Studies in the Geomorphology, Quaternary Palaeoenvironments and Archaeology of the Vel River, a Tributary of the Bhima in Western Maharashtra Sheila Mishra, Sonali Naik, Utpala Adhav, Sushama G. Deo and S.N. Rajaguru Department of Archaeology Deccan College Pune411 006
Abstract The Vel river is a tributary of the Bhima with most of its catchment in the semi-arid zone in Upland Western Maharashtra. It has preserved geomorphic and archaeological features throwing light on tectonic, climatic and archaeological history since the Neogene. In this paper we argue that denudational surfaces, high level gravels (HLG), weathered bedrock and the bedrock incised channel seen in the Vel basin around Talegaon Dhamdhere are relict features of the landscape. A series of aggradational episodes have been dated to around 10 kyr, 7 kyr, 2.5 kyr and 200 yr bp. The episodic aggradational activity of the Vel dates primarily to the Holocene while the trunk stream, the Bhima, was in an incisional mode during the same time period. We suggest here that this different behaviour during the same time is caused by different basin characteristics of the two rivers. Archaeological material is associated with all the aggradational episodes. Microlithic industries are associated with both the 10 and 7 kyr gravels, while Late Chalcolithic pottery is found at 2.5 kyr and a Maratha period temple at 200 yr bp.
Introduction Studies of the Vel have been carried out over the last three years. Repeated visits have been rewarded with many new insights, many of them applicable to the region as a whole.The Vel river, with a length of 64 km has its source about 15 km upstream of Peth, on the 700 m denudational surface. This surface forms the divide between the Bhima and Ghod rivers. The Vel River therefore lacks a catchment in the high rainfall Western Ghats zone, unlike the Bhima, Mula-Mutha and Indrayani (Fig. 1). Talegaon Dhamdhere is on the Vel, about 8 km from its confluence with the Bhima. Some of the features of the Vel river are best explained by tectonics while others reflect the effects of cIimatic change. The present-day landscape has been formed over the last 60 myr years after the eruption of Deccan Trap near the Cretaceous Tertiary boundary .Tectonic forces play a major role in forming the landscape over long time scales while climatic change leaves its imprint over shorter time scales. It is perhaps for this reason that geologists, for whom the Quaternary is a very short time period, stress the tectonic elements in the landscape, while archaeologists, for whom the Quaternary is a long time period stress climatic change as an important element. One group of features discussed, belongs to the Tertiary, and while relict, being formed under a climate different from the present,
are none the less, primarily formed by tectonics. The second group of features all belong to the Holocene, and although exact palaeoclimatic interpretations are not possible with the present data, they reflect the frequent climatic changes of the Late Quaternary. Relict Peatures One of the major findings of our work over the last decade is that many features of the present-day landscape are relict. By relict we mean not only that the features are old but also that they have been formed by processes different from those operating at present. This realization is based primarily on a single observation which has been discussed earlier (Mishra 1995; Mishra et nl. 1993). This observation is that often gravels are found which overlie weathered bedrock or "murrum". Some of these gravels are at a "high level" and so they were called "High Level Gravels" by Foote (1 878). However in some basins, these gravels are not at a high level. Nevertheless, because of the well established use of the term it is retained here. We have taken the most significant feature of the HLGs to be, not their high level, but the presence of "nzurrurn" beneath them. This has important implications. The process of deep weathering of bedrock which produces murrum, requires time and also a humid tropical climate. This weathering must also post-date the gravel deposition as river channels
Man and Emironment X X I V ( I ) - 1999
Fig. 1: The Bhima Basin
erode weathered material and so active channels occupy unweathered bedrock. Gravels with Early Acheulian artefacts at Laxmi Nala and Chirki near Nevasa and at Bori, overlie only slightly weathered bedrock, where the flute marks and smoothing of the bedrock by fluvial erosion is still preserved (Corvinus 1983). Mishra (1 992) and Mishra et al. (1995) argue that these Early Acheulian industries are > 400 kyr and > 670 kyr and so probably of Lower Pleistocene age. This chain of reasoning therefore leds to the conclusion that the HLGs and the weathering that post-dates them belong to the pre-Quaternary period. The denudational surfaces on which the HLGs are found are therefore even older. Denudatiorzal Su faces
