In this article we will discuss about the history of Sohan.

Between the river Indus and Jhelum a stretch of nearly 100 km is bound by four discontinuous mountain ranges. These are the Himalayas in the north, the Salt ranges in the south, Pirpanjal in the west and an extension of the Siwaliks in the east. This raised plateau is named Potwar. Rawalpindi lies almost in the centre of this plateau.

The entire plateau is composed of almost 25,000 feet (approx. 7,600 meters) thick Tertiary sediments almost entirely of lake origin. The top six layers of this huge deposition which is normally referred to as ‘Siwalik divisions’. The river Sohan or Soan (Sobhana in sanskrit) started flowing on this plateau in the Pleistocene period when the trough was almost filled up and there was no lake in existence any more.

There were periodic earthquakes and the region was technically very unstable and, under these circumstances, the older beds of the river often got pushed upwards and the river had to form a new bed and follow a new course at a lower level. The process having repeated itself on a number of occasions, a terrace-like feature is exposed on both the banks of the river.


De Terra and Paterson brilliantly correlated the entire geo-morphological process with the advance and retreat of the Himalayan glaciation. Thus, in an indirect way the Alpine chronology could be imported by them to India. Since a large number of Palaeoliths could also be collected by them, a cultural succession of a kind hitherto unknown for India could be established.

De Terra and Paterson explored the region around Potwar and the valley of Kashmir. Several types of moraine debris, locally called ‘Karewas’ were identified in the vicinity of the Dal Lake by these experts. It was demonstrated that the Karewas at a place called Malshahibagh in Kashmir record the earliest Himalayan glaciation. In comparison the Tatrot deposits, although not formed by glacial effect, were found to have remarkable petrological and also fauna similarity with the Malshahibagh Karewas.

This led them to ascribe the Tatrot zone in Potwar as belonging to the first Himalayan glaciation. Further, upto Dhokpathan the fauna in Potwar did not contain Elephas, Equus and Bos, which appear suddenly in the Tatrot zone and continue onwards. This was taken to mean a Villafranchian faunal ‘break’ and hence marking the beginning of Pleistocene.

Subsequent deposits in the Potwar region have also been likewise compared with the glacial deposits in the Kashmir area and a tentative correlation was derived as regards to both the kind of deposition as well as their faunal contents. Interestingly enough, after the Potwar plateau was filled with the lacustrine deposits of Pinjaur, the river Sohan started flowing over this newly formed plateau.


During the advance of the Himalayan glaciers huge debris was pushed along the course of the glaciers and this was taken down further along the course of the river by the melt water as it descended to lower altitudes. Further, at the lower altitudes the glacial phenomenon is believed to have resulted in a period of heavy rain fall or pluviation. Thus, what the melt water could not carry could now be flushed out by the swollen river fed by pluviation.

Hence, in Potwar region thick deposits (occasionally as much as of 500 meters thickness), made up of these concretized boulders, were laid down. The fauna recovered from these boulder conglomerates are designated typologically as Middle Pleistocene. These include open forest and moisture loving creatures such as elephants, horses, buffaloes, wild cattle and hippopotamus, turtle and crocodile.

In the subsequent phase, there was degradation in the river and, consequently, a large expanse of the earlier bed was exposed by erosion. This rhythmic oscillation seems to have continued with the alternate climatic cycles. This changing regime of the river is termed as aggradation (for raised water level i.e., pluvial climate) and erosion (for depressed water level i.e., inter-pluvial climate) phases. Several tributaries of the Indus are recorded to have formed 4-55 terraces representing these phases of climatic fluctuations.

At places like E. Africa the earliest known creation of man occur beyond 1.58 my. These can be demonstrated as having gradually modified into handaxes-the next group of types. In Europe where Lower Palaeoliths emerge around 0.6 my, there is an indication of these forming a regional trait parallel with handaxes. In India Lower Palaeolithic does not emerge before 0.2 to 0.3 my and here these occur as an integral part of handaxe culture.

