders of the walls at Mapagala had been masters of an advanced technology; and controlled a mass labour force. Theie is a second, lower wall running parallel to the main wall, on the west of the North rock. Bell (1908) says this may have been either a street, or a path. In the same area, in the vicinity of the Archaeological Department’s circuit bungalow, there are several lines of stones which seem to have belonged to yet another small wall. West of the Mapagala rocks is a smaller rock, with bou­ lders running north-south. South and south-west of this rock formation is an iron production site (SO. 66). To the north­ west of this small rock is a rock called Petangala. Ruins of a stone foundation are found on its summit. Unlike the Sigiriya gardens, it is very rarely that cut marks for post holes or brick constructions are seen on the rocks of the Mapagala complex. The tank bund of the Sigiri Mahavava runs from the sou­ thern end of the foot of Sigiriya rock up to Mapagala, and from the southern end of Mapagala up to the Kandalama mountain range and further (see new data in Myrdal: "The Archaeology of Irrigated Agriculture" this volume). The present Sigiriya village is situated on a section of the ancient tank floor. The tank as it is today, situated north-east of Mapagala, merely looks like a large pond, but the ruined tank bund indicates that the Mahavava had once been a macro-scale tank (Blakesley 1876; Bell 1906-1908). The bund between Sigiriya rock and Mapagala has had a roadway upon its crest. There are signs of repairs effected to it from time to time (De Silva, R. 1967/68, 1968/69). THE HISTORICAL EVIDENCE Chronicles Historical sources regarding the history of Mapagala are rare. However, judging from the ruins and artefacts found, it is clear that this site is a place where human activity has taken place during various periods of time. It is not unreasonable to consider Mapagala, Sigiriya and Pidurangala as a single ar­ chaeological complex. These rocks can neither be separated from one another in the research process, nor can they be considered in isolation. Thus Mapagala is part of the history of Sigiriya. From various sources it is evident that royal patronage was afforded to this area in the Early Historical period (Ranawella 1984; Seneviratne, A. 1983:65-66; Ilangasinghe 1987:7-8). The region had witnessed important state activities since the earliest Anuradhapura period, when the Ritigala fortress con­ tributed to the siege of Upatissagama by King Pandukabhaya. According to Geiger, places like Giri Laka, Kasa Pabbatha, Thulaththa Vapi, which are connected to the battle between Dutugamunu and Elara, must have been situated in an area between Sigiriya and Minneriya (Mahavamsa, commentary by Geiger 1950: 200-291). The area also formed an important link in the ancient road system connecting Rajarata with Malaya (see for example Wimalaratne 1984:19). These roads became more important during the Sigiriya Kingdom in the 5th century, in matters of transport and external relations (Wimalaratne 1984). Historical sources reveal constant movement of forces towards the inte­ rior of the country and that there existed a communication network among the population for defence purposes. Inscriptions A large number of inscriptions have been found from Sigiriya and the surrounding areas. (Ranawella 1984; Somadeva this volume). This collection of inscriptions, which is a positive indication of the existence of ancient monasteries, makes it pos­ sible to generalize on the extent of settlements and development of population during the period, since the patronage of a self-suf­ ficient people is necessary to support a large monastery com­ plex. Some of the Early Brahmi inscriptions mention the title parumaka. Although information of elite social groups who were given this title is sparse, the extension of that class and their activities show that they were a privileged class in so­ ciety. This was a traditional title which could be used for both males and the females alike, meaning a ‘leader’, ‘head’, or ‘a senior person’ (LiyanagamageT965:194). A social system of aristocracy, based on lineage, or a group of people in the ruling class who held important positions in society, is envisaged by this title. Village leaders who enjoyed some power in their respective villages, high officials who held titles of senpathi, badagarika etc. and property owners can also be considered to have belonged to this group. Paranavitana is of the view that they were the group of people who assisted the king in governance of the country and implementation of economic activities of the land (Parana­ vitana 1970). Indrapala contends that the word parumaka is a derivation from the Tamil word parumakan, and that they may be regional leaders who were descendants of Dravidians (In­ drapala 1965:306). According to Ilangasingha, the existence of such a large number of parumakayo among those who offered caves in settlements around Dambulla, is strong evidence of the fact that the term parumaka was used to identify only the elite class and not a designation, as was the case during the Kandyan period, when the term nilame was a title used collectively for the elite class. (Ilangasinghe 1987). Apart from religious ac­ tivities, the parumakayo were engaged in tank building (To- nigala inscription), trade (Paranavitana 1970), teaching, etc. At times they were so powerful as even to be in a position to interfere in the state machinery. The above-mentioned inscriptions give evidence that there were parumakayo in possession of the area in and around Sigiriya. Although no clear information is available regarding their homes, references such as ‘parumaka naguliya lena9, ‘parumaka abijiya lena\ etc. may refer to the caves which served as their homes. Those caves were subsequently con­ verted into monasteries by their owners. However, if any con­ structions had been made, they may have been either destroyed or changed owing to later settlements; or else suffi­ cient evidence does not remain whereby they can be identified. There are several inscriptions too, found in this area, which are categorized as Late Brahmi inscriptions (l-5th centuries AD). Inscriptions belonging to the reigns of several kings who ruled during the period from the 5th - 10th centuries have been found (Ranawella 1984; Godakumbura 1964/1965; Tissa Kumara 1990). The common feature of all these inscriptions is that the references are to donations made to monasteries and patronage. However, it is this collection of inscriptions which is the strongest and most authentic evidence of the existence of settlements in this area since protohistoric times. PREVIOUS RESEARCH Although more than a century has passed since research in Mapagala was first