A GEOGRAPHICAL ANALYSIS OF PREHISTORIC SITES IN PANG MAPHA

Edward Richardson

Highland Archaeological Project in Pang Mapha district, Mae Hong Son province

Abstract

This paper will present geographical information relating to the prehistoric sites in Pang Mapha district, northwest Thailand. This information will highlight aspects of the physical landscape and natural environment, so that new data and information presented by the Highland Archaeology Project can be considered within a broader context. This paper will also demonstrate the high diversity of habitats and the abundance of accessible natural resources that exist in this area.

The prehistoric rock-shelter site near the cave of Tham Nam Lod, which has recently been excavated by the Highland Archaeology Project, and its surrounding landscape and environment are the focus of specific description.

1. Introduction

The past decade has seen encouraging developments in the study of Thai prehistory. New empirical data, along with continuous improvements to the technical capabilities of field and laboratory research, have helped to establish clear cultural chronologies and further our general knowledge of the past. However, our holistic understanding of these prehistoric cultures remains extremely limited (Higham, 1998. Shoocongdej 1996a). This is perhaps partly due to the fact that the majority of research undertaken in the north of Thailand has tended to focus upon caves and rock-shelters, with the emphasis being traditionally placed upon the identification, classification and chronology of the materials they contain. While these sites can justifiably continue to be chosen as the focus of research, due to the abundance and quality of archaeological data they are likely to reveal, there is a danger that the broader geographical context of this information will not be fully considered. The data produced by this type of research should and will remain an important source of knowledge, but we must recognise the limitations imposed by the very nature of the data itself. It is well known that prehistoric cultures would only have left behind fragmentary evidence of their activities, and from these fragments only a few will survive in the archaeological record, with even fewer still being recovered and interpreted by research projects. It is therefore easy to see how the focus upon particular artifacts, assemblages or sites, while essential to the overall goal of prehistoric research, can often lead to limited interpretations of the archaeological record in general.

In order to address some of these limitations and to provide a more holistic interpretation of the past, the Highland Archaeology Project has been approaching new research from an anthropological perspective (see Shoocongdej 1996a). This requires the consideration and use of practical and theoretical information from a wide range of disciplines, including social, biological and environmental sciences (Whallon 1982a; Shoocongdej 1996a). Therefore, this paper is presented in the hope that geographical knowledge of the area surrounding the excavated site at Tham Lod will assist with the interpretation of the recovered cultural material, in line with the anthropological framework of this project.

It is well known that Geographical Information Systems (G.I.S.) can be of great value to the modern archaeologist (Westcott et al. 2000). Creating a digital map that can relate archaeological data to the physical landscape can be of value to many fields of research including investigations into settlement patterns, migratory routes and the structure of natural resources, to name but a few. However, the development of such a complex system is currently beyond the scope of this phase of the Highland Archaeology Project, as well as this paper. Instead, I intend to present geographical information, which highlights physical and environmental features of the landscape, with an emphasis on the natural resources in the vicinity of Tham Lod. The practical use of a G.I.S. can often be limited by the information it contains. As such, I hope that this paper can be used as a framework for the future development of a G.I.S. that relates the archaeological sites in this area to the surrounding environment and landscape.

Therefore, in the course of this paper, I will describe briefly the wider geographical context of the Tham Lod site, so that the data and information presented by the Highland Archaeology Project can be considered within its regional context. I will also describe the diversity of habitats, and the abundance of resources, that exist in this area.

I will then finish with a discussion concerning some of the questions and issues highlighted by the information presented.

2. Definitions

Throughout this paper I will refer to a number of general terms that will be clarified and defined here.

The ‘research area’ of this project encompasses the district of Pang Mapha, in the province of Mae Hong Son. However, the project has so far focused upon the rock-shelter sites at Ban Rai (2001/2002) and more recently at Tham Lod (2002). This paper will therefore focus primarily on the sites that are in the vicinity of Ban Tham Lod, and especially the Tham Lod rock-shelter site itself. I intend to outline the basic geographical setting of this site as an example of a rock-shelter habitation site, with specific focus upon the natural resources in the vicinity of the site and the general accessibility of the surrounding area. As such I will also briefly refer to a number of other sites that have been excavated in this region, including the Ban Rai rock-shelter (Shoocongdej et al. 2002a) and the Spirit, Tham Pha Chan and Banyan Valley Cave sites (Gorman 1971; Bannanurag 1988; Reynolds 1992).