The Zone TD is aggradational and is composed of boulder conglomerate. The fauna associated with this is typical Middle Pleistocene type and is commonly referred to as ‘Narmada fauna’. This deposit is comparable with the Karewa deposits belonging to IInd Himalayan glaciation in Kashmir.

The Zone T1 is cut during the next erosional phase and the deposits on it, as such, can be dated to the IInd Himalayan interglaciation period. Being erosional in nature it is mainly constituted of redeposited boulders which are found in a loose form in this terrace.

The Zone T2 is aggradational and hence is datable to the subsequent pluviation, i.e., the period of the occurrence of IIIrd Glaciation in sub-Himalayan Kashmir. It is a height of 120′ from the present river bed and has two distinct units. The lower layers are gravelly and are termed as Potwar basal gravels. This is overlaid by a fine layer of loess or silt which is called Potwar Loess.

The Zone T3 is an erosional terrace and hence shows only redeposition of earlier material. It is equated with the IIIrd Himalayan interglacial phase.

The Zone T4 records the last aggradational phase of the river Sohan. This deposit is mainly constituted of a mixture of pink silt, sand and fine gravel. In keeping with the postulated climate sequence this is equated with the last and the final glaciation phase in the Himalayas. Hence, it also marks the final deposit of Pleistocene.

The Zone T5 is the last terrace recorded at Sohan and it is interpreted as belonging to the early Holocene.

Thus, the second glaciation onwards upto the end of Pleistocene the Sohan terraces provide a very clear climatic and hence chronological succession. A large number of tools of prehistoric man were also collected by the authors in the course of their explorations of the region. These are very briefly described by De terra and Paterson.

It is important, at this point, to mention that a strict typotechnological approach in describing Palaeoliths was yet to be grounded as an accepted methodology of prehistory at this period of time. In spite of this, the description of the tool types and their techniques of manufacture are in many ways commendable.


The tools collected from the surface of the boulders of the TD terrace have been grouped together with a view to propose a cultural stage similar to what Gabriel de Mortillet proposed for the Somme valley in France. The latter proposed the term Pre-Chellean and De Terra and Paterson proposed the term Pre-Sohan.

These tools were picked from at least ten localities. In the language of the authors, these were, “Big flakes made from quartzite, all in very worn condition, with large, plain, unfaceted striking platforms mostly at low angles, varying from 100° to 125°. The bulbs are flat, but the cores (bulbs) are well developed, some of them very large.

The upper surface is usually unflaked except for one or two small irregular scars. There is no secondary working on any but one specimen from Kallar; where small flakes of later date have been struck from the main upper surface”. The absence of any working in these flakes made them a suspect and many alter commentators discarded them as pseudoartifacts.

Both-Movius in 1944 and Fairservis in 1975 however, felt that these could be the manufacturing debitage of an Early Sohan industry which was described from the next terrace (T1). In case one accepts them as a part of Early Sohan then Pre-Sohan would definitely qualify as one of the earliest known human cultures in the whole of Euro-Asia.

Early Sohan:

From a large number of localities tools were collected from within the T1 context. These, according to the authors were datable to the IInd Interglacial. Most of these tools were prepared on good size pebbles of various shapes and sizes. Some flakes were also described. The entire collection was given the cultural name of Early Sohan. Before describing the tools from this group the authors tried to form an internal chronology by dividing the collection in three groups based on the state of preservation.

Early Sohan A:

Specimens collected from the T1 context and found to be heavily worn and deeply patinated in brown to purple colour were grouped in this category. The number of specimens falling within this category was, however, not numerous and hence a true assessment of a true type preponderance will neither be possible nor free of fault.

The specimens described from this group consist of several discoid cores with only a few flakes removed from the borders and a rather big unifacial chopper, again produced by the removal of minimum flakes. No description of flakes is to be found in this group.

Early Sohan B:

It comprises of specimens which are as deeply patinated as in Gr A, but are much less worn. In this phase numerous pebble tools of various kinds are seen to occur. These are prepared on moderately big pebbles and in almost all cases maintain large areas of original pebble cortex. Real massive pebbles, as found in Gr A, although in a single sample, are no longer found.

