This document summarizes the Quaternary geology and prehistoric environments of the Son and Bela valleys in north central India based on fieldwork from 1980-1982. Four widespread alluvial formations ranging from Middle Pleistocene to Holocene were identified and partially dated. The oldest formation contained Lower Paleolithic stone tools, while the youngest contained Neolithic artifacts. A buried volcanic ash layer from the Toba eruption around 75,000 years ago was also found. The formations suggest alternating warm/wet and cold/dry periods that impacted vegetation and river behavior over the Quaternary.
This document summarizes the Quaternary geology and prehistoric environments of the Son and Bela valleys in north central India based on fieldwork from 1980-1982. Four widespread alluvial formations ranging from Middle Pleistocene to Holocene were identified and partially dated. The oldest formation contained Lower Paleolithic stone tools, while the youngest contained Neolithic artifacts. A buried volcanic ash layer from the Toba eruption around 75,000 years ago was also found. The formations suggest alternating warm/wet and cold/dry periods that impacted vegetation and river behavior over the Quaternary.
This document summarizes the Quaternary geology and prehistoric environments of the Son and Bela valleys in north central India based on fieldwork from 1980-1982. Four widespread alluvial formations ranging from Middle Pleistocene to Holocene were identified and partially dated. The oldest formation contained Lower Paleolithic stone tools, while the youngest contained Neolithic artifacts. A buried volcanic ash layer from the Toba eruption around 75,000 years ago was also found. The formations suggest alternating warm/wet and cold/dry periods that impacted vegetation and river behavior over the Quaternary.
This document summarizes the Quaternary geology and prehistoric environments of the Son and Bela valleys in north central India based on fieldwork from 1980-1982. Four widespread alluvial formations ranging from Middle Pleistocene to Holocene were identified and partially dated. The oldest formation contained Lower Paleolithic stone tools, while the youngest contained Neolithic artifacts. A buried volcanic ash layer from the Toba eruption around 75,000 years ago was also found. The formations suggest alternating warm/wet and cold/dry periods that impacted vegetation and river behavior over the Quaternary.
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online from Scribd
Download as pdf or txt
You are on page 1of 27
MEMOlfiS GEpLOGICAL SOCIETY OF INDIA
No. 32. PP. 282-308
Quaternary Geology and Prehistoric Environments in the Son and Bela Valleys, North Central India M.A.J. WILLIAMS AND M.F. CLARKE! Mawson Graduate Centre for Environmental Studies. University of Adelaide. Adelaide. South Australia 5005. Australia I Fonnerly. School of Earth Sciences. Macquarie University. New South Wales 2109. Australia Abstract Four widespread alluvial formations ranging in age from Middle Pleistocene to Holocene have been identified. described. mapped and partially dated in the Son and Belan valleys of north central 'India (covering parts of Madhya Pradesh and Uttar Pradesh). The three oldest formations are capped by aeolian deposits of very fine sandy clay which appear to be slightly to substantially reworked loess. These three formations' contain reworked as well as primary-con.text Lower. Middle and . Upper Palaeolithic stone tools. The youngest formation is of Holocene age and contains primary- context Neolithic artefacts. Beneath the Late Pleistocene Baghor Formation is a buried channel-fill of volcanic ash which was erupted from Toba volcanic caldera in nonhern Sumatra 75 kyr ago. Introduction For twenty or so years prior to 1980, the late Professor G.R. Sharma and his colleagues from Allahabad University had identified and mapped a large number of Lower, Middle and Upper Palaeolithic sites in the Son and Belan valleys of Madhya Pradesh and Uttar Pradesh in north central India (Shaima et al. 1980; Sharma and Clark 1983). This pioneering work deserves high praise: the survey area was vast, poorly served by access roads, with large tracts of dense thorn scrub. Some of the more detailed maps dated from the last century. Air photos were not available and so were never used. In 1980 systematic excavation of selected Lower, Middle and Upper Palaeolithic sites in the Son valley was initiated by G.R. Sharma in collaboration with a team from Berkeley directed by Ptofessor J. Desmond Clark (Sharma and Clark 1983; Kenoyer et al. 1983). Working very closely with the archaeologists, M.A.J. Williams was responsible for describing and sampling a number of key Quaternary stratigraphic sections (Fig. 1) and for attempting to reconstruct the allu vial history of the Son and Belan during the 1980 and 1982 field seasons (Williams and Clarke 1984; Clark and Williams 1986, 1990). The following account is distilled from this preliminary and mostly unpublished fieldwork and from subsequent radiometric dating and sedimentological analyses. Sillll 1'1. ,/( _-I- to K......"., __ '-.,.____ I I I i --/
V "> AmIlia I "-! 9 -- " (, 4 'Meln Belan ' ,/ i / - SeotiN
/' \$> Deoghat (( 5
5km ,1-- J 82'O!) ------:". ---'-' "-"----'.-.::..' 'II 1882 plllogl." aectiana 82"15' , ." Kymero ,- \ _'fang I, G4''':1'I'/ !', {I G4,;",{f) \ Bighor SIte. II '--i.'" ,r r, ________ 1II:::as: to1 __ "'._. 'Nala --- - - Ie- _.- iil1' _ - c.mp;; Ge , ... 7 '",3:.ll10 ""C:>-HAWAI. - -Me",;'r','- Gl.L-., G12', . .s Bighor Bi h ,--'--- o. /!"" ..... _/, c:.tiar /- .q;. -; <,., '1 ,/ p 5km Fig. 1: Location maps of stratigraphic sections discussed in the text 1 I
10 c:: ;l>- I ;} n fi
2
284 M.AJ. WILLIAMS AND M.F. CLARKE One feature of Indian Palaeolithic archaeology noted at this time by both Jacobson (1975) and Paddayya (1977) was an undue preoccupation with fossil bones and stone artefacts recovered from alluvial gravels. The theoretical dangers of this approach are fairly evident. Apart from the fact that gravels are not easy to date except in a relative sense, there is no guarantee that the bones and stone tools are of one age, even when they appear relatively unabraded. The identification of three gravels (designated I, II, III in ascending stratigraphic order) in both the Belan and Son valleys, associated respectively with Lower, Middle and Upper Palaeolithic stone tool assemblages (Sharma 1975), was a useful initial working model. Unfortunately, image and reality were sometimes confused, and observations became clouded with circular reasoning. A gravel could be assigned a relative age on the basis of included artefacts; and artefacts could be dated according to whether they belonged to gravel I, IT or Ill. 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. Results of the 1980 Season's Geological Programme
The outcome of the first season's stratigraphic work in the Son valley (Williams and Royce 1982, 1983) was the recognition of four main alluvial fonnations (Fig. 2) which, for the sake of simplicity, are here called A,B,C and D (for actual fonnation names see Table 1). 3. Terminal Pleistocene sands and clays Lower Proterozoic metasediments 1. Middle Pleistocene gravels and clays Present-day channel sands and pointbars of River Son 2. Upper Pleistocene gravels, sands and clays Fig. 2: River terraces and Quaternary geological fonnations in the middle Son valley, Madhya Pradesh, north central India (after Clark and Williams 1986: Fig. 3) THE QUATERNARY OF CENTRAL INDIA 285 Table I: Quaternary alluvlll formations in the Sorp valley I A. Sihawal Formation Middle Pleistocene colluvial-alluvial clayeygtavels and fanglomerates with Lower Palaeolithic artefacts. Capped by generally sterile grey and yellow mottled very fine sandy clay of aeolian provenance. B."Patpara Fonnation , (?)Late Middle to early Upper Pleistocene fluviatile red-brown clayey fine gravels, gravelly clays and granule sands with "transitional Lower to Middle Palaeolithic artefacts. , C. Baghor Formation' Upper Pleistocene to Early HolQ!:!ene fluviatile pale yellow-brown coarse sands and calcreted gravels merging laterally into and/or capped by horizontally-bedded silts and clays. Some Middle Palaeolithic artefacts near base, but mostly Upper Palaeolithic. with surface and near-surface concentrations of Mesolithic and younger material. D. Khetaunhi Formation Middle to Late Holocene fluviatile fine sands, silts and clays. Neolithic artefacts. Note: Between Formations C and D there are minor remnants of an end-Pleistocene/earliest