A Preliminary Report

John Spies: Independent Researcher: Abstract


Eighty-six open lithic sites have been recently discovered in Mae Hong Son and Chiang Mai Provinces. The majority of the sites are located in highland areas of Tambon Tham Lod in Pang Mapha District, Mae Hong Son Province. Typically, sites are situated on relatively flat and barren, well drained areas on ridges and hilltops, between 350 and 1,300 metres in elevation. A water source, often seasonal, is always nearby.

Other sites have been identified next to hot springs and on elevated areas near rivers. Lithic material found on the surface of the sites includes flaked tools fashioned from river cobbles and other locally available raw materials, such as chert and quartzite sourced from outcrops in the hills. At a few sites elongated, bifacially flaked axes and edge ground tools have been identified. There have been no ceramics or organic remains found. The location of the sites implies that they were used for wet season habitation. The density of sites suggests that the non-karst highland areas of Northwest Thailand were once extensively occupied by hunter-gatherers.

1. Introduction

The first archaeologist to excavate prehistoric sites in Pang Mapha and Pai districts of Mae Hong Son province was Chester Gorman, more than 30 years ago. He identified one of the six key traits of the ‘Hoabinhian’ as "a cultural and ecological orientation to the use of rockshelters generally occurring near fresh water streams in an upland karstic topography" (Gorman 1972).

Since then prehistorical research in Northwest Thailand and highland areas in general, has focused on cave and rockshelter sites (Shoocongdej 1996). This research bias towards sheltered sites has limited our understanding of the overall settlement patterns of hunter-gatherers. Higham states that"The upland groups, known generally as Hoabinhian, took advantage of the shelter offered by the front of caves. But this does not mean that they were not equally if not more at home out in the open, near rivers, or within the forest to exploit seasonal harvests. It is simply that caves provide greater protection for habitation remains and are more easily located." (Higham 1998. 63-64)

In a later publication he reiterates that our knowledge of inland hunter-gatherers comes "almost exclusively from occupation remains in rockshelters. This provides only a partial glimpse of their way of life, for they must surely have occupied many open sites as well" (Higham 2002, 81).

Shoocongdej (1996) has also noted that open sites are virtually absent from archaeological literature on Thailand and has pointed out the need to investigate lifestyles and mobility strategies of hunter-gatherers away from naturally sheltered sites.

Prishanchit (1988), however has previously recorded 52 open lithic sites in Mae Hong Son Province near the Mae Nam Pai and several tributaries at elevations from 300-700m. These sites were found only on hills and mountains close to rivers, or other water sources, and were up to 250m across "covered by stone flaking debris" (Prishanchit 1988). The lithic material scattered over the surface of these hilltop sites included large numbers of bifacially flaked, elongated axes. Prishanchit concluded that the sites were "workshops" for producing stone tool, and mentioned the "problem of finding associated habitation sites." (Prishanchit 1988).

Previous research has established that prehistoric hunters and gatherers sought caves and overhanging cliffs for shelter. There are, of course, large areas of Northwest Thailand devoid of natural shelters where the prehistoric inhabitants had to live in the open. Even in areas dominated by limestone karst, like Pang Mapha district that has hundreds of known caves, there are relatively few cave mouths that are particularly suitable for extended human habitation. Active stream caves are rarely habitable; their mouths often choked with mud, flood debris or boulders. Dry, inactive caves may be already inhabited by bats, snakes, and insects. Until recently tigers and bears also sought shelter in caves. Cave mouths can be covered in guano or be too damp and musty for occupation. Many cave entrances also have rough floors, dangerous boulder piles or limited livable space. Location can also pose a problem due to difficult access or distance from a water source, particularly in karst areas where there are often no surface streams.

Rockshelters, which I define as day-lit places protected from rain by an overhanging cliff, are generally more suitable for habitation. Guano is not a problem and they can be bright, breezy places that often have flat dirt floors and extensive areas under shelter. Virtually every rockshelter recorded in Pang Mapha District that is near a water source and has a dry, flat living space more than a few metres across, contains evidence of prehistoric use.