The archaeological timescale of this paper, which is referred to as the ‘prehistory’ of this area, should also be defined. Radiometric dating from the Tham Lod rock-shelter indicates activity at this site around 20,000 years ago (Shoocongdej et al. 2002b). Previously, carbon 14 determinations from the Spirit Cave have indicated activity at this site around 11,000 B.P. (Gorman 1972; Reynolds, 1992). Radiometric dating has also been recently applied to a number of Log Coffins in this area (Grave et al. 1994) which suggests that the timescale for this mortuary practice is roughly 2200 to 1200 B.P.

The ‘prehistory’ discussed in the coarse of this paper therefore falls within this time frame, which can generally be identified as from the Late Pleistocene period through to the Middle to Late Holocene period.

3. The Paleoenvironment

Any study that attempts to describe the geographical and environmental setting of a prehistoric site must consider the possible environmental and physical changes that could have occurred since prehistory. Investigations into the paleoenvironment of Southeast Asia, and at a more regional and local scale, are ongoing and could potentially highlight some important changes to the environment and vegetation in this area.

However, I believe that for the purposes of this paper we can generally assume the environment, and especially the physical landscape, in this area would have changed very little. The post-glacial climatic amelioration that occurred between approximately 15,000 to 8,000 years ago, would have had a less dramatic effect at higher altitudes (Bellwood 1985).

Paleoenvironmental research has also demonstrated that the extent of habitats can shift spatially (especially in terms of altitude) in response to climatic change, while retaining the overall diversity that still exists (Shoocongdej 1996b). In fact, it is likely that the most significant changes to the environment and habitats of this area have been caused by the long-term exploitation of the natural resources, and the more recent cultivation of the land.

We can also turn to the analysis of excavated faunal and floral assemblages as another indication that the environment has fundamentally changed very little. The analyses of excavated assemblages from this area (Gorman 1972, 1977; Reynolds 1992; Shoocongdej 2001, 2002) have identified a large number of faunal species that are still extant today, including several species of monkeys, civets, and deer.

In addition to this, the present environment also contains many of the species of flora identified from the cultural deposits at Spirit Cave (Yen 1977). This suggests that either all of these species are highly adaptable to environmental change, or that the environment, especially in terms of the overall diversity of habitat, has not significantly changed since prehistory.

This paper will also assume that the physical landscape has changed very little in the last 20,000 years. With the exception of dynamic physical features (such as active cave systems and rivers) the antiquity of the physical landscape, and the underlying geology of the area, extends far beyond the archaeological timescale of the cultural material being investigated.

It is important for us to fully understand the geomorphology of the area, and the site formation processes that have taken place at the Tham Lod rock-shelter. However, as the objective of this paper is to present a more general geographical perspective, I will only highlight limited aspects of this information, and follow the assumption stated above.

4. A Geographical Overview

It has already been mentioned that this project is concerned with the archaeology of Pang Mapha district, Mae Hong Son, which is located in the northwest of Thailand (see Figure 1). The district is well known for its beautiful and dramatic landscape, the aesthetic benefits of a topography that ranges from 400 metres to over 1200 metres above sea level. This area forms part of the mountain ranges that extend along a north-south orientation from the Malaysian peninsular to the Shan States in Burma (Dunkley 1986).

The Mae Nam Lang follows this orientation past the Tham Lod site, through the cave, and then heads southwest until it flows into a sinkhole just to the west of the Ban Rai rock-shelter site. It re-emerges at Tham Nam Lang roughly 4km to the west, which at 8.3km is one of the longest active stream caves in Thailand. From here it soon joins the Mae Nam Khong, which flows southwards to meet the Mae Nam Pai (as shown in Figure 1). This river then flows westward towards the city of Mae Hong Son, and eventually joins the Salween River on its journey southwards to the Burmese coastline. As such, the Mae Nam Lang can be described as a component of the Greater Salween River catchment area.