taken up, actual archaeological fieldwork so far has been very limited. Most has been done at survey level and no further. In 1876 Blakesley evinced some interest in conducting a survey on this site. Emerson Tennent’s Re­ ports provided him with the necessary sources. He strongly believed in earlier historic constructions relating to Sigiriya and Mapagala. He further says that if the wall around Mapagala was meant for the protection of the site, Sigiriya would have of­ fered better protection in view of its location. About the wall on the east, he is of the opinion that such a wall on that side is unnecessary, in view of the fact that it is protected by water, the Sigiriya Mahavava having been built later. Elaborating his point, he says that although the necessity of water was felt for moats.and gardens etc., during the Kasyapan period, there was no necessity for the construction of a large tank, and it was during the time of King Parakramabahu (1164 AD -1197 AD) that such large tanks were built for agricultural purposes. He explains that owing to the very nature of their construction, the dams fell into disrepair on several occasions. Thus it was necessary to build large tanks to facilitate agriculture. Besides the above contention of Blakesley, his other views are rational and reliable. The historical details put forward by him do not seem to be relevant, in the light of our sources, in the context of Sigiriya or Mapagala. However, he is of the view that the original constructions at Mapagala date back to the pre-Kasyapan era, in the Early Historic period. The second scholar who did surveys on Mapagala was J.B. Andrews (1909). Drawing attention to the uniqueness of the wall at the base of the rock, he expresses the view that it may be a structure belonging to the megalithic period. He says this stone wall may belong to the Vijay a Rama period, or even to an earlier period before the structures in Anuradhapura were built. Further, while pointing out that stone walls of such magnitude, belonging to the early phase of the Roman civ­ ilization, are to be found at several places in southern Europe, he cites the example of stone walls near Pyu Ricard in Mo­ naco, referred to as ‘Ligurian’ (Andrews 1909). Although Andrew’s dating is too early, the rationality of his argument can be accepted. He contends that the constructions at Ma­ pagala may date back to the pre-5th century AD. When the finds of quartz, Black-and-Red Ware, iron slag, chert, etc. on the summit of the rocks are considered, it is not difficult to agree with him. However there is not sufficient evidence to agree with his megalithic dating. H.C.P. Bell, who paid attention to the study of Mapagala, reiterates that the history of Mapagala could not date beyond the 5th century AD at all, and emphasizes that it is not prehis­ toric. He stresses that the residence of the Mahadipada or Mapa of King Kasyapa, who built Sigiriya, may have been built there. He is of the view that the vast wall resembles the large wall of the Incas at Sacsahuman, in South America (Bell 1908). There is no reason to build the walls at Mapagala stronger than at Sigiriya, which was more important during the Kasyapan period. Among hypotheses built on the basis of the architecture of Mapagala, predominant are that this was a fortification (La­ wrie 1898; Deraniyagala, P.E.P. 1973:10; De Silva, R. 1974; Bandaranayake 1984; 1990), a monastery (Blakesley 1876), and a palace or a residence of the elite, (Bell 1906; 1907; 1908). According to the above-mentioned research, views on the dating of Mapagala can be divided into three main groups: I. A creation of the prehistoric period. II. A creation of the protohistoric period. III. A creation of the historical period. When the available data is considered, it becomes clear that features of all these periods mentioned above are to be found at various levels. However, no in-depth research has been done with a view to identifying these factors separately. Adikari (this volume) has identified prehistoric camp sites, not only in caves like Aligala (Sigiriya), Pidurangala, Dambul- la and Potana, but also in open places like Millagala and Tammannagala. He maintains that Mapagala may also have been an open area hunting ground, since there are stone imple­ ments among the collection of quartz debris found at this site. Also, the existence of a pond, or water hole, on the flat sum­ mit, strengthens this possibility (Adikari this volume). A stage of inter-relationship or interdependence between the mega­ lithic and Brahmi periods, or a stage of transition, has been revealed through studies in areas like Ibbankatuva, Dambulla and Pidurangala (Bandaranayake 1990:22-23). Particulars of the resources utilized by societies during prehistoric and pr- otohistoric periods in the Sigiriya and Habarana areas have been confirmed. (Seneviratne, S. 1990). Interpretations of the name Mapagala How and why this place was called Mapagala cannot be established with certainty. It may be a corrupt form of another word in use for generations. The name Mapagala is not men­ tioned in written historical sources, but by the time foreign travellers like Bell, Blakesley and Andrews started visiting the area, these two hillocks were known as ‘Mapagla’. Various interpretations are given to the genesis of this name. It has been established by Blakesley that the maha paya of King Kutakanna Tissa, and other constructions, are among some of the ruins found near Mapagala. Blakesley is of the view that ‘Mapagala’ was derived from the words ‘Maha- pay a-gala’ or the ‘Great Precipice Rock’. (Blakesley 1876). The precipitous drop on the west and the south of these two rocks lends evidence to this conclusion. But when one thinks of Sigiriya, which has a steeper slope, one begins to wonder why a small rock like Mapagala was thus named by the people in ancient times. This has to be further examined. The "gala" or rock where the viceroy (Mahadipada or Mapa) resided, may have been called Mapagala. Bell (1886) emphasises that the palace of the viceroy may have been built on this rock by the time Sigiriya became the seat of the kin­ gdom (477-495 AD) and so later came to be known as Ma­ pagala. Although the term adipada, or apa (viceroy) is an equivalent of ‘heir to the throne’, very often all royal princes were known as adipada, or viceroys. This designation is first mentioned during the reign of King Silakala in the 6th century. The term Mahadipada - Mahapa - was a ceremonial title conferred on the ‘heir to the throne’. In chronicles these titles are mentioned during the time of King Dathopatissa (650 AD) (Geiger 1969; Liyanagamage 1965). King Dhatusena appo­ inted his nephew (Migara Senevi) as his commander-in-chief. Migara served King Kasyapa