The numerous associated flakes continue to be high angled and unfaceted but some among them show further flaking. In this regard these flakes compare well with the Clactonian flakes of England. The cores are more neatly worked and three essential types have been identified in them. These are choppers, chopping tools and discoid cores. Much of these flakings-both on the pebbles as also on the flakes-show a degree of control introduced by adopting the step flaking technique.

De Terra and Paterson identify two different techniques of pebble flaking. They call them as -(i) Flat based and (ii) Rounded pebble technique. In the former a flat under-surface is first obtained (or may be only such pebbles are selected which have a fractured flat under surface). Subsequent flakings are done with this flat surface as the platform.

Thus, when a unifacial chopper is prepared by this technique, the working border is always straight. In the rounded pebble technique the flakes are removed from the original rounded pebble surface without any adequate platform. Tools prepared by this technique either have a convex working border or when the blow is delivered from opposite borders, a pointed end. In the latter case, unless unifacially worked, these specimens compared well with primitive pebble hand axes known from elsewhere in India as also broadly with many specimens of Europe and Africa.

Early Sohan C:

It constitutes specimens which are least worn and patinated. Pebbles again dominate the scene, with choppers, chopping tools and discoid cores forming the main typological categories. Some of these pebbles have also been described as ‘side choppers’ perhaps because the working border runs along the length rather than the breadth of the pebble.

The flakes show a larger degree of modification when compared with their counterparts in Group A and B. Single faceted flakes with evidence of greater amount of primary dressing appears for the first time, although these still remain without any typological retouchings.

Late Sohan:

Specimens collected from the T2 context are termed as late Sohan. Since T2 maintains two stratigraphic units- the Potwar basal gravel forming the older layer and the Potwar loess forming the younger layer-the tools from the former group are separately designated as Late Sohan A and those from the latter group as Late Sohan B. These tools are found from several localities within the Sohan- Indus region and virtually merge with another implementiferous zone, around 17 Km.

South of Rawalpindi, called Chauntra. In the latter zone a deposit similar to the Potwar basal gravel records more than 100 specimens of handaxes and cleavers. It would, therefore, seem that a non-biface tradition distantly comparable to the Clactonian stage of England occur in Punjab alongwith the biface tradition somewhat parallely.

Both late Sohan A and B show almost similar core tools in them. The types are all finished in medium sized pebbels of various shapes with neat primary and secondary flakings. The secondary flakings usually show the mastering of a resolved flaking technique. The types include a number of chopper and chopping tool besides the usual types of discoid cores.

Often flattened circular pebbles are chosen for these and alternately flakes are removed from both the surfaces in a circular manner. There are some specimens described as ‘turtle back type’ which are flat split pebbles in which the fractured surface has been flaked all over so that the obverse remains as an evenly convex pebble cortex, with some minimum flaking done at one end or border.

The flakes in Late Sohan A and B show progressive development. Levalloise flakes and tortoise cores are both identified. Some of these flakes are also retouched into such types as side scrapers or similar cutting tools. Blades or ‘bladish’ flakes are also reported from Late Sohan B. Looking at the total character of the Late Sohan, one is left with no doubt that it is essentially a flake dominated industry.

But the core component and the chronological ascription of the terrace would make it difficult to accept it as Middle Palaeolithic as some authors seem to propose in later works. De Terra and Paterson felt that the Chauntra site shows a mixture of Late Sohan with Abbevillo-Acheulian traditions.

The material is divided by them into three groups on the basis of the degree of weathering evidenced on the collected specimens:

(a) Well-rolled and crude specimens – This group includes Abbevillian handaxes, some pebble cores and a few massive flakes without much secondary workings.

(b) These included medium-rolled specimens which show at least a lesser degree of weathering than the earlier group. Tools falling within this group include Early Acheulian and Middle Acheulian handaxes and some cleavers.

(c) This group includes both fresh and unrolled specimens. Besides the Late Sohan type core and flakes Paterson illustrates some cordates, pyriforms and ovates which are the thin bifaces that characterize Upper Acheulian in France. Cleavers are also recorded in this group. Surprisingly, none of the flakes or flake types which normally predominate an Upper Acheulian industry is described by the authors.