Although the climate of this region is clearly monsoonal, there are three distinct seasons. These can be identified as the wet season, from June until October; the cool dry season, lasting from November until February; and the hot dry season, lasting from March until May. During the coldest months the temperature in the Highland regions often falls below 10 degrees and ground frosts have been reported. The annual rainfall in this area exceeds 2000mm a year.

Figure 1 The Permian limestone outcrops, river drainage and fault lines in the Tham Lod Area.

(Source: Adapted from original in Dunkley & Brush 1986. 11)

Following the habitat zone model proposed by Holdridge (1967), Higham (1989 and 2002) has previously described the vegetation of this area as that of “a subtropical wet, evergreen forest” (Higham 2002. 54). However, this generalisation fails to recognise the real diversity of habitats that actually exist within this relatively small area. This diversity is caused by the dramatic topography of the area, coupled with the spatial change in the underlying geology (the spatial distribution of the limestone karst is shown in Figure 1).

Generally speaking, dry dipterocarp and mixed deciduous forest can be found in the lower lying areas and on the lower hill slopes. It is within this habitat that the co-dominant species of Teak (Tectona grandis) can be found, as well as frequent patches of different species of bamboo forest (although many of these patches are a result of secondary growth after human disturbance). At higher elevations the mixed deciduous and dry dipterocarp forest blends into lower montane and pine forest. Patches of evergreen forest, occur around the depressions and outcrops of limestone, for example in the area around Banyan Valley Cave. Evergreen gallery forests also occur at lower elevations, usually in cooler and wetter areas that are shaded from the sun or along stream valleys, for example around the exit of Tham Nam Lod (see Figure 2). Pine and chestnut forests, with very sparse ground vegetation, are located along the high ridges, but once again these forests are often mixed with other species. The presence of limestone karst also increases the diversity of habitats in this area, with some species confined to this unique landscape. For example Mai Bo, the largest species of bamboo in this region, can only be found amongst the karst.

This high degree of habitat diversity is also reflected in the variety of faunal species, many of which have been identified in the archaeological record (as mentioned above). The river and its margins can support a wide range of crustaceans, amphibians, reptiles, shellfish and fish, as well as the mammals and birds that prey on these resources, such as the otter and fishing cat. The forested areas have also been inhabited by a wide range of arboreal and terrestrial species, including the macaque and gibbon, a number of different species of civet and deer, squirrels, small cats and wild pigs; and in the past elephants, rhinoceros and tigers were also present (Higham 2002).

5. A Geographical Perspective Of The Tham Lod Rock-Shelter

The prehistoric site at Tham Lod is located at the base of an over-hanging limestone cliff, which is approximately 20 meters high. The floor of the rock-shelter has a flat area that is approximately 10 metres across and extends about 4 metres from the base of the cliff to the dripline, and is interspersed with large boulders. From the size of the habitable area, I would speculate that the maximum number of people able to occupy this site at one time would have been less than thirty.

The site is currently less than 100 metres from the Nam Lang, and is over 20 metres above the highest water level. A full geomorphological investigation needs to be carried out to establish whether the course of the river, or the relative height of the shelter’s floor surface, has changed significantly since prehistory. Observations at the Tham Wua habitation and rock art site suggest that the living floor of the rock-shelter is now approximately 3 metres lower than it was in prehistory. This may have been caused by a combination of factors that may include the general uplift of the mountain range, the slumping of sediments towards the stream below, and the continual erosion of the valley floor by this stream. Similar forces could also be acting on the Tham Lod rock shelter, and it is likely that the present level of the floor is also lower than it was in prehistory.

The proximity of the river, which follows a known fault line (shown in Figure 1), is an important geographic feature of this site. Generally speaking the area to the east of the Nam Lang valley contains high ridges separated by steep-sided stream valleys and their flood plains, with a few pockets of Permian limestone. Further to the west of the Nam Lang, the landscape is dominated by limestone, creating what can described as a typical karst landscape, containing a large number of caves, pinnacle peaks, dolines, and sinkholes (including the sinkhole that the Nam Lang itself flows into). The area around Tham Lod can be described as lying within the transitional zone that exists between the karst and non-karst areas. The Tham Lod rock-shelter, located next to the river that divides these two geologically distinct areas, is at the centre of this transitional zone. This enables easy access to the surrounding diversity of habitats and natural resources.