too, in the same capacity. There­ fore it can be considered that King Kasyapa had a viceroy, or mahadipada, in keeping with the tradition of the historical line of kings. It can be surmised that subsequently he was known as 'mahadipada' in folklore. The name of the residence of this mahadipada may have lent itself to the place under survey. A structure found near the Basavakkulama tank, in Anur- adhapura, has been identified as a mapa palace. This is con­sidered the best example of a building of the Lankan elite class (Godakumbura 1976). The structural features on the summit of Mapagala, identified by Bell, closely resemble a large st­ ructure, with cells, passages and verandahs, of a summer pa­ lace upon the summit in the ancient city of Sigiriya. (Bell calls them islands: Bell 1908). A large number of ‘attam inscriptions, which contain in­ formation regarding the privileges and benefits enjoyed by mapas during the later centuries of the Anuradhapura period, have been found in surrounding areas. They are at Vayaulpota (Sena KK, 853-887 AD); Mihindal Mahapa, Ramakale-Ma- pagala (Kasyapa IV, 898-914 AD); Kasabal Mahapa, Pid­ urangala (Dappula IV, 924-935 AD); Uda Mahapa (Ranawella 1984, 203-209). Thus the word Mapagala may have come into use after the titles of viceroy, or mapa. The fact that written evidence which contains the names of these titles found in abundance in this location, leaves room to surmise that this area may have served as a centre of ad­ ministration, controlled by holders of the title mapa. It can also be considered that subsequently this name found its way into common parlance in the belief that the ‘attam' pillars in Mapagala are those of the mapa belonging to the period of King Kasyapa. THE SARCP WORK Within the framework of SARCP, fieldwork on Mapagala was commenced in 1989, when test excavations on the sum­ mit of Mapagala were carried out. This was followed by the excavation in 1990 of a trench transecting the main wall, at the foot of the western slope. Questions, hypotheses and objectives Mapagala can be presupposed to be one of the best places in the region, next to the Sigiriya complex itself, for a study of administrative and socio-economic circumstances in ancient times. At the outset of the investigation, several major ques­ tions regarding Mapagala were taken into consideration. Wh­ en were the terrace walls constructed? What was their function? Was the site occupied continuously, or was it aban­ doned at one time? If so, why? The fact that the construction of Sigiriya and Mapagala cannot be considered on an equal footing renders it impossible to apply the archaeological theories developed about Sigiriya in the context of Mapagala. Since there is evidence which reveals that settlements had been in existence in the Sigiriya region prior to the construc­ tion of the citadel, there is a possibility that Mapagala too, dates back to pre-5th century AD, as presupposed by earlier researchers. So far we do not have a clear idea of what purpose was served by Mapagala, with its enormous embankments; whether it served as a regional administrative centre, or the residence of local leaders, or whether it was a religious centre. The existence of quartz in abundance around the area also leads to another experimental approach. Does the presence of quartz have any connection with the activities of those people who built this wall and other constructions? Or does it have any connection with previous, or subsequent, activities? Is it waste from the production of microlithic tools on the site; or were these fragments mixed with soil brought from outside, as fill? Is the quartz debris deriving from some other kind of activity during later periods? The finds of iron slag in considerable quantity from the site and the abundance of iron slag on the west of this rock (SO. 66), gives further significance to the necessity of an ar­ chaeological study of this site. Could this slag be connected with the large-scale constructions? Besides the large outer wall, no other traditional com­ ponents of a fortress are found here. Even though, at first thought, the utility of the large wall may have been to ensure safety from external forces, it may also have served as a social marker. Anyhow, with the enormous labour input evident, it is certainly not to be considered a temporary structure. The fact that an ‘attanf inscription is found in the vicinity of this site allows the theory that the complex was connected with the ‘Mahanaga Rock Monastery Complex’, subsequent to the original phase of construction (10th century). This be­ comes further evident as subsequent monastery settlements too, are found in the caves nearby. Although it can be thought that this may have been in use as a palace, or residence of the elite in the 5th century, it has been hypothesized that its period of origin lies further back. Considerable attention was paid to the protohistoric and Early Historic elite class called ‘parumakas’ (chieftains). Early Br- ahmi inscriptions mentioning parumakas are found at four different sites in the region: Pidurangala, Potana, Dambulla and Kandalama. Thus a hypothesis was formed that the site Mapagala mi­ ght represent a multi-period history of occupation, with func­ tions changing over time: a) the site could possibly have attracted settlement previous to the construction of the terrace walls; b) the terrace wall constructions could antedate the Kasyapan establishment at Sigiriya and may represent an ear­ lier societal development, e.g. the central place of zparumaka, c) the site could have been used after Kasyapa’s reign (hence the name indicating a viceroy); and d) late post-Kasyapan occupation could hve ben of a monastic nature. Excavation sites were selected on the summit and near the wall, with a view to establishing a local stratigraphy and col­ lecting artefacts for interpretation of the dating and function of the site. Mapping Although mapping of the rocks was conducted earlier by H.C.P. Bell in 1907, it was necessary to carry out a new mapping of the site. The survey unit of SARCP surveyed not only the rocks themselves, but extensive areas around the foot of the rocks as well. Detailed plotting of the test pits and the excavation trench was also an objective of this survey (fig. 9:3). TEST EXCAVATIONS OF 1989 Two test pits, 2x2m, were dug on the summit of the North rock of Mapagala. The contextual relationship between the quartz (prehistoric implements?) and other cultural finds dis­ covered here, and the utility of the summit for a settlement were taken into consideration in selecting the location of the excavation. An area at the lowest point was selected for test pit no. 1, in view of the fact that the surface of the rock had a southern slope