DISTRIBUTION, STRATIGRAPHY, CLIMATE

Preliminary Issues

Three issues need clarification before embarking on a review of regional distribution of sites, their stratigraphy and their climatic background. First, the succession of the lower, middle and upper palaeolithic is clear from the stratigraphic profiles obtained from different regions; second, in each case there are some distinct typological features of these successions. Both these points have long been known; in fact, as noted earlier, these three stratigraphically and typologically valid distinctions of the Indian palaeolithic were realized as early as 1930 by Cammiade and Burkitt and somewhat later by Todd. Subsequent explorations all over the relevant areas of the subcontinent have amply highlighted the validity of these distinctions and added considerable local and regional details to them.

Notwithstanding this development, it is important to note that problems related to such distinctions are not yet over. For instance, in the case of the Nevasa area in the Pravara valley, a tributary system of the Godavari, a recent opinion brackets the lower and middle palaeolithic components of the local stratigraphical profile together, arguing that the flake component of the middle palaeolithic could represent the light-duty facies of the early palaeolithic and that the usual heavy-duty component of the Nevasa Acheulian assemblage got disintegrated because these implements were fashioned out of basalt, an easily disintegrating stone. Further, the sorting by fluvial agencies could have been a contributing factor to the situation where the Acheulian implements are absent in the ‘middle palaeolithic’ context at Nevasa. Problems of this kind seem to have been best expressed by M.A.J. Williams and M.F. Clarke in the context of the Son and Belan valleys where three gravels were earlier identified in ascending stratigraphic order:

It was also very rare to find all three gravels in a single stratigraphic section, and field inspection of particular gravels soon revealed that they were highly complex, with substantial vertical and lateral facies variations, making it increasingly difficult for the stratigrapher to accept that there were only three gravels. Little further progress seemed possible until the gravel chronology had been replaced by an alluvial stratigraphy defined and dated independently of the archaeology, and based upon accepted tenets of stratigraphic mapping.¹⁰

Such problems are best left to geomorphologists to sort out, and it appears for the present that it is too early to discard the usual stratigraphic and typological divisions of the Indian palaeolithic into lower, middle and upper.

Such distinctions do not, of course, mean that the Acheulian tool-types based on pebbles and cores do not continue in the middle palaeolithic level; they do, but the basic distinctive feature of the latter assemblage is an emphasis on flakes. Similarly, the distinctive feature of the upper palaeolithic is a blade-and-burin industry or an industry primarily based on flakes and blades. In 1973, A.K. Ghosh pointed out that the palaeolithic culture complex of India might broadly be divided into three basic chrono-cultural elements: pebble-core, flake and flake-blade elements. According to him ‘the terms are based on culture–traditional units placed in the sequential framework, both geological and archaeological. The whole scheme broadly has an all-India character, though of course at places regional variations are marked.’¹¹ In 1974 Sankalia designated the principal tool components of the middle palaeolithic as scrapers, points and borers, whereas the distinctive feature of the upper palaeolithic was a blade-and-burin industry.¹² At the same time, it is important to mention here that the cultural validity of the three-fold stratigraphic division between the lower, middle and upper palaeolith is increasingly being questioned in the palaeolithic literature outside India. What Clive Gamble proposed in the European context, i.e. a distinction between a pre-35 thousand years ago (ka) as early palaeolithic and a late palaeolithic after that date, has been advocated in the context of the Potwar plateau by Rendell, Dennell and Halim:

The Early Palaeolithic in the Soan and the Potwar is defined in this study simply as being earlier than 30,000 BP; it cannot at present be subdivided into chronological stages defined by the presence or absence of certain items such as (Acheulian) handaxes, and (Levallois-Mousterian) tortoise- and disc-cores. What is clearly needed are examples of these items in secure geological contexts that are amenable to absolute dating.¹³

However, the important thing here is the availability of well-dated contexts. Till such contexts are reasonably plentiful in various parts of the subcontinent, it may be less confusing to stick to the labels of the lower, middle and upper palaeolithic.