As an example of this high degree of habitat diversity, Figure 2 shows the vegetation zones, including recently cultivated land, that can be identified in the immediate area around Tham Nam Lod. It was produced as the graphic display of results obtained from a detailed survey carried out by the Tham Lod Nature and Wildlife Education Centre, using a system adapted from the ‘Phase One Habitat Survey System’.

This system can only provide a general overview of the habitats in this area. Furthermore, it must be remembered that this map is a reflection of the current spatial variability in habitats, and that the spatial and temporal effect that long-term human activity may have had on this diversity is still unclear. Even so, it does clearly demonstrate the high degree of general habitat diversity that exists in such a relatively small area.

Figure 2 The habitat diversity in the area surrounding Tham Nam Lod.

(Source: Adapted from results of survey by the Tham Lod Nature &

Wildlife Education Centre.)

This diversity also suggests that an abundance of diverse natural resources would also exist within this area, and I have already mentioned many of the faunal species that are, or have been, extant. However, in order to further demonstrate the abundance of natural food resources in this area, I have compiled a table that lists some of the wild species of flora that are still exploited and consumed by the Shan people of Ban Tham Lod (Table 1).

This table is by no means comprehensive, and many other species of plants, fruits, fungi, and root vegetables can be still be found today. There are also many other species of insect that are seasonally consumed, including cicadas and termites during the dry season, and honey can be collected from the forests and limestone cliffs.

The hunting and trapping of mammals would also have followed seasonal patterns. Natural or man-made fires at the end of the dry season clear the undergrowth from the forest, making it easier to find, track and hunt the terrestrial mammals. The reduction in the overall biomass of the forest canopy would also have facilitated the hunting of the arboreal mammals during this season. At other times of the year, and especially during the wet season when it was less comfortable to stay away from a well-protected shelter, mammals tend to be hunted using traps and snares. This suggestion of year-round access to an abundance of food resources is supported by evidence recovered from excavations undertaken in this area. Gorman (1972) highlighted the fact that the shellfish deposits he excavated at Spirit Cave indicate that they were exploited all year round, and the preliminary identification of the more recent assemblages from the Ban Rai and Tham Lod rock shelters also show no significant signs of seasonal fluctuation (Shoocongdej et al. 2002a, 2002b).

Ethnoarchaeological data on the Phi Tong Luang, the last remaining group of hunter-gatherers in northern Thailand, has indicated that they also collect many similar plants to those shown in the table, including many species of yam (which are dug from the ground), palm, fruits and bamboo shoots (Pookajorn et al. 1992).

Bamboo is perhaps the most versatile of the floral resources in this area. A microwear study of stone tools from Tham Pha Chan has supported the assumption made by Solheim and others that these tools were primarily used for the procurement, fashioning and maintenance of wooden and bamboo products (Bannanurag 1988). The local Shan people currently exploit seven different species of bamboo, which are then used in a number of different ways. From the direct consumption of the new shoots, to the creation of shelters, baskets, containers, traps and weapons – it is clear that in the present day, and perhaps in prehistory as well, bamboo is one of the most heavily exploited resources in the forest. Bamboo is also an extremely durable material, and it is possible that traces could have survived in some of cultural deposits found in the well-protected cave environments. For example, a bamboo post was uncovered in Area III at Ban Rai, which has been identified as belonging to Component 3 and of post-Log Coffin culture origin (Shoocongdej et al, 2002a). Any data concerning this resource that can be extracted from the archaeological record could further our understanding of general resource exploitation and lithic functionality.

I have already mentioned the geological diversity of this area and the proximity of this site to the river. However, another significant feature of the landscape in the vicinity of the rock shelter at Tham Lod is the cave of Tham Nam Lod itself. This massive cave, with 1.6km of surveyed passage including a 600 metre long river tunnel formed by the Mae Nam Lang, is less than 3 minutes walk from the rock shelter. The evidence from the Log Coffin burials, and the example of rock art located deep inside the cave, indicates that this cave has been extensively explored throughout prehistory.