In Pang Mapha the total number of recorded cave and rockshelter sites with flaked stone tools and space suitable for habitation is less than 40. This low number of sites is despite more than 20 years of intensive exploration of the District’s caves and karst areas by this author, international teams of cave explorers, researchers, and local inhabitants. One of the most impressive caves in Pang Mapha is Tham Lod, a huge cavern formed by the Nam Lang, a stream that drains over 400 sq kilometres before sinking underground. Within 5 kilometres of Tham Lod, are scores of known caves, but only 3 cave mouths and 4 rockshelters contain evidence of use by hunter-gatherers. The only significant habitation site recorded in the karst hills west of Tham Lod is a large overhanging cave mouth at Tham Nam Bor Suksit (Sacred Well Cave). At the lowest point in Tham Nam Bor is a natural rimpool that contains water all year, the only perrenial water source for several kilometres. Five of the other sheltered sites are located on the edge of the karst area, close to streams (fig 1).

In contrast to the low number of sheltered habitation sites around Tham Lod, more than 60 open lithic sites have been recently identified in the vicinity. John Spies and Stephen Brown located most of the open sites in Tambon Tham Lod during 2000, and continued exploration of other areas sporadically throughout 2001-2. The total number of documented sites is 86 with many more sites still unrecorded.

This paper proposes that open lithic sites, which I believe were used for habitation, occur in greater density and numbers than previous research has shown. Open sites are far more numerous than sheltered sites, even in regions bordering on, and including, limestone karst. They are easy to locate and widespread. This paper argues that the site locations indicate that hunter-gatherers in highland areas spent the wet season at these sites, where they could live in relative comfort and had easy access to a wide variety of resources.

The recorded open sites in Pang Mapha are listed in a table with a summary of information collected at each location. Lithic sites in the Tham Lod area are recorded on a map. The next section describes how these sites were located. It is hoped that this will be useful for the identification of sites in other regions.

Fig. 1: Lithic Sites in the Tham Lod Area, Pang Mapha, Mae Hong Son

Open Lithic Sites: White Circles, Sheltered Lithic Sites: Black Stars

Notes on the Map: Map area is 100 sq. kms. Contours are 700m to 1000m. Top is north. Areas of exposed limestone are bordered with broken lines. Some circles represent a group of sites.

This is followed by general descriptions of the various types of sites and the lithic remains. The discussion section focuses on the similarities between sites, the lithic materials, settlement patterns and the possible mobility strategies employed by the region’s prehistoric inhabitants. It must be emphasised that this paper is based on preliminary reconnaissance and there have been no excavations or detailed surveys conducted at the sites.

2. Locating Open Sites

The first open lithic site recorded in Pang Mapha (1.1), was located during January 2000, on a low ridge near Huoy Hin Gong, a seasonal tributary of the Nam Lang, in Tambon Tham Lod. About 30 unifacially flaked river cobbles lay exposed on the surface of an elevated area, with more found a few metres away, by a seasonal pool. Further exploration of the Tham Lod area was conducted mainly during the wet season, from June to October 2002, and involved many long walks along ridgeline paths. At this time of the year these trails through open forest remain easy to negotiate, unlike the overgrown and muddy paths near streams. It was during a period of heavy rain, while standing on a well-drained lithic site, when we first developed the theory that the ridges were used for wet season habitation.

A pattern of prehistoric occupation of the area around Tham Lod was forming and we began examining other areas in Mae Hong Son and Chiang Mai Provinces for similar sites. Again we found that elevated flat areas on ridges and hilltops, near sources of water and abundant wild foods, were often littered with stone tools. Sites were found on high ridges above tributaries of the Mae Nam Pai in Muang District of Mae Hong Son and on mountain ridges between Pai and Chiang Mai. On the low hills and ridges in the broader parts of the Mae Nam Ping, Mae Nam Pai and Mae Nam Chaem valleys there are many sites, some very close to the river.

At some thermal springs we have found relatively huge concentrations of stone tools indicating intensive and/or extended use of these special sites. We visited as many of the undeveloped hotsprings as we could find in Northwest Thailand, but not all had evidence of prehistoric use. At some places, such as the springs at Pha Bong near Mae Hong Son, development of the area as a tourist attraction has made the recovery of evidence problematic.

Highland lithic sites are easy to locate, due to the consistency in their geographical location and the relatively undisturbed natural environment at most sites. It has been fortunate that the ridges and hilltops were generally too dry or infertile for farming and, until recently, had few trees worth exploiting commercially. On elevated, undisturbed areas the rate of soil loss appears to be about equal to the rate of build up. Flaked tools are commonly found scattered over the surface of the sites, or partially embedded, with little evidence of deeper deposits. Ridge sites, whilst unsuitable for farming, are however being increasingly threatened by road construction in highland areas. Evidence from many sites has already been disturbed.