It is important at this point to mention that a later exploration on the Indus, conducted by Paolo Graziosi, has yielded almost an entire Upper Acheulian industry, although from a much lower region. Higher up in the Shimla Hills again, recent explorations by G C Mohapatra, have yielded a typical Upper Acheulian industry.

B B Lal’s explorations on the Beas-Banganga region in Kangra, on the other hand, almost duplicate the findings of De Terra and Paterson. Lai found the same number of terraces at almost comparable heights with almost an entirely pebble dominated industry. In a collection of 52 specimens only 4 are counted by him as bifaces.

It will, therefore, appear that at Chauntra we are probably dealing with the northernmost limit of an Acheulian intrusion which at least for Punjab has a further western origin. The Sohanian group, on the other hand, is fairly widespread further north in Tadjikistan and hence may be independent of the Acheulians.

In 1964, K V Soundara Rajan had proposed that Sohan, in the fact of its not showing appreciable internal dynamics, might be representing an ‘endogenous’ culture. It was around the same time that an alternate hypothesis for the pebble culture of India was mooted by A P Khatri.

He claimed that a prebiface existence of pebble tools is stratigraphically evidenced at Mahadeo Piparia on the Narmada and hence pebble tools in India must have evolved into bifaces independently after the pattern of Olduvai Gorge in East Africa. If one can extend Khatri’s argument into Punjab, Soundara Rajan’s proposal would mean a stagnation of an archaic element into the Punjab plateau.

Seen in the present light of our knowledge most of these opinions would seem unlikely. Firstly, no pre-biface layer from Narmada has so far been found. Secondly, the biface layer at Narmada can at the most be dated to early Eemian or Upper Pleistocene if one has to go by all available evidences.

Hence, even if Khatri did find a pebble tool underlying the bifaces, one cannot expect the Olduvai pattern of cultural evolution to repeat the same pattern in India after a lapse of nearly 1.6 million years. In the context of Punjab, the recent studies in Himalayan glaciations seem to bring out certain well-attested facts which not only complicate the earlier evidences at both Sohan and Banganga but also nearly demolish their entire chronology and hence cultural succession proposal.

It would appear that the only glacio-fluvial load that came down to the Potwar region from Kashmir was the Boulder Conglomerate and this episode is estimated to have taken place around 1.9 million years ago. All the terrace deposits are in fact derived from this deposit during a process of extensive and repeated tectonic motion that rocked the entire area during early Pleistocene.

If these terraces are not created by the river the entire logic of a climatic succession fails-and hence the tools from almost every terrace become broadly contemporaneous. The lack of internal dynamics of Sohan culture as alluded to by Soundara Rajan will, therefore, seem to be explainable at one stroke.

One might ask at this stage about the need of continuing with De Terra’s scheme of succession in the North-Western Prehistory, especially in the light of the statements from Dennell and Rendell in the recent years.

To quote, “We conclude that an entirely fresh start needs to be made in investigating the Plio-Pleistocene and Palaeolithic sequences of northern India and Pakistan. What has to be kept in mind is that the area is highly unstable tectonically and so it is highly improbable that the geological deposits contain a sequence reflecting climatic change alone. Secondly many units are time-transgressive and thus units such as the Boulder Conglomerate cannot be used as marker horizon. Instead each local sequence has to be dated independently by palaeomagnetic and radiometric dating techniques.”

It should not be entirely unwise to recommend that we continue with the format suggested by De Terra, if not for anything, at least for historical reasons.

Further the evidences reported by Ranov from Tadjikistan and also the recent discoveries at Dina, Jalalpur and Riwat in north west Pakistan definitely indicates a fairly wide spread Lower Palaeolithic occurrence in the entire region. It is also important to note that some of the absolute dates for these occurrences are very impressive. For instance Riwat is dated to C. 2.0 m.y. and Dina and Jalalpur are put within 500,000-700,000 bracket.

Thus, one might say that both the cultural features and the antiquities recorded by De Terra may not be incorrect but their inner succession seems to be grossly incorrect.

Home››History of India››Sohan››