The Deccan plateau is characterized by a "stepped" character. Flat surfaces at different elevations are separated from each other by "steps". These flat surfaces have been formed by denudational processes and have been separated from each other by episodic uplift. The Vel River originates on one of the most extensive of such surfaces (Dikshit 1970; Kale and Rajaguru 1988) at approx 700 amsl. In the Vel basin two lower surfaces at 600 amsl and 540 amsl are present. Around Talegaon, the 540 level forms an extensive surface while the divides are around 600 m. The 600 m
level also forms a surface, well developed near Pune. The Vel river therefore traverses three surfaces (700 m, 600 m and 540 m) from its source to confluence. The HLGs are on the lowest surface and therefore all the surfaces pre-date the HLG. High level gravels
High level gravels are exposed in three localities around Talegaon. The first is near the Bhairava temple, where a small tributary joins the Vel (Loc 1, Fig. 2), the second is downstream of Talegaon (Loc 2, Fig. 2) and the third is along the road to Vittalwadi (Loc 3, Fig. 2). At the Bhairava temple 2 m of sandy pebbly gravel is exposed. This is cemented and cross bedding is preserved. It is about 4 m above the Vel bed level. The preservation of the sedimentary structures, and calcrete cement give this gravel a similiar appearance to Quaternary gravels such as those containing Acheulian artefacts at Bori and Nevasa. Most o f the HLGs consist o f well rounded exotic pebbles of resistant lithologies in a matrix of clay and carbonate, with little preservation of the original structures unlike the Vel HLGs. However the bedrock below the Vel HLGs is weathered to a depth of over 1 m unlike the bedrock below the Acheulian gravels. A lag of gravel is found scattered in the fields along the river. At Talegaon, downstream of the
Studies in the Geoirtorphologq: Quaternary Palaeoenvironments and Archaeology of the Vel River
Fig. 2: Localities studied around Talegaon on the Vel River
village in the "bazar" locality, a 4 m thick gravel similar in character to that of the Bhairava temple is exposed on the left bank. For some stretch the gravel is exposed on both sides of the river. The bedrock below this gravel varies from compact, well jointed basalt to amydaloidal basalt. The weathering of the bedrock below the gravel varies with the bedrock lithology, with the jointed basalts more weathered than the amygdaloida! varieties. This gravel is also exposed along an untarred road running parallel to the river on the left side and has been exploited for building material, providing good exposures. The third exposure of gravel is on the right side (east) of the road to Vittalwadi from Talegaon. Here we did not observe any vertical exposure, but there is a lag of cobbly, pebbly gravel overlying bedrock. This gravel is richer in siliceous rocks and there are artefacts on the surface. It occupies the highest position in the locality and so indicates an inversion of relief.
Post high level gravel bedrock weathering
The HLGs identified from Talegaon are associated with the youngest denudational surface. The gravels are fairly coarse, but unlike many HLGs, cobbles and boulders are rare, lithology is local and sedimentary structures, such as grading and cross bedding are preserved. The weathering of bedrock below the gravels is only about 1 m compared with over 5 m for many HLGs in the Godavari Valley. For this reason, it is likely that all the HLGs are not of the same age. The Vel HLGs appear to be younger than others observed by us in the Godavari (Joshi et al. 1980 ) and Tapi valleys (Mishra et al. 1993). Nevertheless, the weathering of bedrock below the gravels and the absence of any fossils or artefacts from these gravels leds us to consider them to be pre-Quaternary in age.
Man and Environnzent XXIV ( I ) - 1999 Fluvial incision of bedrock
Holocene aggradational episodes on the Vel River
Near the Bhairav temple (Loc 1 on Fig. 2) the Vel has a rocky bed with a minor knickpoint. The bedrock has well developed potholes and flutemarks. After this knickpoint the river enters a stretch with a verlical bedrock bank on the left side. After Talegaon village it again enters a bedrock stretch with bedrock1HLG on either side. This bedrock incision shows the ungraded character of the Vel River even near its confluence. This bedrock incision may also be a relict feature related to rejuvenation when the present base level was established.
In our explorations around Talegaon we were able to collect shells from two gravels for dating. In both the cases microliths were associated with the gravels from which the shells were collected. Our expectation, based on previous dates from Western Maharastra was that the gravels would date to the Late Pleistocene when many rivers were aggrading (Kale and Rajaguru 1987; Rajaguru and Kale 1986; Rajaguru et al. 1993; Sadakata et al. 1995). The results of the dating however showed that these gravels belong to the Holocene period (Table 1). The gravels dated were from Loc 3 ("Bazar" of Talegaon) and near Loc 2 (Bhairava temple) and dated to around 10 kyr and 7 kyr. In additional, minor episodes of aggradation during the Midand Late Holocene were associated with Late Chalcolithic pottery and the Maratha period Bhairava temple respectively.