When searching for sites, often the first clear indicators of prehistoric use are rounded river stones that must have been brought from elsewhere. The wet season and early winter months are best for finding lithics after forest fires and rain have cleared the dry season litter of deciduous leaves.

3. Hilltop and Ridge Sites

Highland lithic sites can be found in varying densities across Northwest Thailand. Topography has a major influence on the number of locations suitable for wet season habitation in geographically different regions. Karst areas are composed of rugged exposed limestone, caves, dolines (closed depressions) and sinking seasonal streams with few flat places available. Underground drainage and the subsequent lack of surface water add to the inhospitability of karst areas. The valleys of rivers such as the Mae Nam Pai and Mae Nam Chaem, are often steep sided with flat land confined to riverbanks. Highland terrain formed by the smaller tributaries of the region’s rivers however, is characterised by long ridges dividing catchments where there are many places suitable for habitation. The highest density of recorded sites is on these dry ridgelines. Most sites are within open pine, mixed deciduous, dry dipterocarp or mixed forest. Some higher sites are located in lower montane evergreen forest, often again mixed with pine.

All lithic sites are located on an elevated and well-drained, relatively barren area that requires minimal or no clearing. The sites are flat, or gently sloping, with thin or sandy soils, in places where there is no mud during the wet season. The dimensions of the flat areas can vary from a few metres to more than 50 metres across. The forest is generally open-canopied with little shelter provided by the trees, making the sites open to sunlight. The nearest available water source in many cases is a seasonal stream or pool. The pools, typically 3-5m across, are found in depressed areas on ridges where spring water or runoff can collect. They are only active during the wet season and often appear to have been dug or enlarged by humans, sometimes into a rectangular shape. Flat areas with stone tools are usually only a few metres away from the pools.

Lithic material can spread over hundreds of square metres and there can be as many as 10 lithic sites in a square kilometer. The density of lithic material can vary widely between the sites. Some sites contain a few scattered pieces while other sites are densely covered. Lithics are usually concentrated on the flat elevated areas, but can also be spread across nearby slopes.

Paths that follow the natural ridgelines, which can be as narrow as one metre across, connect the sites on ridges to other features in the landscape. Today, as in prehistory, people walking along ridges often have no choice as to which route to take and it is likely many pathways have been in use since the first people inhabited the area.

Most sites are close to raw materials for stone tool manufacture, usually rounded river cobbles or quartzite outcrops. Other valuable resources, that are ubiquitous and are used in the Tham Lod area, include bamboo, pine wood (for lighting), the bark from mai bor (for twine and rope) and gum from mai huk (making composite tools, waterproofing). It is likely that these forest products were particularly important for prehistoric hunter-gatherers.

Around Tham Lod there are 7 varieties of bamboo commonly used by the local Shans that occur naturally near their village- mai bo, mai sung, mai hea, mai kao larm, mai bong, mai bao and mai rai. Each of these varieties is best suited for different functions. Mai sung and mai kao larm, for instance, are the best varieties for cooking and boiling food. Other varieties are used for water containers, building shelters and roofing, making snares and traps, weaving mats and baskets etc. Lengths of dry bamboo, split by hand, make extremely sharp knives, which, amongst other tasks, are effective for cleaning and cutting skin and meat from animals. It is unlikely that stone tools would have been used for this task. The shoots of all varieties of bamboo are edible, with different flavours, and in highland areas today they are an important supplementary food during the wet season.

Most tools and other pieces of stone were often left or discarded on the flat area that I assume was the campsite. This could indicate that shelters had raised floors, being constructed in much the same way as they are by people in the area today. Having a stone tool that could chop, split and flatten bamboo (by chopping many times around the joins between sections) would have made construction of such shelters easy. I believe that hafted ‘sumatraliths’ could have performed these functions effectively, but this theory needs field-testing.

Another observation is that places near precipitous drops that are otherwise suitable for habitation have very few or no lithics. This suggests that dangerous sites were avoided out of concern for the safety of small children.

4. Thernal Spring Sites

The thermal springs near Ban Muang Paeng in the Mae Nam Pai valley and the Theppanom springs in the Mae Nam Chaem valley both have particularly large mounds of lithic material near to the hot pools, more than is found at any other open site and most rockshelter sites. Both these sites are located near to other water sources. Some other thermal springs, such as those nearer to Pai, also have many tools but not in such high concentrations.