and that the shape of a pond was discernible at this place. The inference that the surface collection found near the location may have been remains which were washed down into this lower point, was the reason for selecting this location. This section, which is completely silted up now, is covered with iluk and mana. The surface layers consist of very fine sand and no cultural remains were found there. The surface elevation of this excavation site is 237.34m above sea level. The second test pit was dug in the north-eastern sector, which was the highest point on that front. A few pieces of stone, probably parts of a small stone wall, were found about one meter east of that point. Grass had grown very luxuriantly on this site. Throughout the site quartz, a small quantity of brickbats and pottery were found. Methods The Harris’ matrix system was adopted (Harris 1979). Ar­ bitrary sub-strata were designated when necessary. These were reported under the same context number, e.g. context 1, level 1. Each stratum was sieved using a 5x5mm sieve and sea level was reported at every 10th cm. Every stage, from com­ mencement onwards, was photographed in black and white and colour; and slides were made. The excavation was carried out using small pickaxes, trowels, hand shovels and hand brushes. Results The occurrence of artefacts relevant to our objective was very low at test pit no. 1. Both layers (see fig. 9:4) can be described as alluvial, non-cultural layers. At the end of the excavation there was no possibility of arriving at any conclusion on the basis of surface collections. Therefore test pit no.l did not yield any results. Test pit no. 2 was very rewarding (see fig. 9:5). Excava­ tions were carried out on the hypothesis that the microlithic implements spread around the surface of the layer had em­ erged as a result of the erosion of the uppermost soil layers. The soil strata were divided into three contexts. Quartz and stone implements, which are to be seen even on the surface, were found in large quantities in the third context, where they were mixed with Black-and-Red Ware. The findings can be summed up thus: 1. Microlithic imple­ ments belonging to a stone age technology. 2. Black-and-Red Ware pottery belonging to the protohistoric period. 3. The existence of Black-and-Red Ware in the last context alone is not a coincidence. 4. Features of a common economic system brought about by the mixture of implements belonging to two cultures. 5. Environmental factors in a mixed culture. The existence of microlithic implements belonging to a mesolithic culture, from the surface right down to the bottom, has given rise to some complicated conclusions. Finds of Black-and-Red Ware pottery in the bottom layer compound the issue. Similar relationships were found in excavations of settlements in Pomparippu and Anuradhapura, and there are instances of the mixing of both these economic systems (De- raniyagala, S. 1972; Begley 1981). It is also not easy to build up interpretations based on megalithic implements alone, al­ though they are found extensively. A people with a knowledge of using pottery can lay claim to an agricultural economic system and a system of minor trades, based on small village units with a household economy (Seneviratne, S. 1984). The occurrence of Black-and-Red Ware from the third context alone is not a coincidence. Although microlithic im­ plements are found spread almost everywhere, there is no Black-and-Red Ware pottery. Since this type of pottery was discovered very close to the bedrock, about one meter below the surface, it can be considered an undisturbed deposit. If this is the first instance where Black-and-Red Ware pottery was found at a point very close to Sigiriya, it may be an important incident. It is also interesting in the light of Andrew’s theory on megalithic settlements in Sigiriya and the surrounding area (Andrews 1909) In view of the above factors, it is possible to elaborate on the hypothesis: 1. That the lack of historic artefacts and occurrence of microlithic implements from the surface itself, makes it pos­ sible that historic strata have been destroyed. 2. That the remains of brickbats and tiles mark the last stage of the historic strata. 3. That the remaining strata relate to the prehistoric period. 4. That in view of the fact that Black-and-Red Ware were found in the last context, there is a possibility of the existence of a megalithic (cultural) settlement with microlithic usages. 5. That according to the natural location of the summit of the rock, there is a possibility of its having been a prehistoric hunting ground. Two test pits were excavated on the tank bund, in the southern sector of Mapagala (see report by Gunasiri, PGIAR Archive). Two major factors were revealed through these ex­ cavations: (1) Finds of earthenware pottery belonging to a settlement prior to the construction of the tank bund. (2) That the main wall of Mapagala was constructed before the construction of the tank bund. THE EXCAVATION OF 1990 A test ditch 29x2m, which covered the main wall of the western axis of the Mapagala North rock and the second wall, was selected for excavation this year. In the selection of the excavation site, attention was paid to the fact that walls on the western section were still intact, and that it was necessary to have some idea of the nature of constructions and evolution of local settlements. Methods The context matrix system was followed as the method of excavation (Hanis 1979). The area belonging to three terraces (malakas) was marked in a single line, as a ditch of 29x2m. These terraces were marked A-B-C and divided A 100-115, B 88-99 and C 86-88. The south-north axis was marked from 300 upwards, towards the north, and the west-east axis from 100 upwards and downwards, towards east and west respec­ tively. Thus, in reporting a square meter, a number was in­ cluded (300/100 etc.). Contexts which were extensively thick were divided into sub-numbers or arbitrary strata of 20cm (e.g. context 1 - level 1). Context numbers were given according to the texture of natural and artificial soil layers (colour and composition); pits, cut marks, special features (brick walls, stone walls, termitaria,etc.) etc. Accordingly, for excavation ditch A context no. 1-28, B 30-41 and C 50-53 were Results According to the study of stratification in this excavation, the strata can be divided into seven consecutive phases. They have been tentatively named according to their relation to the main ‘cyclopean’ wall. Phase I: Pre-construction phase. The excavation did not give evidence of prehistoric activity at the site. The first period is depicted in context 7 (see fig. 9:9 and 9:10), the first undisturbed cultural deposit formed on the gravel layer (context 8). It contains Black-and-Red Ware to a relatively small extent only. The nature of a settlement deposit is evident. The main stone wall (context 34) is built in a cut (context 18) through context 7. Therefore it can be considered that the people who lived before the construction of the stone wall are represented by context 7. Also, lumps of iron slag were found. Although it could be surmised to belong to the protohistoric and Early Historic period, the 14C-dating of context 7 gives evidence of a 3rd century settlement (see below). Slag has also been found in large quantities around Mapagala, especially towards the west, at site SO. 66. The dating of the latter site is yet unknown. Cut marks, which could be those made for posts, and the fill in those postholes, were given numbers 24, 25, 41 and 37. These holes may be contemporary with or prior to the period of construction of the main wall. Though this factor is not helpful in our hypothesis, our personal view is that this was done prior to the construction of the wall. Phase II: Construction phase The major and the minor terrace walls (contexts 34,33) were built after the cut context 18 had levelled the ground, and behind the main wall soil was filled to form a terrace (con­ texts 11, 10,14, 15, 12, 6, 5, 28, 23, 17, 9, 22, 21, 19, 20 and 3) (see fig. 9:9 and 9:10). Black-and-Red Ware was found sparsely from these contexts. This can be identified as the main construcion stage of Mapagala. Phase III: Occupational phase Contexts 38 and 52, below the main terrace, have been iden­ tified as occupational deposits. These are cultural layers where plain Red Ware and tiles are found in large quantities and Black-and-Red Ware only in small quantities. The pieces of tile found near the wall, on either side, provide evidence of the existence of some structure, like a roof, along the wall. The 14C-dating of context 38 indicates occupation in the Kasyapan period. Phase IV: Abandonment Contexts 2, 32 and 51 have been interpreted as a period of abandonment. No clear signs of settlements have been retrieved from these contexts. Phase V: Reconstruction phase Above the natural layer (context 2) there are traces of rec­ onstruction of the main wall, in contexts 16, 4 and 13. This probably belongs to the post-Kasyapan period and can be considered the last phase of use depicted in this excavation. No idea can be formed as to the duration of this period. According to artefacts found in context 2, it can be considered as dating back at least to the 10th century. Phase VI: Destruction This phase is represented only by the stones fallen from the main wall (context 35). Phase VII: Post-destruction phase There are traces of settlement remains also in the last strata (contexts 1, 30, 50) but the evidence is not clear enough to identify the nature of such a settlement. However, instead of distinctive features of rural and urban settlements, what is found are those common to monastery settlements. Pottery The collection of pottery discovered from Mapagala can be divided into three groups of wares. 1. Black-and-Red Ware 2. Red Painted Ware and 3. Plain Red Ware. Only a very small quantity of Black-and-Red Ware has been found. Among them there are only two identifiable rim sherds, while those remaining are body sherds. The technol­ ogy displayed is very poor, not as high in quality as those found at the Gedige, in Anuradhapura, or the Jetavana, etc. They are more like some of the earthenware bowls from the Ibbankatuva megalithic culture site. Red Painted Ware too, has been found in very limited quantity. The outer coating is red and has a composition of clay. The pieces seem to have been made of a mixture of finely sieved clay, and are similar to the Red Painted Ware found at other places in Sigiriya. This Red Ware can be divided into two types: Red Ware with a very delicate finish, and with a very rough finish. The clay composition of the first type is very fine and mainly used for very small ware. The clay content in the second type of pottery is very low and the finish not so advanced. For a study of diagnostic rim sherds found in almost all contexts, the pottery can be divided into several basic types. The speciality of contexts 7 and 38, where Black-and-Red Ware are found, is that these artefacts are similar to pottery found from sites like the Early Historic Gedige (Deraniyagala 1972), Tissamaharama (Kuna 1985), etc. These finds are limited to artefacts such as urns, plates, pots, etc. The cultural characteristics revealed through contexts 1, 2, 32, etc. show a continuity from the middle-Anuradhapura period up to the post-Anuradhapura period. However multi-period relation­ ships too, are to be found from certain contexts, i.e. context 32. Certain types represent the early Anuradhapura period. 14 C-datings Charcoal samples from five different contexts were sent to Sweden for analysis (see "14C-datings" this volume). Context 7 is a pre-constructional settlement layer; contexts 5 and 3 are fill belonging to the construction phase; context 9 is fill belonging to the construction phase, or the occupation phase; and context 38 belongs to the occupation phase. The results (see fig. 9:13) must be considered to be fairly coherent - and hence reliable - and fix the construction and first occupation phases between the 3rd and late 5th centuries. CONCLUSION It was possible to gain new knowledge of the chronology of the settlements in Mapagala through the excavations con­ ducted on the North rock of Mapagala in 1989, and at the foot of the rock on the west side in 1990. Although there were not sufficient facts to identify separately the structural relation­ ships among settlements, it was possible to distinguish the qualitative features of different settlement layers. There are three common factors that emerge from the ex­ cavation on the summit (test excavation no. 2) 1. The absence or destruction of strata belonging to the historic period. 2. Extensive finds of quartz, with microlithic imple­ ments from all contexts. 