Until about 20 years ago, the local Shan people used bunches of burning bamboo or pine to explore the cave. Wood cut from the base and roots of pine trees contains a sticky sap that burns like oil, and Hilltribe people still extensively use this resource for lighting fires and providing light. The pine resources have undoubtedly been used for lighting and cave exploration since prehistory.

The exploration of this cave system was probably encouraged by the wealth of resources it contains. There is an abundance of aquatic resources, especially fish that gather in the deep pools and under ledges in the cave. There are also large numbers of birds and bats, which could have easily been exploited. Particularly notably are the thousands of swifts that inhabit the downstream chamber of the cave. The amount of guano and droppings on the cave floor is one reason why no evidence for habitation has been found in the cave. The birds and bats also provide unique hunting opportunities as many reptiles and mammals would have also exploited this resource, especially during the wet season when many birds fall to the cave floor. However, preliminary analysis of the faunal assemblage excavated from the Tham Lod site has not indicated a significant presence of bird or bat species. This could be an indication of an abundance of preferred foods, or that evidence of this resource exploitation is not represented in the archaeological record. In the present day, as there is so little meat that can be cut away from the bones, small birds are crushed and ground using a pestle and mortar before they are consumed. The consumption of bats, birds and small mammals in this way provides a valuable source of calcium.

One significant feature of this cave is the fact that the river has carved a 600 metre long passage through the hillside. This has effectively created a natural land bridge over the river, which would have provided vital safe access to the area east of the Nam Lang, especially during the wet season when the river can reach dangerously high levels. This physical feature enables year-round safe and easy access to the wide diversity of habitats and resources of this area.

This high level of access can be demonstrated by the length of time it takes to walk from the Tham Lod rock shelter to the some other prominent features in the landscape. During the dry season, a journey that follows the river to the sinkhole of the Mae Nam Lang will take roughly half a day, while the headwaters of this river can be reached within a full days walk. The Mae Umong stream, which is one of the major tributaries of the Mae Nam Lang, can also be reached within one hour. However, the path over the land-bridge of the cave and onto the ridge paths to the east increases the accessibility across this area, and especially during the wet season. The Muang Paem stream valley can be reached within one hour, the Pha Mon stream valley within two hours and the Nam Mae Lana valley within 3 hours walk – at any time of year. Figure 3 demonstrates this high accessibility across the area as it highlights the two major natural corridors that exist, namely the ridges and streams.

The paths that are marked on this map follow the natural lines of the ridges and hills, and it is highly likely that they have been used for access across this area since prehistory. In fact, the topography of many river valleys in this area makes it extremely difficult to follow the side of the river for any great distance, and eventually every stream in this area sinks underground. In addition to this, these streams and valleys are even more difficult and dangerous to follow during the wet season. This would further suggest that the ridgeline paths were the most exploited, and perhaps even preferred mode of access across this area. Although it is clear that prehistoric movement would by no means be restricted to these paths, and it is likely that their use fluctuated on a seasonal basis, Figure 3 does demonstrate the overall extent of the area accessible through these natural corridors.

Figure 3 The streams and hilltop paths in the area around Tham Nam Lod.

(Source: Edward Richardson and John Spies, 2003.)

6. Discussion

In summary, the Tham Lod rock shelter is a small but dry and comfortable site that is surrounded by an abundance of easily accessible natural resources. The location of this site within the transitional zone between karst and non-karst areas increases the degree of surrounding habitat diversity. In addition to this, the proximity to the Tham Lod cave system, the resources it contains and the natural bridge over the river, adds another unique dimension to the general accessibility and diversity of resources in this area.

The overall quality of this location as a prehistoric habitation site is suggested by its geographical setting, and proven by the evidence of long-term activity uncovered by the archaeological excavation. However, the abundance of natural resources also suggests that this area could support a far larger hunter-gatherer population than could inhabit the shelter itself. Therefore, we not only need to address questions related to the specific capacity and seasonality of this site, but we also need to identify the locations of contemporary sites. Investigations into these other sites would provide valuable data that could be used as a comparison with the intense and long-term activity that has taken place at this uniquely advantageous site.