Inlplications of the relict features for the Quaternary The relict features discussed above, form the major landscape elements. The distribution and extent of HLGs is much greater than for any of the Quaternary gravels and the HLGs and weathered bedrock provide a ready supply of material for the Quaternary deposits. The reason for the dominance of the relict elements in the landscape is the ineffectiveness of present day processes to alter the landscape due to the semi-arid climate and tectonically stable landscape. The present day landscape therefore cannot be understood without understanding past processes. While dating the various elements of the landscape is not possible, their relative chronology can be determined. The eruption of Deccan Trap close to the Cretaecous Tertiary boundary places a maximum age boundary, so that all the "events"/features post-date this important event. The denudational surfaces have developed by the continuing erosion of the Deccan Traps and the higher surfaces are older than the lower ones A number of higher/older denudational surfaces are seen in addition to the ones in the Vel basin. The high lcvel gravels are associated with the youngest of these denudational surfaces. The weathering of the gravels and the bedrock below them post-dates the gravel deposition. The bedrock channel incision post-dates the lastest denudational surface but in some cases appears to be prior to the Quaternary alluvium. The dating of the youngest "landscape" element to the Pre-Quaternary means that the other features are earlier.
Pleistocene/Holocene transitional period (10 kyr) Evidence for this time period came from Loc 2 (Bazar). On the right bank of the Vel, just downstream of Talegaon, an 8m section is seen (Fig. 3). This comprises 70 cm silt overlain by a pebbly gravel ol' about 2 m and then a silty sand with some sand lenses within it. This silty sand has a few niicoliths and shells, but these were too few to collect for dating. Overlying the sand is a 1.5 m thick gravel. This gravel is horizontally exposed in a number of fields on the bank of the Vel. The surface gently rises with the gravel overlain by 1 m of silt and then 1 m of black soil. Calcrete nodules are being exposed from only one horizon in the silt and are part of a Bk horizon in the soil. Because the gravel has an extensive exposure on the surface many artefacts and shells could bc collected from it. Two shell samples were collected from this gravel and dated separately. The first sample was of Unio bivalves associated with a part of the gravel which is mostly derived form the HLG exposed nearby and the second comprised gastropods associated with a more sandy facies of the same gravel. The dates are 10,390 130 (BS 1427) for the gastropods and 9,420 90 (BS 1228) for the bivalves. This 1000 year difference between the two specics of molluscs within the same gravel might be partly due to the gravel itself having
+
+
Table 1: Radiocarbon dates from the Vel River near Talegaon Dhamdhere Lab No.
Sample No.
Material Dated
Age
Calibrated Age
lo Age Range
-
BS-1427
TLG3
Gastropods
6,930f 120
7,690
7890-7590
BS- 1228
TLG 1
Bivalves
9,420+ 90
10,380
10,780-1 0,300
BS-1427
TLG2
Gastropods
10,39W 130
12,280
12,460- 12,060
-
-
-
Calibration using the programme of Stuiver and Reimer (1993)
Studies in the Geomorphology, Quaternary Palaeoenvironments and Archaeology of the Vel River
Sections from some Localities around Talegaon on the Vel River Metres
mw Eoq 1
S.no 5,7,8
Recent Silt Holocene gravels High Level Gravels Calcareous Silts Compact Brawn Silt Black Fissured Clay Weathered Basalt Less Weathered Basalt Calcrete nodule1Bk hor~zon Vertisol
S.no. 2
LOC1, 200 bp silt
LOC2, HLG
Loc 2. 10 kyr gravel
Loc 6, Bhima alluvium
Fig. 3: Sections from some localities around Talegaon on the Vel River
existed for some length of time, and also due to different fractionation and reservoir effects of the two species. This time period is one of rapid change in the CO, of the atmosphere also, so that calibration of the radiocarbon dates at this time introduces more uncertainty into the dates (Table 1). However, the two dates do show that a significant aggradational phase on the Vel River occurred at the Pleistocene-Holocene transitional period (Mishra et al. 1998).
"relatively humid" might still be a semi-arid climate). It is possible that this river, having less discharge than the rivers originating in the Western Ghats, would become ephemeral during relatively arid phases and have little capacity to either deposit or erode. The depositional events might belong therefore to periods of relative humidity. The period 14-8 kyr was one of rapid climatic change and the dating itself cannot resolve the actual age of the event either to a well known arid phase or humid phase.
Interpretation of the climate during this phase is ambiguous. The local immature nature of the gravels implies a semi-arid climate, but whether this was more humid or arid relative to the present climate is more difficult to determine (i.e. as it is in a fairly arid zone,
Some sedimentological studies of the gravels from this locality, as well as a study of the microliths collected from a 5 m x 5 m area of this gravel was undertaken for an M.A. thesis (Adav 1998; Adav and Deo 1998). The results of these studies show that the while the gravel from which the
Marl and Etwirormerlt XXIV ( I ) - I999 gastropods were collected shows normal distribution with a modal value in the 1-2 mm size range (making it a sandy gravel) the gravel from which the artefacts and bivalves were collected are bimodal with a coarse component mode in the 8-12 mm size range and the main mode in the 1-2 mm size range (Fig. 4) The coarse component is probably derived from the nearby exposure of HLG. It would appear that during this aggradation the presently exposed bedrock bed would have been buried. The artefact collection also has a primarily microlithic industry, but with occasional abradedtweathered flakes with MiddleILower Palaeolithic affinities, which are also probably derived from the surface of the HLGs. Flakes from the gravel have a modal value of 1-2 cm and maximum sizes of 5 cm. Artefacts smaller than 1 cm are rare, but some were collected. The artefact assemblage is a microlithic blade industry. The artefacts are not abraded and flake to core ratio is high showing that although the context is from a gravel, the assemblage has not been severely modified by geological processes.