Another point to note here is that the transition from one stage to another was not sudden. Where clear stratigraphic profiles are available, a clear evolution of the lithic industry undergoing transformations from the lower to the middle palaeolithic and from the middle to the upper palaeolithic, are noticeable. What is equally noticeable is a gradual change in the preference for raw materials: quartzite and similar material in the lower palaeolithic, and fine-grained stone varieties in the middle and upper palaeolithic. The varieties of available tool types considerably increase from the lower to the middle palaeolithic, and again, there is a sharp increase in the variety of tools from the middle to the upper palaeolithic. These general assertions notwithstanding, one need not postulate simple, uni-directional changes through time in the palaeolithic assemblages. The study of inter-assemblage variability within a given period does not support the premise of uni-directional changes.

As long as one can examine the regional and local stratigraphical profiles there is no difficulty in assigning in situ artefacts to their positions in the palaeolithic succession. The problem begins when one encounters surface assemblages, i.e. an assortment of tools lying on the surface in the ground: The distinctive tool types of one stage may also occur in the next, and it is a bold man who can assign a scraper or a group of scrapers and borers found on the surface straight to the middle palaeolithic stage. They can also occur in the next stage, i.e. the upper palaeolithic stage. One also encounters amorphous tool assemblages combining both early and late tool-types; in such cases the best thing perhaps is to judge the industry by its latest tool-type. If such an amorphous assemblage contains blades and burins, it will perhaps be more logical to assign them to the upper palaeolithic.

Our third issue in the present context is related to the methods of the study of palaeoclimates.¹⁴ Temperature, rainfall and humidity—the basic parameters of climate—leave behind their historical records in various living and non-living materials. Trees, ice-cores, lakes, coastal and ocean deposits preserve records of shorter duration (from a few centuries to a few thousand years) whereas deep-sea deposits and continental rocks contain the climatic story up to several million years ago. The methods of dating the relevant geological samples and placing them in a chronological order are many but the most common ones are based on the decay of radioisotopes. The range of applicability is dependent on the half-life of the isotope concerned and the sensitivity of the instrument to measure it. Some well-known radioisotopic methods are ¹⁴C, K-Ar, ³⁹Ar-⁴⁰Ar, ²³⁸U, ¹⁰Be and ²³⁰Th methods. In some cases age is directly calculated whereas in some others the rate of deposition has to be calculated first. Physical and chemical methods include thermoluminiscence (TL), Electron Spin Resonance (ESR), magnetic reversals and a method based on the racemization of the amino acid isoleucine. Unlike the TL, ESR and racemization methods where one can obtain the ages, plot them and obtain deposition rates, the magnetic reversal technique needs a continuous deposit to detect polarity changes in the earth’s magnetic field as recorded by samples which are known to have occurred during the following periods:

Once these events are detected in chronological order in a sedimentary column, it is possible to assign ages to the corresponding depths and calculate the deposition rates. Radioisotopic ²¹⁰Pb technique has been successfully used to date lake and coastal sediments related to the climate of the past century. Geochemistry plays a leading role in reaching the isotope levels of many of the elements which in the present context hold clues to climates. The fields of study are related to mineralogy, elemental composition, isotopic studies, marine sediments and trees.