3. Finds of Black-and-Red Ware from the lowest con­ text (no. 3). The fact that quartz and microlithic implements are found in large quantities and brickbats, tiles and other finds in very small quantities from the remaining soil strata, shows signs closer to an earlier period than the historic. Therefore informa­ tion regarding a historic period is not revealed through the strata on the summit. The existence of Black-and-Red Ware in the bottom context and quartz from all the contexts, has given rise to certain complicated conclusions. That is because the economic structure of the society which used microlithic implements was a society of hunters and food gatherers, while the society that produced Black-and-Red Ware was an ag­ rarian economy. Two suggestions can be made to solve this problem. Firstly, the test pits selected for the excavation are not sufficient to solve all the problems. There is insufficient ev­ idence upon which a complete vision of the social structure of those societies could be built, while the structural relationsip of those finds which were discovered is not clear either. The discovery of Black-and-Red Ware from the lowest bottom strata, while quartz is found everywhere, poses the question whether the strata were turned upside down at a certain stage. But it is clear that these soil layers are not those brought from outside. On the whole, the excavation is insufficient to throw light on these basic questions. Secondly, the entire stratigraphic sequence shows one co­ mmon characteristic. The above finds are seen mixed in vary­ ing quantities in all the strata. There is a similarity between the stratum in colour and composition too. This shows that it is not possible to highlight the transitory features of settlements. What is revealed very clearly are not distinctive features that separate one settlement from the other, but those which show inter-relationship and interdepedence between them. This is evident not only in socio-economic factors, but in the technique as well. Such features are also found in places like Dambulla and Pidurangala (Bandaranayake 1990: 23). In the trench excavation (see fig. 9:9 and 9:10) conducted on the west side of Mapagala (1990), the order of strata, various activities, the order of pottery and the results from 14C-analysis were observed and upon that data conclusions were drawn. There is a settlement sequence here which can be divided into three main categories. 1. Protohistoric settlement 2. Early Historic settlement 3. Middle Historic settlement Proto and Early Historic ‘pre-wall’ settlements are rev­ ealed through context 3 in test pit 2 (1989) and context 7 of the trench (1990), as well as through the stratigraphic sequence, from context 24 upto context 16 in test pit 1 of the Sigiri Mahavava bund excavation, in 1989 (see report by Gunasiri in the PGIAR Archive). Although the factors revealed through the excavations conducted below the rock do not relate direct­ ly to protohistoric settlement features, these are confirmed through the excavation on the summit of the North rock. There is no similarly between the Black-and-Red Ware found on the summit and those found below. The Black-and-Red Ware remains found on the summit are similar to the megalithic cultural ware found in Pomparippu (Begley 1981) and Ib- bankatuva-Polvatta (Karunaratne this volume and forthc­ oming). The Black-and-Red Ware found in the 1990 trench show a certain similarity to Early Historic pottery. They are very similar to those pottery finds discovered from sites like Anu- radhapura Gedige (Deraniyagala 1972), and Tissamaharama . (Kuna 1985). When these similarities and dissimilarities are considered, two interpretations can be made about the original settlements, depending on the possibility of arriving at definite conclusions. 1. The strata on the summit (especially context 3) repre­ sents the megalithic culture. megalithic cultural group using microlithic implements in a mixed economy. Secondly, the major settlement period is the one repre­ sented by the main wall. This shows a stage of a more de- 2. The first cultural layer below (context 7) represents the Early Historic period. The iron slag here is a special feature. The fact that it is found from both strata may be due to some technological relationship. But the existence of microlithic implements and quartz, as on the summit, is not evident from the strata below the rock. And thus it is clear that of the two, the settlements on the summit are the older. It cannot be concluded with certainty that these quartz pieces are remains of a prehistoric usage. Their utility and related activities are not clear. Firstly, what is evident from this is the fact that qualitative characteristics of one settlement are reflected in some manner in the other settlements as well. The continuity of and inter­ relationship between the technology found in the prehistoric settlements and the technology found on the summit, shows a certain connection. But the settlement below is the second stage of its evolution. As a rough reference to the people of the first settlement, it can be said, as a suggestion, that they were a megalithic cultural group using microlithic implements in a mixed economy. Figure 9:13 Mapagala. Radiocarbon dates. Calibrated age ranges from cumulative probability, one sigma (68.26%). CalibETH 1.5b (1991). In the diagram is included age range for sample from cultural deposit south of Mapagala, under the Sigiri Mahavava bund (SMV Con. 16). Secondly, the major settlement period is the one repre­ sented by the main wall. This shows a stage of a more developed and organized nature. According to stratigraphical evidence discovered, it can be concluded that the vast con­ struction stage, which included the main wall of Mapagala, be­ longs to the second stage (see fig. 9:10). However, evidence of its activities apd utility was not revealed, therefore it cannot be exactly determined whether it is a fortress or a monastery. The phase represented by contexts 16, 13 and 4 (see fig. 9:9 and 9:10) can be considered a major reconstruction phase. Rough­ ly, it can be considered a result of activities in the 5th century. Characteristics of monastic settlements can be observed in a study of subsequent stages, contexts and pottery. Although such traits are not distinctive from each other, this place can be considered a monastic centre upto the 9th and 10th centuries, in view of the fact that a monastery complex like Mahanaga Pabbata Vihara is situated on the west of this complex. This is confirmed by the discovery of ‘attani pillars’ close to this site. ACKNOWLEDGEMENT The co-operation extended by Manel Fonseka, who furnished Archaeological Administrative Reports and other necessary particulars, was very helpful in this study. REFERENCES Andrews, J.B. 1909. Prehistoric fortification of Sigiriya. Sp- olia Zeylanica. 6, pt. 22: 75. Bandaranayake, S. (ed.) 1984. Sigiriya project: First Arch­ aeological Excavation and Research Report. Colombo: Central Cultural Fund, Ministry of Cultural Affairs. Bandaranayake, S. 1990. Approaches to the Settlement Ar­ chaeology of the Sigiriya-Dambulla Region. The Settle­mentArchaeology of the Sigiriya-Dambulla Region.Bandaranayake, S., M. Mogren and S. Epitawatte (eds.). Colombo: PGIAR. Begley, V. 1981. Excavations of Iron Age Burials at Pom- parippu, 1970. Ancient Ceylon. 