Holocene dark clay with fresh Upper Palaeolithic and blade-cores. The oldest Formation (A) consists of local pediment gravels capped by several metres of archaeologically sterile aeolian clay. In and on the gravels are fresh and abraded Lower Palaeolithic bifaces of Late Acheulian aspect. . Disconformably over A is Formation B: an upward-coarsening sequence of red- dish-brown gf&velly clays and gravelly sands in which are interstratified Late Acheulian to Middle Palaeolithic artefacts, often quite fresh and sharp. The eroded surface of B is the base of Formation C. Formation C is perhaps of greatest stratigraphic and archaeological interest. Near the main river it consists of two distinct members. The Coarse Lower Member consists of cross-bedded and planar-bedded sands and gravels which contain Middle to Upper Palaeolithic stone artefacts and an abundant fossil vertebrate fauna. Near the river the Coarse Member is usually capped by horizontally-bedded silts and clays. The distinc- tion between a Coarse Lower Member and a Fine Upper Member is not always absolute: on occasions sands and clays may be interstratified. Away from the river both members of C merge laterally into massively-bedded aeolian clays. After the deposition of C, the river cut down into its floodplain, which is some 30-35: m above the low-water river level and 15-20 m above normal flood level. Formation D represents a minor phase of Holocene aggradation and consists of hori- zontally-bedded fine sands, silts and clays. Alluvial History and Quaternary Climates: Initial Interpretation A simple model was proposed by Williams and Royce in February 1980 (Williams and Royce 1982, 1983) to account for the alluvial stratigraphy of Formation C. The argument was based upon four main premises: (a) Indian rivers with densely vegetated catchments tend to be high-sinuosity. suspension-load rivers; (b) Indian rivers with sparsely vegetated catchments tend to be low sinuosity. bed-load rivers; (c) vegetation in north central India would be dense during warm, wet interglacial or postglacial times; and 286 M.A.J. WILLIAMS AND M.F. CLARKE (d) vegetation in north central India would be sparse during cold, dry, full g l a ~ a l times. Our 'initial interpretation of the fine upper member of alluvial Formation C was that it represented vertical accretion of fine overbank deposits during warm, wet postglacial times (Terminal Pleistocene to Holocene), in contrast to the Coarse Lower Member. This we attributed to bed-load aggradation (lateral accretion) by a more seasonal Son during the Last Glacial Maximum (c. 25 kyr B.P. to c. 15 kyr B.P.), when northern India was drier and colder than today (Singh et al. 1972, 1974; Singh and Agrawal 1976). Evidence from Indian Ocean Deep-Sea Cores Before reviewing the' aims and achievements of the 1982 season's geological programme, it is helpful to consider the overall pattern of Quaternary climate change throughout the Subcontinent. The best record of Late Quaternary climatic events in India is that provided by the surrounding Indian Ocean (Prell et al. 1980; Cullen 1981; Duplessy 1982). More limited data from on land include the Holocene pollen spectra of some of the Rajasthan lakes (Singh et al. 1974) and evidence from geomorphic and archaeological reconnaissance studies in Gujarat and Rajasthan (Allchin et al. 1978), none with adequate time control. Duplessy's (1982) elegant study of glacial to interglacial contrasts in the northern Indian Ocean offers encouraging support to our own attempts at reconstrcting Late Quaternary changes in river flow and sediment load in the Son and Belan valleys of Madhya Pradesh. He used differences in the planktonic foraminifera that lived in the northern Indian Ocean during the Late Pleistocene and Holocene to reconstruct the probable climates at those times. Towards 18 kyr B.P. the salinity gradient in the Bay of Bengal was very much steeper than today, reflecting a drastic reduction in freshwater input from the Ganga and Brahmaputra. This inference is consistent with our suggestion, based on sedimen- tary evidence, that the Son was a more seasonal river at this time, with a sparsely vegetated catchment. The upwelling that is now a feature of the southern coast of Arabia (and a contributor to aridity inland) had also disappeared at this time, indicating that the southwest summer monsoon winds were not particularly strong during the Last Glacial Maximum. A weakened summer monsoon and much reduced summer rainfall would account for the aridity evident in northwest India during the Last Glacial Maximum (see Goudie et al. 1973 and the thermoluminescence dates of Rajasthan dunes obtained by Singhvi et al. 1982). In contrast to the weak summer monsoon, the northwest monsoon seems to have been stronger during the Last Glacial Maximum. Evidence includes the clockwise circulation pattern in the Bay of Bengal revealed by a tongue of low salinity water. The combined evidence from rivers, lakes, dunes and pollen spectra accords well with Duplessy's reconstruction of a drier, windier climate towards 18 kyr B.P. over much of India, with less rain in summer and a stronger winter monsoon than today (Williams 1985). During the Early Holocene the marine isotopic record indicates a reversal of the Late Pleistocene pattern of weak summer and strong winter winds (Duplessy 1982). By THE QUATERNARY OF CENTRAL INDIA 187 15 kyr B.P. the ice had disappeared below 5000 m in the northwest Himalayas (Singh and Agrawal 1976). Early Holocene cold upwelling water accentuated the aridity of the south coast of Arabia in response to the stroqger summer winds. The SW monsoon blew vigorously, bringing summer rain to swell the floods and fill the lakes of India. Runoff from the Indus, Ganga and Brahmaputra increased, and the rivers of central India became less seasonal. . . The lakes of Rajasthan began to fill shortly before 10 kyr B.P. With minor fluctuations they remained full and fresh until about 4-3 kyr B.P. after which they became saline (Singh 1971; Singh et al. 1972). To sum up, the Early to Middle Holocene climate of India was generally wet and warm, with heavy monsoonal rain in summer and moderate rain in winter. By about 4-3 kyr B.P. the Late Holocene desiccation of northern India was under way, aggra- vated by the impact of Neolithic herding, land clearance and cultivation. At the present time the margins of cultivation are advancing into the hitherto wooded footslopes of the Vindhyan hills in the middle Son valley, leading to renewed soil loss and gully erosion. The Toba Volcanic Ash in the Son Valley One important outcome of the 1980 fieldwork in the Son valley was the joint discovery by Williams and Royce on February 5, 1980, of a very pure volcanic ash filling a buried channel beneath the Lower Member of the Baghor Formation just below the confluence of the Son and Rehi rivers, on the Son left bank (Williams and Royce 1982). This ash was erupted from Toba volcanic caldera in Sumatra some 75 kyr ago and is discussed in more detail below. The Toba volcanic ash has aroused considerable interest among geologists in recent years (Acharyya and Basu 1993, 1994; Rampino and Self 1992, 1993; Mishra and Rajaguru 1994; Badam and Rajaguru 1994). Given current interest in the various Toba ash deposits found at an increasing number of localities throughout India, and the strong possibility that there are at least two chronologically distinct Toba tephra beds in India, a brief account of its initial discovery in the Son valley in 1980 may be of interest here. A more detailed description of the ash will appear elsewhere. The ash was first observed by the author and Keith Royce just below the Son-Rehi confluence on February 5, 1980. Our initial but erroneous impression was of a hori- zontally-laminated fine sandy dIatomite up to 3.4 m thick and at least 10 m wide. Interstratified within the horizontally-layered ash were lenses up to 45 cm thick of what we at first interpreted as a reworked very fine aeolian sand. The base and eastern margin of the ash deposit were obscured by the talus (lying at the foot of the cliff) in which the section was exposed. A horizontal layer of gravel up to a metre thick capped part of the upper surface of the ash. This gravel was itself in part eroded and replaced by a channel-fill of fine to coarse gravel up to 3 m thick, inset into the upper surface of the ash to a depth of 1-2 m. The enigmatic nature and provenance of the so-called "fine sandy diatomite" prompted us to embark on a detailed physical and chemical examination of the field samples