At this point we should also consider the issue of Neogene—Quaternary boundary or the Pliocene—Pleistocene boundary in the Indian context. Neogene means the geological ages of Miocene and Pliocene and Quaternary means Pleistocene and Recent. So it is an issue of dating the Plio-Pleistocene boundary in India. In the context of the subcontinent the areas which have drawn detailed studies in this regard are the Siwalik beds of the outer Himalayas and the Karewa beds of Kashmir. The Siwalik beds have been studied for a long time whereas the Karewa situation has been examined in detail in recent years. Both these beds show thick sedimentary profiles and thus it is easier to study the problem here than elsewhere. The thick Siwalik section has been subdivided into Lower, Middle and Upper Siwaliks. The Lower and Middle Siwaliks are entirely Pliocene, with the topmost zone, Dhok Pathan, marking the end of the Middle Siwaliks. The Upper Siwaliks comprise, successively, Tatrot, Pinjor and Boulder Conglomerate. The problem seems to lie in determining if the Neogene–Quaternary transition can be put at the base of Pinjor or at the base of Boulder Conglomerate or even within Pinjor. Following D.P. Agrawal, we shall accept the opinion that this boundary in the Siwaliks should be placed at the Tatrot–Pinjor transition of about 2.4 mya, coinciding with the transition to the Matuyama phase in the history of changes in the magnetic polarity of the earth’s field. The change of Tatrot fauna into Pinjor fauna took place at this point and the argument is based on the simultaneous appearance of Equus, Elephas, Bos and Cervids with antlers. Agrawal also puts the N/Q boundary in the Karewa deposits of Kashmir at the Matuyama transition.¹⁵

Regional Survey

In the extreme north-west, the Sanghao cave at the edge of the Peshawar plain in the North Western Frontier Province of Pakistan has yielded evidence of middle to upper palaeolithic occupation, but tower palaeolithic is still unreported in this region. We have already drawn attention to the finds from the Soan valley and the Pabbi hills in the Potwar plateau section of Pakistani Punjab and further noticed that its post-Siwalik stratigraphy is dominated by the Lei conglomerate and loess deposits. The earlier Siwalik strata were eroded, and the quartzite cobbles derived from this process of erosion formed the first post-Siwalik deposit of the Lei Conglomerate. Then, after the formation of a layer by colluvial deposition, the sequence is dominated entirely by the deposition of loess, the oldest level of which perhaps reaches 170 ka. Between 130 ka and 75 ka loess is preserved in isolated pockets, but its deposition and preservation were widespread between 75 ka and 18 ka. Loess is basically a wind-blown silt material, and in this case its source has been presumed to be local, within 100–200 km. The whole phase of loess deposition may here be associated with locally drier periods and weaker monsoonal circulation. It has been inferred that the south-west monsoon penetrated this area only after 18 ka in the late Pleistocene. This picture is said to be corroborated by the work on Indian Ocean sediments which shows wetter conditions after 18 ka. Environmental information seems to be tentative for the pre-loess stage, but the upper Siwalik strata in this area have been supposed to connote the existence of a flood-plain zone, peripheral to major channels.

In Sind, the distribution of such sites is conditioned by the locations of the outliers of the Kirthar range, Milestone 101 and the Sukkur-Rorhi hills being the two well-known centres of palaeolithic distribution here. Only surface occurrences have been noted but they cover the whole range of the palaeolithic from the lower to the upper. The Rorhi hills provide a good source of the raw material used for making palaeolithic implements, and on the basis of the study of the Acheulian workshop site of Ziarat Pir Shaban it has been inferred that it was used for the manufacture of handaxes which were later taken to the (contemporary) Acheulian base-camps. Occurrence of palaeoliths has also been reported in the Las Bela plain to the west of Karachi, and although the report is not very specific, both lower and middle palaeolithic materials are said to be represented.¹⁶Since the earlier reports of lower palaeolithic artefacts from the Pahalgam sector, only upper palaeoliths have been reported from Kashmir. Considering that Ladakh has lower palaeolithic sites, there is no reason why they should not occur in Kashmir. The Kashmir upper palaeolithic has been placed around 18,000 BP and thus it coincides with a phase of climatic amelioration or warming. There is apparently an extensive literature on the loess-palaeosol sequences of the Kashmir valley, but till we get well-defined archaeological correlates, such literature does not have any archaeological significance.