The extremely hot water was undoubtedly the attraction of thermal springs for hunter-gatherers. Rather than water for washing, the springs were mostly likely used as a convenient way to cook and process foods. Today, people living near thermal springs cook bamboo shoots and eggs and use the water to remove feathers from birds and fur from animals. In prehistoric times the only natural containers that could boil water were lengths of green bamboo and these are still used for cooking by people camping in the forest. However, bamboo is not so effective for the extended boiling of bamboo shoots or large sections of meat, as it tends to burn through after a while. Having the ability to boil large amounts of food for long periods and to easily clean animal carcasses would have been a definite advantage for hunters and gatherers. Boiled bamboo shoots and meat can also be dried in the sun and stored.

Of importance for future research is the fact that, unlike at other open sites, there is the possibility that prehistoric materials have been preserved in sequential deposits, cemented by the mineral rich waters. At the Muang Paeng springs in the valley of the Nam Pai, ‘sumatraliths’ and other flaked tools are partially cemented into the rim of several hot pools and it is possible that more tools and possibly bones, charcoal and even hair have been preserved in the lower mineral deposits.

Table 1: Open lithic sites in Pang Mapha. Notes on the table: The table is a summary of information recorded at some open lithic sites in Pang Mapha District. It is neither complete nor comprehensive.

The elevations and map co-ordinates have been determined from 1:50,000 maps with some GPS readings. The heights are generally accurate within 20m and the co-ordinates within 20-100m.

Ridgepath access (RPA) for most sites demonstrates that ridgeline paths undoubtedly crossed the landscape allowing for easy exploitation of wide areas during the wet season.

The density of the lithic materials at each site is arbitarily given as low, medium and high. This preliminary grading reflects the amount of lithics relative to other sites. The lithic symbols simply illustrate whether rounded river pebbles and cobbles or quartzite and chert sourced from hill outcrops were the predominant raw materials.The forest types are the dominant species in the immediate vicnity of the site.

The site dimensions are an approximation as are some of the distances and times given to the nearest sources of water, during the wet season.

Table 1: Open lithic sites in Pang Mapha.


1.1 to









4648 1

4648 11

4647 1

4647 1V


TL- Tham Lod

NL- Nam Lang



Width of flat drained area in metres.

RPA-ridgepath access to wide areas


P- Pine

DD-dry dipterocarp

MD- Mixed deciduous

LME-Lower montane evergreen







HQ-hill quatzite


S-unifacial discoids


SS-seasonal stream

PS-perennial stream


1.1 Low ridge 730 25.10/62.20 East of Nam Lang, near T.L. Above seasonal stream. 10m. RPA MD.DD. H. R. Q. S. bifacially flaked axe S.pool-5m


1.2-1.11 High broadridge 960-1040 26.50/62.70-63.10 Doi Hin Gong, east of T.L., 20 mins. from N.L. 9 sites near pools and SS. 20-30m. Barren, open. Broad, flat areas. RPA P H.R.Q.S. S.pool at top of ridge, SS near sites.
1.12 High broad ridge 1100 25.30/60.45 Doi Hin Gong, looks over NL. Small flat area on edge of precipice, barren, rocky. RPA p L. R. Q.

Minor site

Spring-10 mins walk
1.13-1.16 High narrow ridge 940 22.65/66.10 3km NW of TL. Edge of karst Nearby sites, 5-10m. Open forest, barren RPA P. LME. M.HQ.R. PS-5-10 mins walk
1.17-1.18 High narrow ridge 820 23.55/65.60 2km NW of TL. Edge of karst. Nearby,20-30m, open forest, RPA P. DD M.R.Q.C. SS-5-10 mins walk
1.19 High narrow ridge 980 22.45/67.55 West of Huoy Haeng, near karst 10 by 30m, cut by road. Narrow ridge.RPA. P.LME. M.R.C.HQ Next to spring PS-10 mins
1.20 High narrow ridge 900 21.95/68.80 Between Huoy Haeng/Nam Yappanair 10 by 40m, cut by road, elevated area on ridge RPA P.LME M.R. Perennial spring-400m
1.21-1.22 High narrow ridge 1040 21.30/70.35 Between headwaters of Huoy Haeng and Nam Mae La Na 20 by 20m. rounded hills on ridge, gravel. Barren. RPA P. LME H.R.C. S pool near. P spring 300m
1.23 High broad ridge 1060 21.70/70.90 North of headwaters of Huoy Haeng on ridge east of Nam Mae La Na 70 by 10-30m, broad flat arrea, RPA MD. LME. P. H. R. Q. S.