after our return to Australia. Microscopic examination of our field samples by D.A. Adamson (Macquarie University) confirmed that the "fine aeolian sands" and "fine sandy diatomites" were 288 M.A.J. WILLIAMS AND M.F. CLARKE in fact the glass shards of a volcanic ash. Detailed grain-size analysis by G. McTainsh (Griffith University) revealed that all samples were similar, with modes lying within the 30-50 J.lIl1 range. Strontium isotope analysis by D. Whitford (CSIRO Division of Mineralogy, Sydney) indicated a 87SrJ86Sr ratio for our sample S44 of 0.71504 0.00006, consistent with an origin from Toba volcanic caldera in Sumatra (Whitford 1975). Two years before our 1980 field season in the Son Ninkovich et al. (1978) had described a widespread volcanic ash layer recovered from piston cores taken in the northeast Indian Ocean and the Bay of Bengal.Tlwy concluded that this ash erupted from the Toba volcano in northern Sumatra some 75 kyr ago. They also noted that this Late Pleistocene eruption was the largest magnitude explosive eruption recorded for the entire 2 million years of Quaternary time. The map they produced shOwing the distri- bution of Toba tuffs in Sumatra and the equi valent ash layer in deep-sea cOJ;'e${NJnkovich et al. 1978: Fig. 1) revealed a fan-shaped ash plume radiating out to the northwest and west-northwest at an angle of about 50. Extrapolation of the deep-sea Ilh)ayer clearly suggests that every part of India would have received some ash, as is becoming more and more apparent (Acharyya and Basu 1993: Fig. 1): ' Samples of the Son volcanic ash were also sent to C.A. Chesner (Eastern Illinois University) who confinned that they belonged to the youngest of the Toba eruptions (Rose and Chesner 1987; Chesner et al. 1991). Current work on this eruption, now dated to roughly 75 kyr ago, is increasingly focused on its cooling impact on the world climate at that time (Rampino and Self 1993). It is probably no coincidence that the inception of the rapid build-up of the Laurentide Ice Sheet and of a major cooling signal in the Vostok ice core are also dated to about 75 kyr age (Jouzel et al. 1987; Williams et al. 1993). Aims of the 1982 Geological Programme In the Belan valley it was apparent that we n,eeded to examine the status of the three tool-bearing gravel units identified hitherto. A rapid survey of certain Belan valley sections at Deoghat, Mahagara, Chillahia and Chopani-Mando in February 1980, in the company of D. MandaI and B.B. Misra, revealed both similarities with and contrast to the alluvial sequence mapped by us in the Son valley. It was thus quite possible that each of the three units designated as gravel I, IT and III was more complex than suspected, and could well be polygenic. The Ganga alluvium at Sarai Nahar Rai (Shanna 1975) contained Mesolithic artefact assemblages comparable to those at Chopani-Mando in the Belan valley. Apart from its potential for Quaternary stratigraphic correlation, this site might allow us to detennine when the Holocene Ganga began to cut down. Shell-bearing silts inspected by us at Kurha in the Ganga valley and at Deoghat in the Belan valley, suggested that it may be possible to establish a partial radiocarbon chronology of Late Quaternary events in these two Valleys. However, isolated shell dates would not in themselves be of much help in reconstructing the pattern of pre- historic environmental changes. Our aim was therefore to detennine accurate strati- graphic histories for each major river valley in this archaeologically-rich sector of north central India. 290 M.A.J. WILLIAMS AND M.F. CLARKE Results of the 1982 Season's Geological Programme (a) Alluvial Stratigraphy of the Ganga and Yamuna Rivers at and Above Their Confluence Three major sections were logged and sampled for radiocarbon dating (Fig. Sa: sections 1 to 3). Section 1 was at Jhusi, 1 Ian upstream of the Ganga-Yamuna confluence, on the Ganga leftbank. About 22 m of yellow-brown calcareous silts were capped by a protohistoric mound. Pottery from the base of this mound antedates the Northern Black Polished ware dating to 600-700 B.C. suggesting that the base of the protohistoric mound may be about 3 kyr old. Yellow-brown silts similar to the Jhusi silts are also known at Kuiha and Sarai Nahar Rai where they contain microliths dating to c.1O kyr B.P. (Sharma 1975). Beneath the Jhusi silts there are at least 5 m of cross-bedded reworked carbonate nodules, with foreset dips of 23_32 and palaeo-current directions of 35 (NNE). SYMBOLS USED IN GEOGRAPHICAL SECTlONS U ..... 1. JVV'V -.Ionel boundIIry b IIr--. congIoIMr". Iherp boundery bu buff ~ debrielow "--'1 boWldery " .,., m .. wi In .. ...
...... .:: D UncI ....... ' ... 00 bone 0 -.... D . ~ >: =. and t. 8ItIIId 01 ...
.............. R .... ~ .. - - ..... ..,. W .....
~ ., W .., ... Y ,..... 'M. ........ (III' ......... rn IDMI FL ........ LB ....... K radio ....... RB ..... t8IuI, [] dIIUIIed 1n1C1D., .. , , .......... 1.2.3.* ....... c .. an Indu ~ -.., .. __ ~ " a i t Fig. 3b: Symbols used in geological sections THE QUATERNARY OF CENTRAL INDIA 291 Section 2 was on the Ganga left bank roughly 1 km upstream of the Sringaverpura archaeological site. The horizontal silts and silty bands exposed in the banks could be traced in continuous outcrop for at least a kilometre above and below the ancient city, the tank of which dates to the first centur.y A.D. (La] 1981; Lal and Dikshit 1978-79). The January 6, 1982 river level was 82.90 m AMSL; normal flood level at the Sringaverpura gauge is 87.98 m, and the maximum level attained by the 1978 floods was 88.68 m. There are thus at least ten metres of alluvium above the present flood- level. It seems probable, therefore, that the Holocene Ganga has entrenched roughly 10 m into its end-Pleistocene or Early Holocene floodplain. Such entrenchment is clearly not related to a fall in sea level, since the sea was rising steadily until about 6-7 kyr B.P. when it reached its present level. Section 3 on the Yamuna right bank at Telauli village 2-3 km downstream of Mau, in Banda District, has interesting.parallels with Ganga section 2. Both sections appear to belong to an upward-coarsening alluvial sequence, with massive horizontal beds of sand in the upper part of the section resting conformably over horizontally-bedded clays and loams. Present flood level in the Yamuna reaches within 4 m of the top of the logged section, but there is a further 4-5 m thickness of eroded deposit stratigraphically above the top of section 3, again indicating some 4-10 m of incision into the Late Pleistocene or Early Holocene floodplain. A shell-bed at +14.8 m above the January 10, 1982 river level was sampled for radiocarbon dating. Its 1"C age of 15,540170 B.P. (Beta-4788) thus allows us to specify a minimum age for final aggradation and a maximum age for river downcutting. All shell samples were tested by X-ray diffraction for possible recrystallisation from aragonite to calcite. Preliminary field identifications (for which we are indebted to Mark Kenoyer of Berkeley) suggest that there are at least five species of mollusc within the shell-bed: Parreyssia favidens (Benson): a small freshwater mussel Lamellidens marginalis (Lamarck): a large freshwater unionid lndoplanorbis exustus (Deshoyes): a small gastropod characteristic of terrestrial and swampy habitats Digoniostoma cerameopoma (Benson): a gastropod characteristic of swampy and terrestrial habitats Viviparus bengalensis (Lamarck): characteristic of ponds and run-on sites. (b) Alluvial Stratigraphy of the Belan River between Chopani-Mando and Amilia Eight major sections in the middle Belan valley were described and sampled in detail, and five others were sampled specifically for radiocarbon dating but were not logged in detail because they seemed to be lateral variants of sections already de- scribed. Simplified logs are given in Figures 4 and 5. At Mahagara (section 4, Fig. 4) the uppermost gravel unit contains shells dated to 10,030 115 B.P. (SUA-142). All shells were from the lower 40 cm of the cross- bedded gravels which form a bed up to 115 cm thick. Banked against the main section, with its surface 3.80 m lower than the top of, section 4, is a Neolithic midden (Mahagara Index Trench) of grey-brown (10 YR 31 3) gritty loam with lenses of rolled, poorly-sorted carbonate gravel, interbedded with broken potsherds. A perforated mussel shell in the north wall of the Index Trench was
BELAN RIVER : GEOLOGICAL SECTIONS 100
:[ RB A RB
200m upltrum 500m downstream 201-
4 b RB LB 10K. _ 3 5 b
, .. 18 I-
>- 16 I- .