The palaeolithic evidence is more explicit and widespread in the Indian Siwalik belt from Jammu to Kangra and the lower areas of Himachal Pradesh and Punjab. There have been terrace-related studies of the relevant finds in the Beas-Banganga valley of Kangra and the Nalagarh-Chandigarh area of Punjab, but in view of the fact that such terrace-related studies have been discarded in the Soan valley, which provided the model of such studies in the Indian Siwaliks, no emphasis may now be put on the implementiferous ‘terraces’ of the Beas-Banganga and other valleys. Perhaps a more useful approach will be to study the implementiferous occurrences of this region in terms of the basic Siwalik stratigraphy. The tremendous significance of the Siwalik sediments, ‘the thickest and most complete known Neogene sequence of fluvial deposits and palaeosols’, has recently been extensively commented upon.¹⁷

There is an extensive spread of palaeolithic material in Rajasthan on either side of the Aravallis which also contain palaeolithic material as far north-east as Haryana and Delhi. In east Rajasthan the major work was done by V.N. Misra in the Berach valley and its tributary valleys of the Gambhiri, Wagan and Kadamli, leading to the establishment of the following general stratigraphy:

Around Ajmer at the edge of the desert, the lower palaeolithic material occurs in sandsheet with angular slopewash debris, which was covered by another sandsheet deposit with calcification and rootcast development between this and the earlier sandsheet deposit. The middle and upper palaeolithic material comes from buried soil which covers the later sandsheet deposit. Between this soil formation and the covering deposit which is another sandsheet formation, there is trace of calcification and rootcast development. The Hokra basin to the north-west of Ajmer showed rich middle palaeolithic working floors and upper palaeolithic working material. The middle palaeolithic was a flake industry with levallois cores or cores derived from the use of ‘prepared core technique’. In the upper palaeolithic industry here one finds burins, carinated scrapers, blades and blade cores of quartz and quartzite. In the Luni valley, also in the arid zone of Rajasthan, the middle palaeolithic tool-bearing sandy gravel directly overlies bedrock and is topped, successively, by a silt and a sandy deposit. In this area lower palaeoliths also occur, although as stray finds. Middle and upper palaeoliths are found near Jaisalmer as well, although the upper palaeolithic is not as abundant as the middle palaeolithic here. On the basis of their work Bridget Allchin and her group inferred the following climatic and Stone Age sequence in the Thar desert.

In western Rajasthan, including the Thar desert, prehistoric research has significantly advanced in recent years under the leadership of V.N. Misra. In the Didwana area of the Nagaur district of west Rajasthan there are three major stratigraphic units, the first of which—the Jayal Formation—is a very extensive and continuous gravel ridge dating from late Tertiary-early Pleistocene and earlier than the human presence in the area. The two successive post-Jayal Formations, separated from the Jayal Formation by a disconformity, are the Amarpura and Didwana Formations. The Amarpura Formation, as exposed in the Singhi Talav quarry near Didwana, is a strongly mottled and well-laminated greyish silty calcareous clay, containing a rich early Acheulian material. It is a 15 to 20 m thick deposit exposed in various places and contains a succession of lower to upper palaeolithic material. Geologically, it is placed in late middle Pleistocene to late Pleistocene. The Amarpura Formation overlaps with the succeeding Didwana Formation which continues into the post-Pleistocene Holocene age. The reconstructed environmental picture shows that the Jayal Formation, which does not contain any cultural material, is a 60 m thick well-rounded pebbly-bouldery gravel, interlayered by a 2–10 m thick calcrete. It was deposited by a high-energy bed-load braided stream system and was uplifted in the form of a ridge some time during the early Pleistocene. The drainage system of the post-Jayal formations consisted of highly meandering and low-gradient streams with their wide flood-plains and their associated pools and lakes. The abundance of kankar beds in various forms and the presence of mottling in clay loams indicate that the climate was essentially hot semi-arid throughout the long period from the middle Pleistocene to early late Pleistocene. This flowing drainage ceased to exist some time in the early part of the late Pleistocene. In fact, the terminal pleistocene in this region has been supposed to denote one of the driest climatic phases. The early Holocene was relatively wetter.