2 ground adzes

S stream-100m. P spring 300
1.24 High small hill on ridge 1080 21.35/71.10 300m north of 1.23. On north-south ridgeline east of Nam Mae Lana 5m across, on hill top, disturbed by ‘ring ditch burial’ mound. RPA P.LME L.R. Q.

Minor site

1.25-1.26 High broad ridge 1040 21.80/70.80 100m SE of !.23. Above headwaters of Nam Huoy Haeng. One site is 60 by 10m, and other is 20 m. Broad, open area. RPA P. LME M.R.Q.S. SS next to site. P spring-5 mins.
1.27-1.29 Low broad ridge 740 25.55/65.20 On ridge between Nam Lang and Nam Paem Highest site is 20 by 10m, others about 5-10m across. Broad open area. RPA DD.P. H.R.Q.S. s.pool next to site. NL-5 mins
1.30 Low ridge 720 25.66/64.80 Beside road to Muang Paem 15m across. DD M.R.Q. PS.-10 mins
1.31 Low ridge 740 25.90/64.70 Cut by road to muang Paem 10 by 15m.RPA MD M.R.Q.C. SS-5-10 mins
1.32 Low hill 750 26.45/65.25 Above Nam Pam !0 m, barren, next to precipice P.MD M.R.Q. PS.- 5 mins.
1.33- 1.35 Low ridge 790 25.40/63.95 East of NL, opposite Ban Tham Lod Sth. Sites 50m apart 10m across, barren, open. RPA P.MD. H.R.Q.C. S.spring-300m, NL- 500m
1.36-1.41 High ridge 745-953 25.65/66.15 –


3 major, 3 minor sites along ridge between NL and Nam Huoy Haeng Minor-3-5m

Major- 20m across

Barren, open. RPA

P.DD. MD. H.R.Q.S. S.pool at main site. P.streams 10-20 mins.
!.42-1.45 Low ridge 650-688 25.50/60.00-24.65/60.50 4 sites, one major, east of NL, near river. Major-flat area 100m by 30m. RPA P Major- H.R.Q. S.pool next to main site. PS-5 mins.
1.46 Low broad hill 650 24,60/64.00 Tham Lod sth village area Broad flat hilltop,flat area 2000sq.m. MD Significance unknown. Sub surface tools. R.Q. P.spring-200m. NL-500m.
1.47 Low hill 680 24.30/62.90 Near TL-Soppong road, edge of karst 10m across, rocky MD. DD. L.R.Q. Near S.pool
1.48 High ridge 964 25.75/61.65 East of NL, high on junction of ridgelines 20m by 15m, barren, open. Next to steep slopes. RPA P.MD. H.R.Q.S. PS-5-10 mins
1.49-1.51 Low broad hills 650 24.65/61.00 200m east of NL, on gently sloping hillside. Broad flatter area about 200m by 30m RPA P M.R.Q. small flaked bifacial axe PS and NL nearby.
1.52-1.54 Low ridge 740 25.35/63.30


East of NL, 3 sites along ridge south of Huoy Pa Chang, 0pp. TL. 10-50m across, sites about 50m apart. RPA P. DD. M-L. R.Q.HQ. PS-5-10 mins.
1.54 Mineral spring 700 28.10/63.90 South of Muang Paerm, near tributary of Nam Paem, low by stream Spring-100m across

Possible living area about 20m across.

Stream gallery evergreen H.R.Q.

Small ground adze

PS nearby
1.55 Hill slope 700 28.00/64.80 On path 500m sth of Muang Paem Tools found on slope MD L.R.Q. PS nearby
1.56-1.58 High broad Ridge 800-880 26.50/66.00



Along highest ridge dividing NL and Nam Paem, north of Muang Paem. 3 major sites, 20-30m across, 300-500 m. apart. Some tools embedded. RPA P.MD. H.R.Q.S. SS & PS in valleys, 5-10 mins.
1.59-1.61 High ridge 860-884 26.90/64.2



Along ridge dividing NL and Nam Paem, east of Ban Tham Lod. 3 sites about 10m across. RPA MD.P. M.R.Q. SS-5-10 min.
1.62-1.64 High Ridge 780-956 29.80/65.00



Along ridge, sth of Ban Eyla road, above headwaters of Nam Paem. 3 sites on hilltops along ridge. 20m across. RPA MD. H.R. Basalt? More tools than flakes. PS in valley-5-10 mins
2.1 High ridge 1003 22.60/45.60 Near road to Ban Manoula, 1.5km from Luk Pa Kor. On hilltop on ridge, 15 by 20m. RPA MD. LME. H.R.Q.S. S.pool/spring. 50m.
2.2 High ridge 1230 22.20/50.35 On ridge SW of Ban Nam Rin. High, far from P. streams. RPA P.LME. H.R.Q. 2 S.pools nearby.
2.3-2.5 Low ridge 730-780 07.75/60.45



Sth of 1095 Highway, about 1km east of Nam Khong. 3 sites along ridge, about 10m across.RPA DD H.R.Q.S. several biface axes SS in valley-10 min.