. -:- 8 b
. -:-
! 14 I-
4 t'" 3 b 6 b t'" ... 12 I- > 5 rIl 10 f-- 4 b
3 b CI aI- I 0 2B b ' A
, : gap 5 !OJ o D ..... ?-.: v 4 Rb ('l 81- I 0 3 3
o 0 2 2 Rb
41- I 0 2Ab 0 =
2 b :;, 21- kD .. 1 b
1 b iii':" 1 1 w.. .. w..
OL- ...... <150m> bank, cliff 4 MAHAGARA 5 CHILLAHIA 8 DEOGHAT BRIDGE 7 Fig. 4: Quaternary stratigraphic sections, Belan river "'"' .f>. BElAN RIVER (abMdoned chann'l) : CHOPANI MANDO m - down - channel I4'-ChInneI - 14 ' t LB 13 LB 12
LB 11 II. 2 b
5 b rn
10 l- 5 b 0 c:: e .-=- > ..
tl 4 b tl .. 4 b 8 l- LB 1:-:- 1 Vb .. 7 3 ..
> " 7 I ....."l 0< 3 b 0 6 2Yb "'I !"l l!I!J 5 l- r-,- z 2Yb
4 l- [I I"" t"' u .... 3 l-' I .. '""
.... " u >
0 2 f- go IJ: ... " I': o 0 1 l- ..... . 1 bIdroc:k" . 0 0 L '--.J LJ .. Unit, cliff ........ 800m .......... 8m 200m bank. cllf SA 8B Be e N Fig, 5: Quaternary stratigraphic sections, Belan river, Chopani Mando 294 M.A.J. WILLIAMS AND M.F. CLARKE of interest in that similarly perforated mussel shells are used today forskinning green mangoes (B.B. Misra: personal communication, 16.1.82). It is highly probable that the Index Trench deposit accumulated rapidly, for there is no significant difference at one standard deviation between any of tht" four charcoal radiocarbon dates, despite the fact that the deepest and shallowest samples are up to 2.35 m apart vertically (see Table 2). Table 2: Radiocarbon dates from Mahagara Index Trench, Belan valley (data from Sharma et al. 1980: 199- 200) Sample, Lab. No. PRL-409 PRL-408 PRL-407 BAIIAu/ALLDIMGR-77-1 Age Depth below datum 1400 150 B.C. -1.15 m 1330 120 B.C. -1.25 to -1.35 m 1440 100 B.C. -2.4 to -2.6 m 1480 110 B.C. -2.8 to -3.5 m From these dates we can draw the following conclusions: (i) Until 10,030 B.P. or shortly thereafter the Belan and its tributaries were actively aggrading their flood- plain. (ii) Between about 10,030 B.P. and 3,430 B.P. (1480 B.C.) the Belan had cut down at least 7.3 m (3.5 + 3.8 m) into its Pleistocene floodplain. Beneath the 10,030 B.P. shell-bearing rolled carbonate gravels there are 12 m of brown (5 YR 4/4 to 7.5 YR 3/4) massive clay loams and fine sandy clays with up to 20% irregular pedogenic carbonate nodules - the probable source of the rolled gravels in the gravels capping this section. The absence of cut-and-fill structures, the lack of any obvious strtigraphic breaks, and the gradational nature of any colour or grain-size changes in the Mahagara brown clay loam prompted the inference that we were dealing with a loess. A sample collected from +16 m gave a thermoluminescence age of 29,900 ' 4500 (Alpha-897), indicating a Late Pleistocene age for this unit. Stratigraphically beneath the Mahagara brown clay loam was a bench of sandstone gravel and gravelly sand at least 8 m thick. Banked up against both gravel and clay loam was a recent terrace of very fine alluvial sand, the surface of which was' up to 11.2 m above the January 12, 1982 river level. The sequence of events at Mahagara was as follows: (i) Deposition of fluviatile gravels to +8 m (ii) Loess deposition (primary or secondary) to +19 m (iii) Pedogenesis during and after (ii) (iv) Local erosion of loess, concentration and redeposition of carbonate nodules towards \0 kyr B.P. (v) River incision of at least 7 m between 10 kyr B.P. and 3400 B.P. (vi) Deposition of recent sandy alluvial bench to +11 m (vii) Continued incision. The "Epi-Palaeolithic" assemblage (Upper Palaeolithic blades, non-geometric microliths) recovered by excavation of the 10 kyr B.P. gravels is broadly similar to the assemblage in archaeological layer 10 at Chopani-Mando (Sharma et al. 1980:143; B.B. Misra: personal communication, 12.1.82), as well as to that seen eroding from the uppermost beds of section 5 at Chillahia (Fig. 4). THE QUATERNARY OF CENTRAL INDIA 295 Chillahia section 5 (Fig. 4) was again on the Belan right bank, about 3 km upstream. The Belan flows westwards so that Chillahia lies east of Mahagara. The basal 2.5 m ohhe section consist of a brown (7.5 YR 4/4) clay loam with up to 50% calcium carbonate. Not seen by us, but rumetimes exposed by drought is an underlying conglomerate from which V.D. Misra and B.B. Misra collected one cleaver and a biface on June 22, 1971. Above the brown clay loam are three gravel units with a total thickness of 4.4 m. The lower and upper two units form resistant benches of horizontally-bedded sandstone or quartzire pebbles in a matrix of silica and iron hydroxide. The middle unit displays large-scale trough crossbeds. Foreset dips of 10-20 show variable flow directions between 120 and 260, consistent with the 250 palaeocurrent direction inferred from imbrication. The flow of the modern Belan at this site is to the southwest. From the lower gravel unit have come Middle Palaeolithic quartzite tools and bifaces (B.B. Misra: personal communication, 14.1.82). The assemblage seems broadly comparable to that excavated from the Patpara Formation ("FormationB", Table 1) in the Son valley during 1980. From the middle cross-bedded unit have come over 200 artefacts, mostly of Middle Palaeolithic aspect, with fewer fashioned from quartzite and far more of chert. Artefacts in the upper gravel unit were entirely of Middle Palaeolithic aspect, with abundant chert scrapers and no bifaces. What time gaps are represented by the erosional contacts between the three gravel units we cannot gauge. Above the gravels is a massive unit of brown (7.5 YR 4/4-4/6) calcareous very fine sandy clay to clay loam similar in colour and texture to the Mahagara loessic unit in section 4. The basal 5.75 m of the clay loam is capped by a 1.3 m thick bench of rolled carbonate gravel with minor clasts of shale and sandstone and occasional recrystallised mussel shells. Above this intraformational gravel the brown clay loam continues for a further 8.70 m, bringing the total thickness of the loess formation to 15.75 m. Eroding from the upper few metres of the loess is a non-geometric microlithic assem- blage, in which the dominant raw ~ a t e r i a l is chalcedony with subsidiary black chert. From section 4 and 5 (Fig. 4) it is evident that deposition of the calcareous brown loess was accompanied by carbonate segregation within the profile, and was inter- rupted by minor phases of erosion and reworking of pedogenic carbonate derived from the loess. The resulting intraformational gravels appear to be local phenomena and so shouid not be given the status of regional lithostratigraphic markers. There is no valid reason why any of the six gravel units exposed in (and below) section 5 should be referred to as Gravel I, II or III, unless the numerical suffix refers specifically to the local stratigraphic section. Section 6 and 7 are respectively 200 m upstream and 500 m downstream of the Deoghat bridge, with 6 on the right bank and 7 on the left bank (Fig. 4). At the base of section 6 there are at least two metres of rolled sandstone and carbonate gravels which are capped by three metres of brown (7.5 YR 4/6 and 5/6) clay loam and a further 2.25 m of sandstone pebbles, fining upwards to fine gravel. Stratigraphically above the gravel (which may be a local channel-fill deposit) and especially well exposed 200 m above the bridge, is a brown loess-like very fine sandy clay loam, the vertical face of which is distinctively fluted by rainwash rills. Within the deposit, which is over 12 m thick, is an intraformational lens of fine carbonate gravel stringers (Fig. 4). . Section 7 comprises two distinct fine-grained formations separated by a wide- spread shell-bearing gravel unit 85 cm thick dated at 18,055 150 B.P. (Beta-4789). 