A continuous cultural evolution has been traced from the early Acheulian to the middle palaeolithic in the Didwana area. The massive and thick handaxes with large unflaked areas and cleavers of convergent type with narrow working edges come from a well-defined horizon of a soft, loamy sediment at a depth of between 40 and 80 cm in the Amarpura Formation. Such tools were made, used and discarded on the banks of water pools or lakes, and incorporated in the soft sediment during periods of lake flooding. In the 50 km stretch from Didwana to Jayal several Acheulian sites have been found, with the possibility of many more to be discovered. In contrast, there are fewer middle palaeolithic and later assemblages, suggesting an era of disorganized drainage and severe aridity.¹⁸

Fig. 1 Site distabution around Didwana (Misra, 1982)

Gujarat has been well researched from the point of view of prehistory, and here one can offer only a representative section of this research. In mainland Gujarat, the Sabarmati valley and its tributary valleys of Orsang and Karjan reveal a straightforward section of bedrock–gravel conglomerate–reddish silt–yellowish sandy silt or loess, with the lower palaeoliths occurring in the gravel conglomerate. Middle and upper palaeolithic tools have been reported from the eastern margins of the mainland Gujarat plain. In a study of the Quaternary stratigraphy of the Narmada, Mahi and Sabarmati valleys in the Gujarat mainland, it has been observed that the average thickness of sedimentary deposits exposed in the three basins is about 35 m, the basal part of which is a bluish-grey clay deposited over bedrock. This may be marine in origin in the lower reaches of all these rivers. This is followed by a compact and cemented gravel horizon with lower palaeolithic implements. This gravel is overlain by a red soil horizon which suggests a more humid climate and is correlated with a particular stage (stage 5e) of the deep-sea oxygen isotope record, dated 125 ka. The red soil horizon is succeeded by sediments of aeolian origin or loess and finally, there is soil formation over the aeolian deposits. An analysis of the middle to late Quaternary deposits of the Sabarmati basin has shown the presence of semi-aridity throughout the stratigraphic sequence. In the Kutch peninsula, the Bhukhi valley profile shows the succession of bedrock, cemented conglomerate (implementiferous), yellow brown silt and Black Soil. This is the representative section of this region. In Saurashtra, the Bhadar valley profile is the following: bedrock > gravel conglomerate (lower palaeolithic tools) > reddish soil with kankar> gravel conglomerate (middle palaeolithic tools) > brown silt. The section offered for the Hiran valley in Saurashtra is important because of its association with Quaternary limestone deposits known as miliolite deposits: bedrock > fluvial gravels and silts (lower palaeolithic tools) > miliolite deposit 1 > fluvial gravels and silts (middle palaeolithic tools) > miliolite deposit 2 > fossil soils, etc. > dunes, etc. The miliolites occur not merely in coastal Saurashtra but also about 170 m inland and this has led to an extensive literature on the origin of these deposits and their mode of deposition so far inland. The inland deposits are generally considered wind-borne, although there are opinions to the contrary. According to N. Bhaskara, ‘These deposits were formed under sea water in close proximity to where they are seen today.’ In fact, the changes in the sea-levels of the Saurashtra coast have witnessed considerable research in recent years. According to one opinion, about 87.2 ± 9.7 ka the sea-level rise is estimated to be +7 m after making correction for a tectonic overthrow, whereas around 126 ± 5.7 ka the sea-level was −13 m as compared to the present. In the Kalubhar valley in the Bhavnagar district of Saurashtra the implementiferous gravel has been put in the late Pleistocene, and more importantly, ‘geologically the Acheulian occurrence does not appear to be stratigraphically separated from the implementiferous middle palaeolithic formation.’¹⁹