5. Lithic Remains:

From a general reconnaissance level perspective there appears to have been a wide range of raw materials and techniques employed in the making of stone tools at different sites. There are also a variety of sizes and shapes of tools observed including some sites with bifacially worked tools. The overall impression however is that most sites have at least some unifacially or bifacially flaked river pebbles or cobbles that could be considered typical ‘Hoabinhian’ tools. These include ‘sumatraliths’, ‘half axes’ and ‘scrapers’. Some of these tools have been manufactured from cores broken from large cobbles or boulders, rather than from pebbles and small cobbles. Most sites also have pointed tools, usually triangular in shape. All of the sites show evidence of at least some tools being fashioned in situ, with unused flakes, broken cores and debris making up the bulk of most assemblages.

Lithic raw materials generally appear to be whatever suitable stone was easily obtainable. If a stream was nearby, then rounded river stones were usually the raw material used for most tools. Pink or grey, fine to medium grained quartzite appears to have been the stone of choice in the Nam Lang area, but many other rocks, such as basalt, were also used. There are also sites, far from a source of river stones, where the raw materials were not so uniform or symmetric. Some tools have been fashioned from pieces of quartzite, broken from nearby outcrops, irregular pieces of quartz and occaissionally chert, which is common on some high ridges. Sites near the Mae Nam Pai and Mae Nam Chaem contain tools made from a larger variety of raw materials, including fine-grained quartzite, basalt, schist and chalcedony.

At a few sites, natural or human excavation has provided cross-sections that clearly show the depth of lithic material. This is particularly well illustrated at site 2.5, where a road cut through the centre of the ridge has exposed a steep profile. The lithic material is less than 10 cms deep and this appears to be the case at most barren ridge top sites. There are a few sites on less well drained or less open areas where there has been an accumulation of soil with many tools partially embedded. There is a possibility that an excavation could reveal deeper deposits.

At several sites edge-ground axes and shouldered adzes were found amongst the flaked pebble tools. Two complete shouldered adzes, flaked and ground from a soft, cream coloured rock, possibly siltstone, were found at site 1.23. The larger adze is 10 cms long, 4.5 cms wide and 2 cms wide. They have both been fashioned from the same rock and are identical in design. They have a rust coloured coating, presumably due to oxidation after long exposure on the surface. Another adze was found next to a path 300m northwest from the site. At a mineral spring near the Nam Paem was found a small, heavily used, shouldered tool. Originally made as an adze, the tool was later ground on both sides of the cutting end, effectively making it into an axe.

Several bifacially flaked axes made from a cream coloured stone, have been found amongst unifacially flaked river cobbles at sites 1.1 ,1.51, 2.4 and 2.5. These appear to be similar to axes found near Mae Hong Son (Prishanchit and Dissayadej 1983).

At site 1.37 is half of a broken stone ring, with a man made hole in the centre. This ring, made from a flat round river pebble, is similar to many more found at the Tham Lod rockshelter excavation site (Shoocondej pers. com.). It appears to have been drilled and one possible use for such a ring could have been as a weight for a bow drill for starting a fire or for drilling stone, wood and bone.

Whether these artifacts are contemperaneous with the other tools at the sites or are from a later occupation or use is a matter of speculation. The finds do suggest however that people who may have been amongst the first agriculturalists in the area used some of the same highland sites.

No ceramics were found at any of the sites. This could be interpreted as being due to weathering and erosion over thousands of years rather than as confirmation that the sites were exclusively used by an aceramic population

6. Discussion:

It is likely that the artifacts found at the open sites represent only a small fraction of the material remains left by hunter-gatherers. Most stone tools, particularly those made from rounded river cobbles and pebbles were probably fashioned at the source. Broken or blunted tools used away from the campsite were probably discarded anywhere. Without any surviving organic material, what remains are the lithics that weren’t swept away during campsite clearance or by erosion. Still, much material survives in situ and provides valuable information on settlement patterns.