296 M.A..J. WILLIAMS AND M.F. CLARKE The lower formation, a reddish-brown (5 YR 4/8) sandy loam with 3% black pisolites, is largely obscured by talus, and may overlie the conglomerate bench which crops out a water level. A thin band of ironstone pisolites, .sandstone fragments and bits of broken shell in a reddish-brown sandy clay matrix htrms a minor gravel unit 25-30 cm thick near the top of this formation, which here comprises a further 65 cm of dark reddish-brown (5 YR 3/5) sandy clay. The shelly gravel unit sitting on the eroded surface of the red-brown sandy clay varies lithologically over quite short horizontal distances. At the site of section 7, the gravel consists of ironstone pisolites and ferruginous sandstone fragments in a brown clay loam matrix. Fifty metres upstream it becomes an indurated bed of rolled carbon- ate nodules with crude planar bedding. The most common shells within the gravel are two species of freshwater mussel similar to th,ose found in the Belan today. These shells were sampled at section 7, as well as at 7B (in gully 150-200 m upstream and inland from 7), and 7C near Amilia, 2-3 kIn downstream. In all three places the shelly gravel had an abrupt erosional contact with the underlying red-brown clay. Above the shelly gravel bed at section 7 there are 11 m of alternating dark and light brown (7.5 YR 4/3-4/4) calcareous light clays, often with a strong subangular blocky structure. The uniformity of colour, texture (grain size) and structure in the upper 11 m, the lack of shells, and the absence of fluviatile gravels again suggest that the parent material of the stratified soils and sediments above the shelly gravel was a calcareous wind-blown dust or loess. Since it is now technically possible to obtain fairly reliable thermoluminescence dates from loess and other aeolian sediments (Wintle 1981; Wintle and Huntley 1982; Singhvi et al. 1982) we have sampled every Belan section with this possibility in mind. We hope to be able to test the accuracy of the TL dates against the radiocarbon ages obtained on unrecrystallised shell samples. Both TL and additional 14(: dating programmes are in progress now. Shells were also collected from gully section 7D near Koldihwa archaeological site on the high terrace opposite Mahagara (section 4). The shells were from a carbonate gravel at least 5 m higher in elevation than the 10 kyr B.P. gravel of section 4. They were thought to be laterally equivalent to a gravel band at the base of the Koldihwa excavation which has yielded Upper Palaeolithic blades and a 14(: age of either c.25 kyr B.P. or 19 kyr B.P. (Sharma et al. 1980: 71; B.B. Misra: personal communication, 15.1.82). The actual 14C age for the shells was 25,430 350 B.P. (Beta-4877). Our final investigations in the Belan valley were focused upon four sections on the south bank of a cut-off channel of the Belan opposite Chopani-Mando archaeological site (Fig. 5, sections 8A,B,C and 9). We also revisited the Belan Main section (Fig. 4, in Williams and Royce 1982) on the Belan right bank just downstream of the bifurcation of the Belan and its cutoff channel. The latter is functional during flood, and re-enters the main Belan just upstream of section 4 at Mahagara. At section 8A the cutoff channel is a bedrock channel cut in horizontal quartzose sandstones of Upper Vindhyan age (=Upper Proterozoic). At 3.2 m above the exposed bedrock, and probably sitting directly on bedrock which is here obscured by talus, is a resistant conglomerate bench of flaggy sandstone and gravelly sand which has yielded Middle Palaeolithic quartzite flakes, scrapers and discoid cores. Roughly a hundred metres upstream, at section 8B, bedrock crops out from the cutoff-channel water level to +6:5 m. Above bedrock is a 2.65 m thick unit of pale yellow-brown (10 YR 6/4) clay loam overlain by 3.8 m of dark brown (10 YR 3/4) light clay from which Upper Palaeolithic chert blades have been collected (B.B. Misra: THE QUATERNARY OF CENTRAL INDIA 297 personal communication, 16.1.82). The lowest stratigraphic unit in section 8C, 80 m further east, is laterally equiva- lent to the lowest unit in 8B, and overlies a thin lag gravel said to contain Middle Palaeolithic quartzite flakes. Its upper surface is the Bca horizon of a now truncated palaeosol which developed within the yellow-brown clay loam during a pause in deposition. Above the eroded palaeosol there is a metre of dark brown light clay equivalent to unit 2 of section 8B. It too has yielded Upper Palaeolithic and early Mesolithic artefacts. Capping the brown clay is a brown (10 YR 4/4) fine sandy clay loam at least 3 m thick, the surface of which is being actively eroded. This unit contains microliths but no pottery. Perhaps the most complete of the Chopani-Mando sections is step trench 9 (Fig. 5). Above a 2 m thick conglomerate bench of platy sandstone and nodular calcrete is a yellow-brown (10 YR 4.5/4) calcareous clay loam overlain by a brown (10 YR 4/4) clay loam from which have come Middle Palaeolithic quartzite and chert blades. Above the brown clay loam is a brown (7.5 YR 4/3) fine sandy clay unit some 3 m thick with Upper Palaeolithic blades and microliths. Capping the section is an eroded brown (10 YR 4/4) clay laom unit at least 0.8 m thick, also with microliths. One feature common to all of the Belan sections studied by us is a brown to yellow-brown clay loam to fine sandy clay formation which predates the 10 kyr B.P. shelly gravel at section 4 and appears to be a slightly reworked Late Pleistocene loess, the bulk of which accumulated during the Last Glacial Maximum, on the evidence of our shell dates and of the two shell dates of 23,840 B.C. (PRL 86) and 17,765 B.C. (TF 1245) available from the shelly gravels which immediately underlie this formation at Deoghat and Koldihwa (Sharma et ai. 1980: 71; B.B. Misra: personal communica- tion, 15.1.82). Our present interpretation of this formation differs slightly from that of Mujumdar and Rajaguru (1970) who thought that "the possibility of these finer soils being deposited solely by wind was very meagre" (Mujumdar and Rajaguru 1970: 102) and considered them to be overbank deposits or fluvio-Iacustrine deposits associated with a floodplain environment. The operative word is "solely". We consider the parent material to be loess, reworked to a minor degree by slopewash and local runoff. Some of the more highly stratified beds may well be overbank deposits laid down by a loess- choked Late Pleistocene palaeo-Belan. Until we have completed our dating, micromorphological and granulometric studies, further speculation seems unprofitable. Table 3: Provisional lithostratigraphic sequence in the middle Belan valley, based on fieldwork in January 1982 and February 1980 I. Tabular sandstone conglomerate over Upper Vindhyan sandstone bedrock. Lower Palaeolithic cleavers and bifaces. Age probably Middle Pleistocene. 2. Calcareous brown clay loam. Archaeologically sterile. (?)Middle Pleistocene. 3. Planar and cross-bedded sandstone gravels. Middle Palaeolithic artefacts. (?)Late Middle to (?)early Upper Pleistocene. 4. Reddish-brown sandy clays and clay loams. (?)Upper Pleistocene. Archaeological associations not known. 5. Shell-bearing gravel(s), mostly rolled carbonate nodules and black ironstone pisolites. Upper Palaeolithic artefacts. Upper Pleistocene (c.25 kyr to 19 kyr B.P.). 6. Brown and yellow-brown calcareous clay loam and sandy clays. Minor intraformational gravels, mainly rolled carbonate nodules. Upper Palaeolithic to Epi-Palaeolithic and (?)Mesolithic artefacts. Age late Upper Pleistocene (c.25 to c.1O kyr B.P.) 7. Holocene clays, loams and fine alluvial sands. Mesolithic, Neolithic and protohistoric artefacts. 