The Narmada tract is the most well-researched prehistoric area of Madhya Pradesh, though research has also been done in Damoh, Raisen, Upper Son valley and Mahanadi valley. In the Narmada valley, the stratigraphy offered by de Terra and Paterson showed three cycles of deposition on the bedrock and overlying laterite: gravel > pink silt; > comparatively loose gravel > pink silt; > fine gravel > pink silt. H.D. Sankalia’s stratigraphy was derived from the Maheshwar area: thick bouldery gravel (lower palaeoliths) > thick deposit of silty clay > a second boulder bed with the same types of lower palaeolithic tools > a second silt deposit > a finer gravel deposit with smaller pebbles, which contains middle palaeolithic tools > and finally, a clay deposit. Sankalia and his student S.G. Supekar, who undertook excavations in this area, did not find evidence of any stratigraphic evolution of the typology of the lower palaeolithic industry. However, G. Armand, who later undertook excavations in the area of the Durkadi stream, a part of the Narmada system, thought that the handaxe in the Narmada could evolve from a pebble base, and in this he was anticipated much earlier by A.P. Khatri, who thought that there was an Oldowan-like pebble industry at the base of the Narmada sequence. At Samnapur in the central Narmada valley, V.N. Misra and his group found a middle palaeolithic industry in what they call a semi-primary context of implementiferous rubble embedded in silt. At Bheraghat on the Narmada, Sheila Mishra and S.N. Rajaguru studied the following section: sandy pebbly gravel with Acheulian cleaver > break > bouldery gravel > sandy pebbly gravel with upper palaeolithic material.

In the Damoh area near Sagor in Madhya Pradesh, R.V. Joshi found two gravels, the lower one with the lower palaeolithic and the upper one with the middle palaeolithic. Acheulian surface clusters in Raisen district were studied on the basis of their sizes and locations by J. Jacobson. In eastern Madhya Pradesh, R.P Pandey has located a number of palaeolithic sites in the Mahanadi valley. The lower palaeolithic material comes only from surface occurrences, whereas the middle palaeolithic specimens are found associated with the river gravels. The upper palaeolithic industry of the area also seems to have been derived exclusively from the surface occurrences.

Fig. 2 Excavated scrapers of various kinds from Samnapur (Misra et al., 1990)

The rock-shelter-rich area of Bhimbetka near Bhopal in Madhya Pradesh has yielded evidence of a continuous prehistoric occupation from the late Acheulian onwards. Out of 8 layers in the 4 m thick deposit of its rock-shelter III.F-23, the layers 8–6 are late Acheulian, layer 5 is middle palaeolithic and the rest (i.e. the layers 4–1) are microlithic.

In the Son valley section (Sidhi district) of Madhya Pradesh a clear assessment of the stratigraphic situation has been made by Williams and Clarke. Their reconstruction of the Quaternary history of this part of the Son valley is as follows:

An important aspect of the prehistoric research conducted in the Sidhi district by G.R. Sharma, J.D. Clark and their group is the excavation of the upper palaeolithic site of Baghor I.²⁰

Palaeoliths occur all over Maharashtra, including the Konkan coast on the west and in the Wardha–Wainganga valleys in the east. The well-known stratigraphic profiles in archaeological literature are those of the Mula–Mutha rivers in Pune, the Godavari in Nasik, the Pravara (a tributary of the Godavari) in Nevasa and the Tapti river system at Patne near Chalisgaon. The Dattawadi locality on the Mutha river in Pune shows the succession of gravel bed (12 ft thick and with lower palaeolithic tools) > dark red silty bed (35 ft thick) > fine gravel (middle palaeolithic tools) > ashy grey silt, with the last two deposits abutting against the older deposits. The Gangawadi locality on the Godavari in Nasik shows the succession of bedrock > pebble gravel (lower palaeolithic tools) > thick brownish silt. The Hathi Well section at Nevasa on the Pravara shows a loose gravel over the earlier gravel with lower palaeolithic tools. This second gravel contains a middle palaeolithic assemblage of scrapers, points and borers made on finer raw materials like jasper and chert. The fact that this ‘middle palaeolithic’ assemblage may be the small tool component of the Acheulian has already been pointed out. Further, it has been argued that the Acheulian and this ‘middle palaeolithic’ are parts of the same geological unit. A thick bed of bouldery/pebbly gravel at the junction of the Chirki and Pravara streams was excavated for a number of years beginning in 1967 with the expectation of discovering a camp site. Although this expectation was not fulfilled, it was realized that the bouldery terraces near the river were used to manufacture tools made of local and imported materials. The Patne section begins with the middle palaeolithic tools at the bottom and passes on to the mesolithic through four stages of the upper palaeolithic. The subdivisions of the upper palaeolithic based on typology are corroborated by the local stratigraphy and there is no doubt that there is a steady continuity with a remarkable degree of overlap with the mesolithic level. By the time the Patne sequence reaches the mesolithic, the primary raw material is chalcedony instead of jasper; there is also a disc-bead made of a fragment of ostrich shell among the upper palaeolithic tools. Lower palaeolithic and later artefacts have been obtained in the Konkan coast from the Gujarat border to Goa, and the sequence shows a gravel silt succession.²¹