4: 49-142. Bell, H.C.P. 1886, 1906-1908, 1908. Archaeological Survey of Ceylon Annual Report (ASCAR). Blakesley, T.H. 1876. On the Ruins of Sigiriya. Journal of the Royal Asiatic Society, Ceylon Branch 8. Chandralatha, M. Mapagala Excavations 1990. Unpublished Excavation Report. PGIAR Archive. Cat. No. 1990/06. Deraniyagala, P.E.P. 1973. Some sidelights on the Sinhala monastery-fortress of Sihagiri. Spolia Zeylanica. 26, pt. 1: 69-75. Deraniyagala, S.U. 1972. The Citadel of Anuradhapura 1969: Excavations in the Gedige Area. Ancient Ceylon 2\ 48- 164. De Silva, R. 1967/68, 1968/69. Archaeological Survey of Ceylon Annual Report (ASCAR).De Silva, R. 1970. Sigiriya. Colombo: Department of Ar­ chaeology. Geiger, W. 1969. Madhyakaleena Lanka Sanskruthiya. (Sinh, transl. of "Culture of Ceylon in Medieval Times" by M.B. Ariyapala) Colombo: Gunasena. Godakumbura, C.E. 1964/65. Archaeological Survey of Ce­ ylon Annual Report (ASCAR). Godakumbura, C.E. 1976. Architecture of Sri Lanka. Colom­ bo: Ministry of Cultural Affairs. Harris, E.C. 1979. Principles of Archaeological Stratigraphy. London: Academic Press. Ilangasinghe, M. 1987. Dambulla Viharaya. Kelaniya (Wa- ragoda): Empier Printers. Indrapala, K. 1965. Samaja Torathuru. Anuradhapura Yug- aya. Liyanagamage, A. and R. Gunawardhana (eds.). Ke­ laniya: University of Vidyalankara. Kuna, M. 1987. Local Pottery of Anuradhapura: A Way to its Classification and Chronology. Pamatky Archeologicke LXXVIII:5-66. Lawrie, A.C. 1898 (1988). A Gazetteer of the Central Pro­ vince of Ceylon. Vol I and II. Colombo: National Mus­ eum. Liyanagamage, A 1965. Anuradhapura Yugay a. Liyanagam­ age, A. and R. Gunawardhana (eds.). Kelaniya: University of Vidyalankara. The Mahavamsa. Geiger, W. (ed). 1950. Colombo: Govern­ ment Press. Paranavitana, S. 1970. Inscriptions of Ceylon. Vol 1. Colom­ bo: Department of Archaeology. The Rajawaliya. Suraweera, A.V. (ed) 1976. Colombo: Lake House Investments Ltd. Ranawella, S. 1984. Epigraphy. Sigiriya Project: First Ar­ chaeological Excavation and Research Report. Bandar­ anayake, S. (ed). Colombo: Central Cultural Fund, Ministry of Cultural Affairs. Seneviratna, A. 1983. Golden Rock Temple of Dambulla. Colombo: Central Cultural Fund. Seneviratne, S 1984. The Archaeology of the Megalithic- Black-and-Red Ware Complex in Sri Lanka. Ancient Ce­ ylon. 5: 237-307. Seneviratne, S. 1990. The locational significance of early iron age sites in intermediary transitional eco-systems: A preliminary survey study of the upper Kala Oya Region, North-Central Sri Lanka. The Settlement Archaeology of the Sigiriya-Dambulla Region. Bandaranayake, S., M. Mogren and S. Epitawatte (eds.). Colombo. PGIAR. Tissa Kumara, A. 1990. Perani Lakdiva Atthani Vidhy. Kel­ aniya: University of Vidyalankara.Wimalaratne, W. A 1984. Palath Seema Mammawath Ha Matale Ithihasaya. Ithihasika Matale. Wijesuriya, S.S. (e- d.). Matale: Matale District Cultural Board. Appendix: Context Descriptions Test excavations in 1989. Test pit No. 1 The excavation was done down to bedrock and the depth was 80cm at the end of excavation. This is described from the lowest context up to the surface stratum from the bedrock (see section drawing fig. 9:9). Context No. 1 This is a reddish brown (2.5 YR 4/4) layer with a high composition of sand subject to erosion. It is very fine and soft. Slopes from the north to the south. Texture: Clay 55%, Sand 40%, Silt 4%, Pebbles 1% Cultural finds are very rare in this context. Thickness 70 cm. Excavations were done upto 10cm, according to artificial strata. At the last stage (6th stage) a few brickbats, a piece of tile and small fragments of pottery were discovered. Context No. 2 A reddish brown stratum (2.5 YR 4/4) with a high quantity of pebbles. Moisture was high and hence the soil was soft. Texture: Clay 50%, Pebbles 40%, Sand 8%, Small Pebbles 2%. This takes a gradient from north to south. Cultural artefacts are a little more than those in context 1. Very limited quantity of potsherds was found. All were body parts made thin by erosion. There was some quartz and granite too. Test pit No. 2 Context No. 1 Dark reddish brown and very fine (2.5 YR 3/4). Sand grain is very rare. Very small fragments of potsherds, a few tiles and brickbats were found here. Microlithic artefacts were found too. Slopes from north to south. Texture: Clayey soil 70%, Cultural debris 13%, Sand 5%, Silt 5% Context No. 2 A dark reddish brown soil with a little fine sand (2.5 YR 3/4). A fine soil layer with high content of clay, 70cm thick. From the upper part, level 1-5, small brickbats, potsherds, tiles, quartz and two lumps of iron slag were found. From level 5 no other remains were found except quartz. Slopes from north to south. Texture: Clayey soil 65%, Sand 20%, Cultural debris 10%, Silt 5%. Context No. 3 A reddish brown (2.5 YR 4/4) and lighter brown than context 2. Full of fine sand. Takes a southern gradient from the north; thickness is 10cm. Washed away stratum. Microlithic artefacts were discovered in large quantities and Black-and-Red Ware to some extent. Texture: Clay 75%, Sand 20%, Silt 5%, Cultural debris 5%. The excavation in 1990. Excavations were done up to the sterile gravel, to a maximum depth of 3m and minimum depth lm. For convenience of reporting, each context from the bottom up to the surface (recent period) is described upwards. Context No. 1 This is the humus soil on the surface and the surface soil layer of the excavation ditch. Dark brown. 7.5 YR 3/4. Tex­ ture: Clay 83%, Silt 10%, Sand 5%, Cultural debris 2%. Context No. 2 An alluvial soil layer. Dark brown. 7.5 YR 4/4. Texture: Clay 77%, Silt 15%, Cultural debris 5%, Sand 2%, Stones 1%. Context No. 3 A filled soil layer about 50cm thick, where small stones are found in plenty. Context 4 seems to have been built on this. Grayish-brown layer. Texture: Clay 33%, Silt 2%, Cultural debris 5%, Sand 5%, Pebbles 5%, Small stones 50%. Context No. 4 This is a structure located behind context 34 (main wall) and built at a secondary stage. It seems that context 4 is built upon context 3. Context No. 5 This layer, which is about 30cm thick, slopes westwards, extending throughout the entire trench. Reddish brown .7.5 YR 6/5. A layer filled with soil for the construction of the terrace. Texture: Pebbles 35%, Clay 30%, Fragmented stones 21%, Sand 8%, Cultural debris 5%, Silt 1%. Context No. 6 A layer extending up to the wall from 300,302,107 limit. A filled layer about 50cm thick. Yellowish red. 7.5 YR 6/6. Texture: Clay 67%, Sand 6%, Pebbles 4%, Cultural debris 2%, Silt 1%. Context No. 7 A soil layer upon context 8. Situated between 300/102- 109 and 302/102-109. Strong brown. 