The density of sites in some areas, and the relatively large amounts of lithics found at some locations tends to indicate that either there was extended use and re-use of sites, or that some favourable areas were more heavily populated. At some of the well-used ridgetops, such as Doi Hin Gong, the close proximity of lithic sites suggests that some could have been inhabited contemporaneously. In other areas where there are several places suitable for a campsite in close proximity, usually only the most comfortable site has lithic remains. This may indicate that only one small group of people periodically used the site, possibly over an extended period.

The higher ridges were probably used as campsites only during the wet season. The evidence supporting this theory comes from the lithic artifacts left at the sites, their geographical locations and the distances to the nearest water sources.

The lithics are always concentrated on flat or gently sloping elevated areas that are naturally well drained. The lithic assemblages generally include a variety of well-used or broken tools, lots of flakes, debris and large river cobbles or quartzite rocks with peices broken off. Many of these sites are close to water sources that are only active in the wet season, usually the headwaters of small streams or pools.

The existence of flat, well-drained, places on the ridges and hilltops in the region, gave the inhabitants a distinct advantage over populations in lowland areas. Movement from stream valley camps during the dry season to higher sites during the wet season might have only involved a 20 minute walk, but would have resulted in a noticable improvement in campsite comfort levels. Many sites are several hundred metres higher than the valley floors and the differences in airflow, humidity and temperature at these elevations would have been a factor in their choice for habitation. In addition to this the vast majority of sites occurr within sparce or open forest on thin soils. Ground cover is minimal or requires little clearing; there is no mud or leeches and a greatly reduced number of snakes and insects, including mosquitoes and ticks. Malaria is not present above 900m, where many of the sites are located. If malaria was prevalent, then having wet season camps at high elevations would have helped to reduce the transmission rate.

Another advantage of having campsites on ridges is the relatively quick access to a wide variety of food resources. Paths along ridgelines are easy to follow all year and are rarely overgrown. In the Tham Lod area, for example, the inhabitants of campsites on the Doi Hin Gong ridge (sites 1.2-11) could walk along paths to Tham Lod and the Nam Lang valley in 15 minutes. The Nam Muang Paem valley is 20 minutes away and the Nam Huay Haeng and Nam Pha Mon valleys are less than an hours walk. They could also use the natural bridge over Tham Lod to safely cross the Nam Lang during the wet season, and exploit vast areas of limestone karst to the west of the river. The area around Tham Lod is particularly diversified with different forest types found on the valley floors, hillslopes, ridges, higher mountaintops and within the karst (Richardson 2003a). The easy access to all of these areas of evergreen and deciduous forests and several streams from ridge top campsites would have enabled the exploitation of a wide area and a large variety of wild foods.

Assuming that dry season camps were generally located near streams, often the only water sources available, then a move to nearby hills or ridges during the wetter months was not primarily in response to the availability of seasonal food resources or materials for everyday use. People moved uphill simply to find a more comfortable campsite location, rather than to change or expand their areas of exploitation.

Further more, the advantages of using ridge paths to exploit a variety of ecosystems and natural resources would have lessened the need to move wet season campsites. Mobility strategies employed by hunter-gatherers during the wet season in seasonal tropical environments have been hypothesised as being residential, entailing more frequent moves to enable better exploitation of seasonal resources (Shoocongdej 1996). The archaeological remains at such camp sites would be expected to be predominantly local lithic types, with an expedient stone tool technology, where at dry season camps there should be evidence of long distance use of lithic raw materials and a more curated stone technology (Shoocondej 1996). During the dry season it has been reasoned that logistic strategies were preferred with less movement of camps required. However, much of the archaeological evidence supporting this hypothesis has been derived from naturally sheltered sites. The evidence from open lithic sites suggests that hunter-gatherers in the seasonal tropics may have employed different mobility strategies than has been previously thought.

A problem with the residential mobility model for the wet season, when viewed in the light of this new evidence, is that it appears more logical to assume that wet season shelters, constructed on ridges, should be relatively well built to protect from wind and rain. If roofed with split bamboo, a likely construction material, then shelters could last for several seasons. It seems unlikely that people would frequently move camp and build new shelters. It is more plausible that logistic strategies were used, with people using the ridge paths to easily access food and other resources over wide areas, returning to the main camp each day. During the dry season it again seems more logical for residential mobility to be higher, as shelters are often not required and stream valleys are easier to negotiate. During the driest months long sections of streams and other valleys could be more easily exploited by moving camp.