In Karnataka, the Malaprabha–Ghataprabha system, a tributary system of the river Krishna, has been well known for its prehistoric sites since the nineteenth century. The lower palaeolithic—middle palaeolithic succession is represented by two succeeding gravels at various places such as Taminhal. The Taminhal section shows the following ordering of deposits: mottled clay > gravel > clay > fine sandy gravel > fine gravel > modern clay. In the Hunsgi—Baichbal valleys, the Acheulian artefacts are located either within the black soil of the surface or within a brown clay which underlies the black soil. At Yedurwadi on the Krishna the artefact—horizon is a gravel with a number of buried trees preserved as calcrete casts. The idea that most of the Hunsgi–Baichabal sites are in ‘non-fluviatile sedimentary contexts’ has been disputed by Sheila Mishra who thinks that they were buried by fluvial silts which subsequently became pedogenized.

It may be noted at this point that the entire prehistoric stratigraphy of the Deccan is now being visualized in terms of a few geological ‘formations’, basically several deposits containing some common properties. The implication of this approach is not archaeologically very clear yet, but the ‘formation’-wise arrangement of the Deccanese prehistory certainly deserves a mention at this place.

Climatically, the late Pleistocene picture in the Deccan seems to be that of semi-aridity till the end of the Pleistocene. There was a rejuvenation of streams in the early Holocene, and obviously the climate then became wetter. Interestingly, the calcrete formation antedating the gravel with Acheulian tools in the Malaprabha valley has yielded the following information: ‘Nontronite and traces of kaolinite at the base of the calcrete and an increase in the content of montmorillonite and dolomite upwards in the profile are indicative of a transition in rainfall regimes from higher to lower during its formation.’²²

Not much is yet known about the stratigraphy of quartz pebble tools from the Palghat district of Kerala, the only area with reported palaeolithic tools in the State.²³ In Tamil Nadu, palaeolithic research is centred in the area near modern Chennai²⁴ and here the stratigraphy has been made explicit by Shanti Pappu:

We are not sure how this stratigraphy relates to the sequence offered by K.D. Banerjee in the cliff section near Budidamanuvanka, which showed the following succession: shale (bedrock) > clay (handaxes and cleavers) > detrital laterite (post-Acheulian flake industry) > brownish silt (sterile) > surface (microliths).

Both coastal and inland Andhra are rich in palaeolithic sites. Various Andhra districts (Cuddapah, Kurnool, Chittoor, Nalagonda, Nellore, etc.) have been surveyed for palaeolithic remains, the sequence showing the usual two-fold gravel–silt succession. However, the Kurnool area of Andhra possesses evidence of cave-sites occupied during the upper palaeolithic stage and the use of bone tools during this period, the only such categorical evidence available from the subcontinent (the Sanghao cave in the north-west has upper palaeolithic occupation but no bone tools). In the Visakhapatnam area of the coast Acheulian industries have been related to an old sea level which is 7+ m above the modern mean sea level.²⁵