14* 298 o 1'-30 - .. J .eo cI -120 M.A.J. WILLIAMS AND M.F. CLARKE - 1 . ~ ~ ~ - - ~ - - ~ - - ~ - - ~ - - ~ - - ~ ~ ~ ~ ~ - - ~ - - ~ - - ~ ~ o 20 40 10 100 120 140 1. 110 200 220 240 210 Age(1P) Fig. 6: Sea level curve for the past 260,000 years (after Aharon and Chappell 1986: Fig. 4.5) The red-brown clay loam to fine sandy clay exposed in the Belan Main section and in sections 7 and 7b may well be a weathered loess. In colour and soil texture it is similar to the finer units of the Patpara Fonnation (Fonnation B) in the Son valley. The results from our dating programme may resolve this issue which for the present is best left open. Major stratigraphic units and their associated archaeological occurrences are summarised in Table 3. If units 2, 4 and 6 were originally laid down as wind-blown dust, then they would presumably reflect times when the prevailing climate was drier and windier than today. Duplessy's (1982) reconstruction of the last glacial climate in India is consistent with our suggested later Pleistocene age and loessic origin for unit 6 in Table 3. Given that times of low seal level were times of maximum ice volume, and accepting the equation of loess deposition with glacial (or stadial) aridity, the curve of sea level changes over the past 260 kyr (Fig. 6) suggests possible ages for units 2 and 4, assuming they are of aeolian provenance. Unit 2 could date to roughly 160 10 kyr B.P. and Unit 4 could date to 70 kyr B.P. or earlier. However, the shape of the curve indicates a progressive trend towards full glacial conditions, and there seems no good reason why wind-blown dust could not have accumulated in the Belan valley throughout much of the long interval between successive interglacials. The present climate in north central India is probably only representative of, at most, 10% of the preceding 120 kyr, so that throughout much of the Middle and Upper Palaeolithic conditions may have been significantly colder, drier and windier than they are today. (c) Alluvial Stratigraphy of the Middle Son Valley Between Byawaharkhand Village and Jogadha Bridge Our first priority in the Son valley was to relate the archaeological excavations at Baghor I,ll and III to the local and regional geological history (see Fig. 7 and 8). Section G 1 is 1 krn west of Baghor I and G2 is in the same Nala as G I, but about 350 m upstream. The main unit in both sections is a yellow-brown loess, the field texture of which is a calcareous very fine sandy clay to heavy clay loam. Colours range from 10 YR 5/6 near the exposed base of G 1 (which has up to 20% red-brown mottles) through 10 YR 5/4 to 10 YR 4/4 (brown) near the top of G2, i.e. from yellow brown to brown. All colours are moist colours and particular care was taken in describing m 9 8 7 6 5 4 3 2 1 o 2 Yb 1 Yb G1 SON VALLEY : GEOLOGICAL SECTIONS A. GF in the, vicinity of Baghor excavations G2 4 b 3 gb 2 Yb 1C 1B 1A Yb .=11 2 Rb 1 b G3 Fig. 7: Quaternary stratigraphic sections, Son valley near Bagbor ... . ~ I:) ~ i!>- ~ ... l"!I 5 Rb ~ A . ~ " ' 1 4 i!>- 3 qp ill:! >< 0 b "'l 2 gb n l"!I Z ... 1C ~ t"' 1B Rb 2 ~ i:1 '-0 - - o 0 i!>- - 1A '':'0 - - : 0 0 GF:....::.-:::i G4 ~ 300 M.A.J. WILLIAMS AND M.F. CLARKE field colour and texture of all deposits, since even quite subtle differences proved. diagnostic. Within the loess there are minor intraformational lenses of shale rubble (mostly debris flows) and rolled carbonate gravels. Conformably above the loess at G2 there is a vertisolic dark grey-brown (l0 YR 3/2) sandy clay. Above the dark clay is a thin band of shale gravel with agate and chalcedony flakes and microliths. The microlithic horizon is partly concealed beneath a layer of red-brown sandy clay loam (see also section G3, 0.5 km east of G2) which sometimes contains Neolithic implements. Exposed beneath the yellow-brown loess at G2 there are three gravel units. These range from shale fragments in 'a clay matrix (lA) to quartzite conglomerates with crude fluviatile bedding (lB, 1C). In the bed of the seasonal stream which is exhuming these gravels we found one large quartzite core (c.25 x 25 x 20 cm) and one large quartzite flake, possibly Lower Palaeolithic. Fanglomerates also crop out 0.1-0.2 km northwest of Baghor I, at section G4 (Fig. 7). The site is closer to the shale quartzite escarpment, and the gravels differ from those at G2 in being matrix-supported. Units 1A and 1C have a characteristic debris- flow fabric. Interstratified within the debris flows there are several beds and lenses of red-brown (5 YR 5/6) sandy clay and sandy clay loam, often with a strongly pedal structure quite unlike the massive structure of the yellow-brown loess at Gland G2. Erosion of this material, which is reminiscent of the Patpara (or Formation B) clays studied in 1980, has provided the matrix material of the debris flows. We consider that these scarp-foot clays represent an eroded and weathered loess which predates the younger yellow-brown loess which is so widespread in this region. We sampled and described the three archaeological test trenches at Baghor I, II and III (Figs. 8 and 9), and an additional geological section (G5) as a further check upon 'Baghor III. It soon became clear that the'mottled grey-brown sandy clay/clay loam unit at the base of Baghor I and II and comprising almost all of Baghor III was stratigraphically equivalent to the upper metre or two of the yellow-brown loess at G2. The dark vertisolic clay (bed 3) at Baghor II is a lateral counterpart of the dark vertisolic clay at G2 (bed 3). The red-brown ("orange" when dry) sandy clay loam at Baghor II (bed, 2) is the same horizon as the top bed at G3 and G4 (Fig, 7). In descending order, the stratigraphic sequence is as follows: (i) red-brown ("orange") sandy clay loam with Neolithic artefacts (ii) brown clay loam (iii) grey-brown sandy clay with Mesolithic artefacts (iv) debris flows (shale rubble) (v) very dark grey-brown vertisolic clay (vi) grey-brown mottled clay loam/sandy clay with Upper Palaeolithic artefacts grading downwards into (vii) brown loess (clay loam/sandy clay) greyer upwards, yellow-brown down profile; minor intraformational shale and carbonate lenses (viii) interbedded red-brown clays (weathered older loess?) and sandstone/shale debris flows in red-brown clay matrix (ix) water-laid sandstone cobbles and gravels: large core and flake, possibly Upper Palaeolithic, (Note that these gravels could be a lateral equivalent of viii, but we think this is unlikely given the absence of interstitial red-brown clay) Our subsequent work in the Son was designed to answer two important questions which had now arisen:
- - BAGHORlTE8TTRENCH IOUIhflcl ......................... . 2 m , , ' . '.' 2 ..... --.S / ... ... 1 gracIIng downward, ID griIly brown qndy clay wIItI ....w F. & Mn noduln lA Urk grey-brown undy clay .... 1()1j6 F. & Mn nodule, S foooooooooooooooooooooooooopoooo 0 I , o 1m
4 lrIfact ight clay-loam
fine .Ie 811M! i'I M NnUpOrt bind of clay ("",.,,1) S ,."ple for .n.lyail 10001
1A Ih.Irp bound8ry . grldetionll' bouncSwy .. ion.1 bound8ty Fig. 8: Baghor I and n test trench sections, Son valley IS 5 undy dllyWllh IOIIhIc III'II1CtI & NndItone INIrtUpOItI .. "' ............ _ ... __ ............................................. .. _ .. 9 ... _ .. _ ....... _ ....... .. .. ,NIle gnMtI i'I dirk grey-brown cIIIy matrtx . ... .............. . ''" _ ........................................ i 10 a
.. 1IIght cIIIy-lOlm
.. o 1m 2m
> BAGHORIITESTTRENCH ......