The map below (Figure 4) provides an overview of the prehistoric sites identified in the vicinity of Ban Tham Lod, following extensive surveys conducted by Mr. John Spies and Mr. Stephen Brown, as well as the Highland Archaeology Project. This map clearly highlights the abundance and diversity of prehistoric sites that can be found in this area.

Figure 4 Prehistoric sites in the area around Tham Nam Lod.

(Source: Edward Richardson and John Spies, 2003.)

The sites have been identified by the clearest function or prehistoric evidence found there (for example habitation sites, rock art sites and burial sites). However, it is clear that not all of these sites relate to the same period of prehistory, and that many of them could have been the focus of different phases of activity. For example, the evidence from the Ban Rai rock shelter indicates that this one site was used for several different functions, during different periods of prehistory; with early evidence of habitation and a flexed inhumation followed much later by the Log Coffin mortuary evidence. Even so, the map makes it clear that this area was extensively exploited throughout prehistory, and by groups of people who were well adapted to their environment.

One significant feature of this map is the relative abundance of open-air sites along the ridges and lower hills, in relation to the limited number of cave or rock shelter habitation sites. The surveys carried out so far in the area around Ban Tham Lod have identified only seven cave or rock shelter habitation sites compared to 60 open-air sites. At present only preliminary reconnaissance of these sites has been carried out (see Spies 2003). The similarity between the type and style of lithics identified on the surface of these sites and the excavated lithic assemblages suggests that these two different types of site could be contemporaneous. However, we must avoid an interpretation of the chronological sequence in this area based upon the typological and technological features of lithic assemblages, as no clear sequence of lithic technology has yet been established (Shoocongdej 1996a). Instead, as demonstrated by this paper, we can perhaps further our knowledge of these sites by examining geographical information related to the surrounding physical landscape and environment. Such an approach provides a more holistic interpretation of these sites and avoids an over-emphasis on the lithic assemblage data.

Even so, the question of whether these sites are indicative of seasonal variations in the mobility organisation of prehistoric groups will remain unanswered until further research on these sites is carried out. Although I personally believe that this is the case, a model of the hunter-gatherer strategies and organisation that relates to the evidence in this area still needs to be established. While I have demonstrated many of the advantages of the Tham Lod rock-shelter site (which could suggest permanent habitation at this particular site), we should also recognise the disadvantages to living at these lower elevations. During the wet season, the river valleys are less accessible, more humid, muddy and infested with insects. However, the sites identified along the higher ridges offer a greater degree of comfort, while at the same time providing easier access to the natural resources of food and water. This would suggest that these sites were used for wet season habitation, but it is still unclear whether the Tham Lod rock-shelter was inhabited on a seasonal basis.

Would prehistoric groups be reluctant to leave a site with such unique advantages, despite the seasonal fluctuations in the level of comfort it offers? Can we therefore use this site in support of any model of prehistoric mobility organisation in this area, or should it be recognised as a particularly advantageous anomaly?

These questions, and many others, can really only be addressed after further archaeological, geomorphological, ethnographic and geographical investigations are carried out in this area.

Furthermore, this paper has demonstrated that any G.I.S. developed for this area needs to incorporate a wide range of information. The proximity of water, the diversity of habitats, the abundance of food resources and the general degree of accessibility all need to be included. In addition to this, specific features of the environment, especially in terms of the physical landscape, also need to be considered. Once this is achieved, a useful analytical system can then be produced. Such a system would not only be able to highlight patterns in the spatial distribution of archaeological sites, but would also be able to critically examine the relationship that exists between specific archaeological sites and the geography of their surroundings.

References

Bannanurag, R.

1988 “Evidence for Ancient Woodworking: A Microwear Study of Hoabinhian Stone Tools”. In Charoenwongsa, P. and Bronson, B. (editors). Prehistoric Studies: The Stone and Metal Ages in Thailand. Amarin, Bangkok.

Bellwood, P.