Up along the east coast, Orissa is rich in palaeolithic remains too, some of her more prominent areas in this regard being the districts of Mayurbhanj, Dhenkanal, Sundargarh, Sambalpur, Phulbani, Bolangir, etc. In the Mayurbhanj area, the sequence above the bedrock shows the succession of mottled clay > lateritic conglomerate or what has been called a secondary laterite > and a sandy deposit or a thin layer of clay, whereas in many other areas the sequence shows a two-fold gravel-silt succession. All the three phases of the palaeolithic occur widely.²⁶

The western part of West Bengal, which is an upland region as opposed to the Gangetic delta, contains a large number of palaeolithic sites in the districts of Birbhum, Burdwan, Bankura, Purulia and Midnapur, primarily within the framework of the following section:

However, along the Tarafeni river in Midnapur, and possibly along a few rivers in Purulia. one finds the usual gravel-silt succession too, the lower palaeoliths being confined to the cemented gravel at the base of the sequence. The only climatic inference available is in the context of excavations at the microlith-bearing site of Birbhanpur where a comparatively dry climate has been suggested for the end phase of the Pleistocene, with the earlier strata being deposited ‘under conditions of large quantities of moisture and a high mean annual rainfall with a downward percolation of water’.²⁷

The entire Chhotanagpur section of Bihar and a few areas beyond this plateau have yielded palaeolithic sites. The Chhotanagpur section comprises the former districts of Singhbhum, Ranchi, Palamau, Hazaribagh–Giridih and Santal Parganas whereas the non-plateau sections comprise the Nawadah plateau section of the Gaya district extending up to Rajgir and the Kharagpur forest area of Munger. The palaeolithic stratigraphy of Bihar varies widely, but in eastern Singhbhum it is basically an extension of the bedrock > mottled clay > lateritic conglomerate > sandy deposit section that we have observed in the western part of West Bengal and the Mayurbhanj–Keonjhar section of Orissa. At a number of places in Singhbhum itself, the two-fold gravel–silt cycle occurs as it does in other areas of Bihar. The site of Paisra in the Kharagpur forest of Munger has been excavated, and the concentration of upper palaeolithic sites in the Damin area of the Rajmahal hills deserves notice. Here, rich upper palaeolithic levels can be observed in the surface soil overlying the local bedrocks. Another area which has shown a rich upper palaeolithic assemblage is Palamau.²⁸

The palaeolithic remains of UP are basically concentrated in the Lalitpur–Jhansi area, the Giri–Govardhan area near Mathura, the Singrauli area beyond Renukut and the Belan valley section near Allahabad.²⁹ The gravel–silt succession has been noted in Lalitpur, Singrauli and elsewhere, but the area where detailed stratigraphical studies have taken place is the Belan valley, first by G.R. Sharma and his group and later by Williams and Clarke. The Williams–Clarke stratigraphy for the middle Belan valley is as follows:

More positive reports from Assam and elsewhere in the north-east are needed before that region can be included in the palaeolithic distribution map of the subcontinent. An upper palaeolithic industry in fossil wood has been found in the Lalmai hills of Bangladesh and in the Haora and Khowai river valleys near Agartala in the western part of Tripura. Implements—handaxes, handadzes, blades, burins, backed knives, points, etc. made of a type of fossilwood which possesses a flinty core beneath the fibrous woody surface—were found within the depth of 1 m below the present surface. This Lalmai–Tripura fossil wood industry has been found to be identical with the Late Anyathian 2 fossil wood industry of the upper Yarrawaddy valley of Myanmar (Burma). This identity brings to light the existence of an extensive and distinctive upper palaeolithic situation covering the India–Bangladesh–Myanmar borderlands.³⁰

Finally, Gudrun Corvinus has placed on firm footing the palaeolithic prehistory of Nepal. The stratigraphic evidence comes mainly from the Dang and Deokhuri valleys in the Siwaliks of western Nepal.³¹