... SON VALLEY : GEOLOGICAL SECTIONS m 9 8 7 8 5 4 3 2 1 o in the vicinity of Baghor eXC8YationI 6 ':11 3 ' ~ 4 Ab -- 6 I- - gb+b -- 3 gb -- 2 I:. 2 - ~ l' gb -- 1 b+Rb -- -- 100m BAGHORIII OS Fig. 9: Quaternary stratigraphic sections, Son valley, Khunderi Nala Khunderi Hala right bank e b 7B b m 7A b 4A b b 4B =
3 .. Vb 2 -" 1 415 Vb I 3 II 2Vb 1 W. G8 G7 ~ == > ~ ~ t"' t"' ... > ; > 2: t:I == !oJ ~ t"' > '" ~ ttl THE QUATERNARY OF CJj:NTRAL INDIA 303 (i) What is the relationship- between the piedmont stratigraphic sequence evident around the Baghor excavations and the alluvial stratigraphic sequence exposed in the banks and tributary nalas of the Son? (ii) How many loess fonnations are there, and what is their relationship to artefact-bearing units? The answers to these two related questions are illustrated in section G6 to G 12 (see Figs. 9-11). Above the basal gravel units at G6 and G7 (Fig. 9) there are recent alluvial sands and laminated clays which belong to the youngest terraces of Khunderi Nala. They are banked up against the main yellow-brown loess formation which contained shells with a 14C age of 20,135 220 B.P. (Beta-4791). On the south bank of the Son at Nakjhar Khurd we sampled a step-trench located 20 m downstream (east) of the 1980 archaeological trench (G9 in Fig. 10). Above the eroded siltstone bedrock were exposed 4.5 m of mottled grey and yellow-brown sandy clay with a thin interbedded debris flow unit (bed 3 of G9). Beds 1 to 5 we interpret as a slightly reworked loess mixed with some fluviatile coarse sand and bedrock derived shale fragments. A sample from near the top of bed 4 has yielded a thermolu- minescence age of 103,800 19,800 (Alpha-899). This loess is the Fine Member of the Sihawal Formation (Formation A) and in the 1980 step-trench it overlies a discon- tinuous basal conglomerate - the Sihawal Gravel Member - which contains Lower Palaeolithic bifaces . This gravel may be a lateral equivalent of the indurated gravels which overlie limestone bedrock in Khunderi Nala north of the Son at sections G6, G7 and G 1 0 (Fig. 9 & 10). In all three sections the gravel is the lowest exposed unit, and underlies a mottled grey and yellow-brown fine sandy clay. Resting on the eroded surface of the Sihawal Loess Member at Nakjhar Khurd are two metres of horizontally-bedded red-brown (7.5 YR 5/6) and brown (10 YR 4/6) coarse sandy clay, the upper 0.55 m of which has a coarse prismatic structure, and may be part of an eroded palaeosol. Above the "palaeo sol" (unit 6D, Fig. 10 section G9) is a fluviatile gravelly coarse sand, the colour of which ranges from brown (10 YR 4/6) to dark red-brown (5 YR 3/3-4). This unit (7) is the base of a channel fill, so that formation and member thickness can differ by several metres over a lateral distance of 20 m. A further 11.75 m of gravelly sands overlie unit 7, so that the level summit of the section is +25 m relative to its ~ x p o s e d base, which is 1 m above low river level. The stratigraphic status of the red-brown gravelly sands above unit 6D is still in doubt. They could belong in the upper, eroded part of the Patpara Formation or in the subsequent Baghor Formation. They differ from the Baghor sands in colour and in the absence of carbonate, and may represent a stratigraphic transition from the red-brown gravelly clays and clays of the main Patpara Formation to the unconsolidated pale yellow-brown fluviatile sands and calcreted gravels of the Baghor Coarse Member. Alternatively, the Patpara Formation may consist of a clay-rich facies and channel- sand facies, with the latter more common along the axis of the valley, and the former more extensive on footslopes and palaeo-floodplains. In order to correlate the stratigraphic sequence on either side of the Son we ran a traverse from near Rampur archaeological site across the river to Semara Nala on the south bank (Fig. 11). The oldest formation exposed on both sides of the river in the form of eroded remnants of red-brown sands and clayey sands is the Patpara Formation, here present in its coarse fluviatile facies. SON VALLEY : GEOLOGICAL IECTIONI 1CIutdIri .....
11 8gb 10 r 1b1M1trl a.pr c:.mpdI , 5Yb " b f 7 8r-- 8 Nt 5 4 t --j' ...
:: .. Ybtt 3 2 Ybtb 1 H 3Nttg OF l:..!...j 1 0 [ OF :::": 1.2Ob 010 011 Fig, 10: Quaternary stratigraphic sections, Son valley, Nakjhar Khurd 1km-*rI .....1Oud. .: .. '., : :: .. ::. ::.:. .:' .. :':1 ::":::: 012 .Vb 5 b .. b 3 b+oIg 2 b 1
>
>
p >
= ./ ::/ : .. a Q ... , ; " ~ . .. II e" ./ a :j I I : ... 1 ~ i ... .. i ~ .... o I i , EIII a THE QUATERNARY OF CENTRAL INDIA o E I iii 101-1' ... 1 ..... ' ..... ...... ~ A........ .. i 1111 I ..... 1.1 ,,1 I .... .. ' .. II Ii G.:S: I:' . ')? A'::" c':.:: ".,: 305 o - - 306 M.A.J. WILLIAMS AND M.F. CLARKE Overlying the eroded surface of the Pat para Formation are the channel sands of the Baghor Formation Coarse Member. These are interstratified with silts and clays on the north bank, as might be expected in the laterally more distal facies of channel sands. On the south bank the Baghor sands are clean, well-sorted sands with no interfingering units of overbank clays except near the contact with the overlying Fine Member clays. Both the Fine and the Coarse Members grade laterally into the massive brown to yellow-brown loess described earlier in this report. Three dates allow us to place the Baghor Formation Fine and Coarse Members and associated loess firmly in the Late Pleistocene. A 1"C age of 11,870 120 B.P. (Beta- 4793), is in very good agreement with the 1L date for the stratigraphically equivalent unit 50 m further north (Fig. 11). There are sporadic relatively fresh Middle Palaeolithic flakes near the base of the Baghor sands, and Upper Palaeolithic artefacts near the top. The Baghor clays which overlie and/or interdigitate with the Baghor sands also contain Upper Palaeolithic blades and blade-cores. Banked against the Baghor Formation (at Khunjun) or else sitting on the 15 m erosional terrace at the foot of the main 35 m surface (at Rampur) is a brown clay in which there are very well preserved Upper Palaeolithic blades and blade-cores. Using the information discussed above (see Fig. 7-11) we are now in a position to summarise what is still preserved of the Quaternary history of the middle Son valley. (i) Bedrock erosion and pedirnentation of Lower Vindhyan metasediments (limestones, shales, cherts, sandstones), probably during the Middle Pleistocene. Deposition of debris flow rubble and alluvial fan gravels and clayey gravels with some reworking by the palaeo-Son near present low r i v e r ~ l e v e l . Lower Palaeolithic flakes, cores and bifaces of quartzose sandstone and quartzites on and in these Sihawal Formation gravels. (ii) Deposition of fine sandy clay loess during and after accumulation of Sihawal Formation gravels; i.e., a Sihawal Fine Member and a Sihawal Coarse Member, (?)Middle Pleistocene. (iii) Erosion followed by deposition of Patpara Formation gravelly clays and fluviatile sands. Upper Acheulian to Middle Palaeolithic artefacts interstratified among both fine and coarse facies of the Patpar Forma- tion. Syn-depositional and/or post-depositional reddening of clays and sands. Middle to Upper Pleis- . tocene. (iv) Erosion followed by deposition of Baghor Formation channel sands, overbank clays and yellow-brown loess. Middle to Upper Palaeolithic. Upper Pleistocene. (v) Several intervals of episodic downcutting, lateral planation and inset terrace formation. Upper Palaeolitbic, Mesolithic and Neolithic. Upper Pleistocene to Holocene. The most important geological achievement of the 1982 season was the identifi- cation, mapping and sampling of a very widespread Upper Pleistocene loess in the Belan and Son valleys. Shell samples collected above, within and beneath this forma- tion have allowed us to date it with a reasonable degree of accuracy as well as enabling us to calibrate some of the thermoluminescence dates so far obtained on samples of this loess. We are thus able to specify limiting ages for the younger