1985 Prehistory of the Indo-Malaysian Archipelago. Sydney: Academic Press.

Dunkley, J.R. and Brush, J.B. (editors)

1986 Caves of north-west Thailand. Report of the Australian Speleological Expeditions, 1983 – 1986. Speleological Research Council Ltd, Sydney.

Gorman, C.F.

1971 The Hoabinhian and after: subsistence patterns in Southeast Asia during the late Pleistocene and early recent periods. World Archaeology 2(3): 300-20.

Gorman, C.F.

1972 “Excavations at Spirit Cave, North Thailand: some interim interpretations”. Asian Perspectives 13.

Gorman, C.F.

1977 “A priori models and Thai prehistory: a reconsideration of the beginnings of agriculture in Southeast Asia”. In Reed, C.A. (editor). Origins of Agriculture. Mouton, The Hague.

Grave, P. et al.

1994 Dating the Log Coffins of Northwest Thailand. Paper presented at IPPA conference, Chiang Mai.

Higham, C.F.W.

1989 The Archaeology of Mainland Southeast Asia. Cambridge University Press, Cambridge.

Higham, C.F.W. and Thosarat, R.

1998 Prehistoric Thailand. River Books, Bangkok.

Higham, C.F.W.

2002 Early Cultures of Mainland Southeast Asia. River Books, Bangkok.

Holdridge, L.R.

1967 Life Zone Ecology. (revised edition). Tropical Science Centre, San Jose, Costa Rica.

Pookajorn, S. and Staff.

Phi Tong Luang (Mlabri): A Hunter-Gatherer Group in Thailand. Odeon Store, Bangkok.

Reynolds, T.E.G.

1992 “Excavations at Banyan Valley, Northern Thailand: a report on the 1972 season”. Asian Perspectives 31.

Shoocongdej, R.

1996a. Working towards an anthropological perspective on Thai prehistory. BIPPA 14: 119-32

1996b. Forager mobility organisation in seasonal tropical environments: A view from Lang Kamnan Cave, Western Thailand. PhD dissertation, University of Michigan, Ann Arbor.

Shoocongdej, R. et al.

2002a. Ban Rai Site Report (Thai Language). Highland Archaeology Project.

2002b. Tham Lod Site Report (Thai Language). Highland Archaeology Project.

Spies, J.

2003 Open Lithic Sites in Northwest Thailand. Highland Archaeology Project.

Whallon, R. and Brown, J. (editors)

1982 Essays on Archaeological Typology. Center for American Archaeology Press, Evanston.

Westcott, K. (editor) et al.

2000 Practical Applications of G.I.S. for Archaeologists : A predictive modelling toolkit. Taylor and Francis, New York.

Yen, D.E.

1977 “Hoabinhian horticulture: the evidence and the questions from Northwest Thailand”. In Allen, Golson and Jones (editors). Sunda and Sahul. Prehistoric Studies in Southeast Asia, Melanesia and Australia. Pages 567 – 599. Academic Press, London.

Figures & Table

Figure 1: Adapted from original in Dunkley, J.R. and Brush, J.B. (editors)

1986. Caves of north-west Thailand. Report of the Australian Speleological Expeditions, 1983 – 1986. Speleological Research Council Ltd, Sydney. Page 11.

Figure 2: Drawn by Edward Richardson, 2003. Physical features reproduced from original 1:50,000 scale map:

Edition 4-RTSD, Series L7017, Sheets 4648II and 4648III.

Habitats identified in survey carried out by the Tham Lod Nature & Wildlife Education Centre Interpretation Team.

Figure 3: Drawn by Edward Richardson, 2003. Physical features reproduced from original 1:50,000 scale map:

Edition 4-RTSD, Series L7017, Sheets 4648II and 4648III.

The routes of paths were identified and located by John Spies.

Figure 4: Drawn by Edward Richardson and John Spies, 2003. Physical features reproduced from original 1:50,000 scale map:

Edition 4-RTSD, Series L7017, Sheets 4648II and 4648III.

100 metre contours are shown. Darker shading indicates higher altitude.

Site locations identified on map by John Spies. Some site locations have been verified using ‘e-trex’ G.P.S.

Table 1: Created with information from Supaporn Singnakphum and John Spies