7.5 YR 5/6. Average thickness about 20cm. Among cultural debris found from this layer are charcoal, tiles, potsherds, etc. People who settled in this context have dug context 34 and built the main wall. Texture: Clay 75%, Sand 10%, Cultural debris 5%, Granite 4%, Silt 3%, Other 3%. Context No. 8 This is the lowest bottom soil layer. Reddish brown in colour (7.5 YR 5/4). Both context 39 found in the second excavation ditch, and context 53 found in the third ditch to­ gether form the same layer. Context 33 and 34 walls are built upon this layer, which slopes to the west from the east. W.A. KUMARADASA Context no. 9 This is the sub-layer on the surface of the above mentioned pit. Mostly gravel is seen. Reddish brown. 75 YR 6/8. Texture: Stones 75%, Clay 14%, Sand 5%, Pebble 5%, Silt 1%. Context No. 10 This too is filled soil in the above pit. Dark brown. 75 YR 5/6. Texture: Clay 43%, Sand 35%, Pebbles 21%, Cultural debris 1%. Context No. 11 Filling of the pit cut for the construction of context 34 wall. Reddish brown. 5 YR 4/4. Texture: Clay 90%, Cultural debris 5%, Silt 3%, Sand 2%. Context No. 12 A sub-layer with a thickness of about 5cm extending up to the wall from 300,302/106 limit. Yellowish red. 5 YR 4/6. A filled soil layer. Texture: Clay 73%, Sand 20%, Pebbles 5%, Silt 1%, Cultural debris 1%. Context No. 13 The soil filled in the pit cut for the construction of context 4. Brickbats are common. Dark reddish brown. 7.5 YR 3/4. Texture: Clay 50%, Cultural debris 40%, Sand 5%, Silt 5%. Context No. 14 The surface layer filled in the above pit parallel to context 7. Dark, reddish. Texture: Clay 87%, Sand 5%, Pebbles 4%, Small stones 2%, Cultural debris 2%. Context No. 15 A layer of about 5cm upon context 7. Spread over sub­ squares 302/100-302/108. Small stones mainly found. Dark brown 75 YR 4/4. A filled soil layer. Texture: Clay 50%, Small stones 40%, Pebbles 4%, Sand 4%, Silt 1%, Cultural debris 1%. Context No. 16 A cut mark in the pit dug for the construction of context 4 wall. Context No. 17 Boulders in the terrace fill, deposited before context 9. Context No. 18 This is the cut mark of the foundation cut for construction of the main wall, context 34. This is dug from context 7. It seems filled with sub-layer contexts 10,14 and 11. Context No. 19 Filled soil of the above mentioned pit. Situated between 300/111-112. Small stones are found in large quantities. Yel­ lowish red. 5 YR 5/8. Texture: Small stones 70%, Clay 9%, Pebbles 5%, Sand 5%, Cultural debris 1%. Context No. 20 A filled soil layer on the surface of the above pit. Reddish yellow. 7.5 YR 6/6. Texture: Pebbles 41%, Small stones 25%, Clay 20%, Sand 16%, Silt 2%, Cultural debris 2%. Context No. 21 Cut mark of a pit dug from context 5 into context 8. Context No. 22 A sub-stratum situated below context 3 and upon context 9. Strong brown. 7.5 YR 6/6. Texture: Sand 52%, Clay 40%, Pebbles 5%, Silt 2%, Cultural debris 1% Context No. 23 A sub-layer situated on the lowest bottom of context 28. Dark brown. 75 YR 5/6. Texture: Sand 90%, Pebbles 5%, Clay 5%. Contexts No. 24 and No. 25 Context 24 is a cut pit. Context 25 is its cut mark. It has been cut into context 8 from context 22. Yellowish red. 5 YR 5/8. Texture: Clay 50%, Sand 50%. Contexts No. 26 and No. 27 Context 26 is a pit dug with context 27 as the cut mark. It was noted that it is cut into context 8. Dark brown soil; 7.5 YR 4/4. Texture: Clay 46%, Sand 20%, Grain sand 20%, Pebble 10%, Silt 3%, Cultural debris 1% Context No. 28 A cut mark. The pit cut into context 8 is filled with sub­ layers, context 23 and 9. Context No. 30 This is the surface soil layer between 90-97. Strong brown in colour. 7.5 YR 4/6. Thickness is 5 cm. Context No. 31 This is a termite hill, found at the foot of the main wall between 97-100. Context No. 32 An alluvial soil layer about 50cm thick. Dark brown. 7.5 YR 4/4. Cultural debris found in large numbers, among which are Black-and-Red Ware. Appears to be a few stones fallen from the main stone wall (context 34). Texture: Clay 70%, Cultural debris 20%, Silt 6%, Sand 4%. Context No. 33 This is the small wall on the second terrace. Built upon contexts 39 and 53. This wall, about one meter in height, is medium sized compared to the main wall. The measurements of the blocks of stone used for its construction are appr­ oximately 1.50mxl.00mx0.80m. The type of stone, used in the main wall, is used here too. Context No. 34 This is the main stone wall, 3.34m high and 1.80m wide. Each stone used generally measures about 2x1x1 l/2m. A single stone may weigh approximately 5 tonnes. This wall is constructed by connecting large, live stones. Of the walls built round Mapagala North rock, which are largely washed away, only this wall is left. Built entirely of granite.Context No. 35 A few pieces of stone from the main wall, fallen upon context 32. Sunk up to context 38. Found spread over a space of about 3m, 180cm away from the main wall. Context No. 36 A line of stones which shows marks of a construction upon context 32. Located horizontally between 302/94-95. Built of stone about 20cm in diameter. Context No. 37 A cut mark dug into context 39. Probably a garbage pit. Context No. 38 An alluvial soil layer between contexts 39 and 40. About 20cm thick. Dark brown. 7.5 YR 4/4. Cultural debris found in large numbers: Black-and-Red Ware, potsherds, tiles and ch­ arcoal. The majority of Black-and-Red Ware is found from this context. According to the cultural artefacts and location of this layer, it can be that this shows a stage of decay of the wall. Texture: Clay 50%, Sand 20%, Cultural debris 20%, Silt 5%, Small stones 5%. Context No. 39 Same as context 8 and 53. Context No. 40 This is the soil layer upon the one on context 39. Expands upto 97-92/300. On the 302 front it ends at 94. It has the appearance of a pit, but cannot be considered as soil filled in apit because it is very hard, with kabook too. Therefore it is not certain whether it is a transition of the context 39 pebble layer itself. It cannot be decided whether it is a decayed kabook stone. Yellowish red. 5 YR 5/8. Texture: Kabook 75%, Clay 15%, Pebbles 15%, Sand 5%, Cultural debris Nil. Context No. 41 Soil filled in the pit dug for context 37. Dark brown. 7.5 YR 4/2. Context No. 50 Same as contexts 1 and 30. Context No. 51 An alluvial soil layer on context 52. Dark brown in colour and 30cm thick (7.5 YR 4/4). Among cultural debris found here are Black-and-Red Ware and Red Ware. This layer, being a natural deposit, resembles contexts 2 and 32. Context No. 52. An alluvial layer on context 53. Dark brown in colour. 7.5 YR 4/4. This layer, which is 45cm thick, had a steep slope towards the west. The Black-and-Red Ware, Red Ware and brickbats are similar to those found in context 38. According to the nature of the deposit of cultural debris, this layer has some similarity with context 38 and 7. Texture: Clay 55%, Sand 20%, Cultural debris 18%, Silt 7%. Context No. 53 Same as contexts 8 and 39.