Middle Palaeolithic and Upper Palaeolithic occurrences in this part of India. If the as yet unfinished 1L dating programme proves successful, we will also be able to obtain a better idea of the timing of the transition from Lower to Middle Palaeolithic by further dating the Sihawal Fine Member loess. Ultimately, we would hope to be able to correlate the continental loess stratigraphy of north central India with the Quaternary marine stratigraphy obtained from Indian Ocean deep-sea cores as well as with Chinese loess stratigraphy. These are exciting possibilities. THE QUA TERNARY OF CENTRAL INDIA 307 Acknowledgements We remain indebted to Professor J. Desmond Clark and the late Professor G.R. Sharma for inviting us to participate in their joint archaeological investigations, as also to our cotIeagues from Allahabad and Berkeley. Our grateful thanks go also to the Smithsonian Institution for air fares and per diem funds, to Allahabad University for arranging transport and accomodation in the field, and to Macquarie University for special leave, and research funds for the XRD, 14C and sedimentological analyses. We offer this essay as a tribute to our friend, colleague and erstwhile field com- panion, Dr.S.N. Rajaguru. More than any other single person, he has worked to place Indian Quaternary geology on a modern footing through his careful and unflagging field research throughout the Subcontinent. M.AJ. Williams will long remember the joy and privilege of working with Raja in the Rajasthan desert in 1983. References Acharyya, S.K. and P.K. Basu 1993. Toba Ash on the Indian Subcontinent and Its Implications for Correlation of Late Pleistocene Alluvium, Quaternary Research 40: 10-1 9. Acharyya, S.K. and P.K. Basu 1994. Reply to Comments by S. Mishra and S.N. Rajaguru and by G.L. Badam and S.N. Rajaguru on "Toba Ash on the Indian Subcontinent and Its Implications for Correlation of Late Pleistocene Alluvium", Quaternary Research 41: 400-402. Aharon, P. and J. Chappell 1986. Oxygen Isotopes, Sea Level Changes and the Temperature History of a Coral Reef Environment in New Guinea Over the Last I O ~ Years, Palaeogeography Palaeoclimatology Palaeoecology 56: 337-79. Allchin, B., A. Goudie and K.T.M. Hegde 1978. The Prehistory and Palaeogeography of the Great Indian Desert. New York: Academic Press. Dadam, G.L. and S.N. Rajaguru 1994. Comment on "Toba Ash on the Indian Subcontinent and Its Implica- tions for Correlation of Late Pleistocene Alluvium", Quaternary Research 40: 398-399. Chesner, e.A., W.1. Rose, A. Deino, R. Drake and J.A. Westgate 1991. Eruptive History of Earth's Largest Quaternary Caldera (Toba Indonesia) Clarified, Geology 19: 200-203. Clark J.D. and M.AJ. Williams 1986. Palaeoenvironments and Prehistory in North Central India: a Prelimi- nary Report, in Studies in the Archaeology of India and Pakistan (1. Jacobson Ed.), pp. 19-41. New Delhi: Oxford-IBH. Clark, J.D. and M.AJ. Williams 1990. Prehistoric Ecology, Resource Strategies and Culture Chang! in the Son Valley, Northern Madhya Pradesh, Central India, Man and Environment 15(1): 13-24. Cullen, J.L. 1981. Micro-Fossil Evidence of Changing Salinity Patterns in the Bay of Bengal Over the Last 20,000 Years, Palaeogeography Palaeoclimatology Palaeoecology 35: 315-356. Duplessy, J.e. 1982. Glacial to Interglacial Contrasts in the Northern Indian Ocean, Nature 295: 494-498. Goudie, A.S., B. Allchin and K.T.M. Hegde 1973. The Former Extensions of the Great Indian Sand Desert, Geographical Journal 134: 243-257. Jacobson, J. 1975. Early Stone Age Sites in Eastern Malwa, Proceedings of the American Philosophical Society 119: 280-297. Jouzel, J., e. Lorius, J.R. Petit, e. Genthon, N.I. Barkov, V.M. KotJyacov and V.M. Petrov 1987. Vostok Ice Core: a Continuous Isotope Temperature Record Over the Last Climatic Cycle (160, 000 years), Nature 329: 403-408. Kenoyer, J.M., J.D. Clark, J.N. Pal and G.R. Sharma 1983. An Upper Palaeolithic Shrine in India, Antiquity 57: 88-94. Lal, B.B. 1981. Two Indian Epics Vis-A-Vis Archaeology, Antiquity 55: 27-34. Lal, B.B. and K.N. Dikshit 1978-79. Sringaverapura: a Key-Site for the Protohistory and Early History of the Central Ganga Valley, Puratattva 10: 1-7. Mishra, S. and S.N. Rajaguru 1994. Comment on "Toba Ash on the Indian Subcontinent and Its Implication for Correlation of Late Pleistocene Alluvium", Quaternary Research 40: 396-397. Mujumdar, G.G. and S.N. Rajaguru 1970. Investigations of the Pleistocene Sediments from the Belan Valley, U.P., Indian Antiquary 4: 96-105. 308 M.A.J. WILLIAMS AND M.F. CLARKE Ninkovich, D., N.J. Shackleton, A.A. Abdel-Monem, J.D. Obradovich and G. lzett 1978. K-Ar Age of the Late Pleistocene Eruption of Toba, North Sumatra, Nature 276: 574-577. Paddayya, K. 1977. New Research Designs an51 Field Techniques in the Palaeolithic Archaeology of India, World Archaeology 10: 94-109. . Prell, W.L., W.H. Hutson, D.F. Williams, A.W.H. Be, K. Geitzenauer and B. Molfino 1980. Surface Circu- lation of the Indian Ocean During the Last Glacial Maximum, Approximately 18 000 Yr. B.P., Quaternary Research 14: 309-336. Rampino, M.R. and S. Self 1992. Volcanic Winter and Accelemted Glaciation Following the Toba Super- Eruption, Nature 59: 50-52. ' Rampino, M.R. and S. Self 1993. Climate-Volcanism Feed Back and the Toba Eruption of -74000 Years Ago, Quaternary Research 40: 269-280. Rose, W.l. and C.A. Chesner 1987. Dispersal of Ash in the Great Toba Eruption, 75 Ka, Geology 15: 913- 917. Sharma G.R. 1975. Seasonal Migrations and Mesolithic Lake Cultures of the Ganga Valley, in K.C. Chattopadhyaya Memorial Volume (G.R. Sharma Ed.), pp. 1-20. Allahabad: Allahabad University, Sharma G.R. and J.D. Clark 1983, Palaeoenvironments and Prehistory in the Middle Son Valley. Allahabad: Abinash Prakashan, Sharma G.R., V.D, Misra, D. Mandal, B.B. Misra and J,N. Pal 1980. Beginnings of Agriculture. Allahabad: Abinash Prakashan. Singh, G. 1971. The Indus Valley Culture Seen in the Context of Post-Glacial Climatic and Ecological Studies in North-West India, Archaeology and Physical Anthropology in Oceania 6: 177-189. Singh G. and D.P. Agrawal 1976. Radiocarbon Evidence for Deglaciation in North-Western Himalaya, India, Nature 260: 232. Singh G., R.D. Joshi and A.B. Singh 1972. Stratigraphic and Radiocarbon Evidence for the Age and Devel- opment of Three Salt Lake Deposits in Rajasthan, India, Quaternary Research 2: 496-505. Singh, G., R.D. Joshi, S.K. Chopra and A.B. Singh 1974. Late Quaternary History of Vegetation and Climate of the Rajasthan Desert, India, Philosophical Transactions of the Royal Society of London B267: 467-501. Singhvi, A.K., Y.P. Sharma and D.P. Agrawal 1982. Thermoluminescence Dating of Sand Dunes in Rajasthan, India, Nature 295: 313-315. Whitford, 0 J. 1975. Stronium Isotopic Studies of the Volcanic Rocks ofthe Sunda Arc, Indonesia, and their Pevogenetic Implications, Geochimica et Cosmochimica Acta 39: 1287;p02. Williams, M.AJ. 1985. Pleistocene Aridity in Tropical Africa, Australia and India, in Environmental Change and Tropical Geomorphology (I. Douglas and T. Spencer Eds.), pp, 219-233. London: George Allen and Unwin. Williams, M.AJ. and M.F. Clarke 1984, Late Quaternary Environments in North Central India, Nature 308: 633-635, Williams, M.AJ. and K. Royce 1982. Quaternary Geology of the Middle Son Valley, North Central India: Implications for Prehistoric Archaeology, Paleogeography Palaeoclimatology Palaeocology 38: 139-162. Williams, M.AJ. and K. Royce 1983. Alluvial History of the Middle Son Valley, North Central India, in Palaeoenvironments and Prehistory in the Middle Son Valley (G.R. Sharma and J.D. Clark Eds.), pp. 9-23. Allahabad: Abinash Prakashan. Williams, M.AJ., D.L, Dunkerley, P. De Deckker, A.P. Kershaw and T. Stokes 1993. Quaternary Environ- ments. London: Edward Arnold. t< Wintle, A.G. 1981. Thermoluminescence Dating of Lji'te Devensian Loess in Southern England, Nature 289: 479-480. . Wintle, A.G. and OJ. Huntley 1982. Thermoluminescence Dating of Sediments, Quaternary Science Reviews I: 31-53,