The lithics found at the highland sites are generally comprised of tools expected to be used by people using logistic mobility patterns, with extended stays at one campsite, rather than more simply made, expedient tools expected from briefly occupied sites. Shoocongdej has theoriesed that toolkits used at logistic camps: "should show a relatively high level of diversity of functionally specialised tool types...and less concern over transport of complete residential inventories."(Shoocondej 1996a).

There are many ridges sites that I believe were inhabitated during the wet season and also contain a diversity of typical ‘Hoabinhian’ pebble tools; the type of assemblages associated with the logistic mobility strategy model.

The antiquity of the sites recorded in this paper can not be accurately determined without the recovery of datable material. This may prove impossible due to the open nature of the sites and the unlikelyhood of any organic material surviving. The best possibility of directly dating these open sites may be the extraction of organic material from thermal mineral deposits. A comparison of the open lithic material with the assemblages from dated sheltered sites in the region could help establish if these types of sites were used contemperaneously.

I believe that the sites we found are just a few of many more similar sites that undoubtedly exist across Asia. It is plausible that Post Pleistocene hunter-gatherer occupation spread across the whole region from river valleys to the highest habitable mountains. The Tham Lod area is a good example of the intimate knowledge prehistoric populations had of the region. The Nam Lang, like many other streams that flow into karst, sinks underground and the only access to the valley and its tributaries is over high mountains. The wealth of natural resources and the relatively large number of suitable campsites spread across the landscape were a big incentive for people to inhabit this area. It is also plausible that the people who stayed or moved in and out of the region eventually had no where else to go as the landscape filled with people. The idea that highly diversified highland areas could support continuous hunter-gatherer habitation is attractive. Long term habitation and familiarisation with an area and its resources should lead to more efficient exploitation. The accummulating archaeological evidence shows many different cultural periods in Northwest Thailand, extending from well over 10,000 BP until about 400 BP. It is quite possible that more evidence, including DNA analysis, may show that the original inhabitants of the region never left, until historical times.

7. Summary and Conclusions:

A clear pattern of extensive occupation by hunter-gatherers of the non- karst highland areas of Northwest Thailand has emerged. Caves and rockshelters, while being invaluable as preservers of cultural remains, tell only a small part of the overall prehistoric settlement of the region. The vast majority of the people must have spent their lives in the open, away from natural shelters.

The density of open lithic sites, even in areas where there are caves and rockshelters suitable for habitation, suggests that some areas supported larger populations than has previously been thought. Dry hills and ridgelines forested with pine or deciduous species, near to, or dividing, small streams at elevations of 400-1200 metres, were the most favoured locations for wet season camps. It is very likely that detailed surveys of similar terrain in other regions will reveal the same pattern of site location.

The people who left the lithics at these highland sites had several distinct advantages over people who inhabited lowland areas, with more variety in elevation and microclimates, forest types and food resources. Importantly, they were able to ameliorate the discomforts of the wet season, possibly reduce the transmission of malaria, and still have easy access to food and material resources via ridgeline paths.

The location of the sites on the drier hills and ridges implys that mobility during the wet season was primarily in response to a human desire for a more comfortable campsite. In the Tham Lod area, a variety of ecosystems and the raw materials for tools and commonly used items were all readily accessible during every season. It is most likely that residential mobility strategies were employed during the dry season and logistic mobility was favoured during the wet season.

The lithics found at the sites need to be surveyed and studied in depth, and compared with assemblages found at dated sites, before any firm conclusions can be made. The use of a variety of lithic raw materials and the different styles of flaking also needs more investigation, as does the presence of edge-ground tools. Lithic functionality experiments, especially with bamboo, could also increase our understanding of these sites, and assist with the interpretation of lithic assemblages.

The thermal spring sites are particularly important as their accumulated mineral deposits may contain a more complete record of the earliest habitation of the region.

I expect that many more similar sites will eventually be recognised, spread across the highland topography of the distinctly seasonal climatic zones of Asia.


Gorman, C.F.

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

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.

Prishanchit, S.

1988 A preliminary survey of lithic industries in Mae Hong Son, Nan and Uttaradit, Northern Thailand. in Prehistoric Studies: The Stone and Metal Ages in Thailand, edited by Pisit Charoenwongsa and B. Bronson, Amarin, Bangkok, Thailand.

Richardson, E.R.

2003 Geographical Analysis of Prehistoric sites in Pang Mapha. Highland Archaeology Project.

Shoocongdej, R.

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



Enquiries  CAVE LODGE 15 Moo 1, Pang Mapha District, Mae Hong Son Province, 58150, THAILAND Ph. 053617203