The well developed soil overlying the gravel is laterally fairly extensive. The thickness of this soil is over I m and a well developed Bk horizon is seen. The dating of this gravel therefore also gives us a maximum age for the soil development. Early Holocene aggradation (7 kyr) Upstream of Talegaon village there is a minor knickpoint. Upstream of this knickpoint an alluvial fill is seen. Most of this fill comprises calcareous silts and pebbly cobbly gravels in which no artefacts or other dateable material could be found. A small sandy gravel, similar in characters to that from Loc 1 which contains the gastropods is seen. This forms a 4-5 m fill inset to the older alluvial deposits. Microliths were found in this gravel. It is also rich in gastropods which were extracted for dating. These gastropods were dated to 6,930 120 b.p. (BS 1427). This dating shows that the two gravels, although they have similar
+
Locality 2, Sample 5, Talegaon
Locality 2, Sample 2, Talegaon rercentage
Lower gravel with major
_
_ - - corneom!-C.~!rew_o&e_~_HLG_ pebbles.
-
_.- - -
-
_._
- - -
-4
.2
-3
-1
1
0
2
3
4
4
44
phi value phi value
Locality 2, Sample 7, Talegaon
Locality 2, Sample 8, Talegaon
-3
-2
-1
0
1
2
3
phi value
Fig. 4: Size distributions for gravel samples from Locality 2
1 64
4
-4
-3
-2
-1
0
1
phi value
2
3
Studies in the Geomorphology, Quaternary Palaeoenvironrnents and Archaeology of the Vel River
characters do not belong to the same time period. The gravel near Bhairava temple therefore belongs to a later episode of aggradation compared with the one from the Talegaon Bazar locality. Mid-Holocene aggradation (2.5 kyr)
Chalcolithic pottery was found at Loc 4 (Fig. 2). This was earlier reported by Naik and Mishra (1997). This Chalcolithic pottery was collected from a 20 cm colluvial gravel sandwiched between silty clay which is exposed at the margins of the zone affected by the Vel in flood. The pottery therefore was incorporated into a colluvial deposit, probably shortly after the abandonment of the site. This colluviation implies an unstable land surface, probably denuded of vegetation. The total extent of this site is less than 3 m x 3 m, and so is quite unlike any other Chalcolithic site so far reported. The pottery assemblage is similar to that collected from the nearby site of Sheriwadi (Naik and Mishra 1997) and has some similarities to that of Theur reported by S.R. Rao (IAR 197213: 35). It has a link with the Chalcolithic in that one ware is related to Jorwe ware, but has a coarser fabric, slightly different slip and cruder paintings. The associated wares are black and red ware with abundant graffiti marks and the red slipped ware which is also found in the Early Historic period. This pottery assemblage obviously has affinities to both the Chalcolithic and Early Historical assemblages and Naik and Mishra place it in the transitional period between the two phases. This probably belongs to the so-called "gap" period identified from western Maharastra during this time (2200-1 800 b.p., 1200-800 B.C.). The landscape instability indicated by this site provides limited evidence for relative aridity during that period. Late Holocene aggradation (200 yrs bp)
The Bhairava temple - about 200 years old - is built on bedrock near the HLG exposure. The temple is now about 100 m away from the river. The intervening area is filled with a very recent appearing brown silt. Quite a bit of pottery can be seen in this brown silt, both on the surface and in the section exposed on the river bank. This appears to be deposited after the construction of the temple which might have been on the bank of the river when it was first built. While this aggradation is minor and may have no climatic significance, similar fills have been observed at a number of other sites in Maharashtra such as Songaon on the Karha-Nira confluence and Navgaon on the Virabhadra River near its confluence with the Godavari near Paithan. Late Pleistocene Aggradation of the Bhima This study has been mainly concerned with the Vel alluvium. At Loc 5 (Fig. 2) about 1 km from the Bhima a small man-made pond has exposed a thick black fissured clay. This fissured clay has well developed calcrete
nodules. A few microliths were collected from this locality. This alluvium is most probably the older alluvium of the Bhima. The presence of microliths also indicates that at least some of it probably belongs to the Late Pleistocene. At Loc 6 (Fig. 2), an almost 10 m thick section of brown clays sandwiching a 2 m gravel was observed (Fig. 3). This gravel also had a few microliths. A few bivalves were found but they were fragile and weathered compared to those from Loc 2 and we could not collect enough for radiocarbon dating. The Bhima alluvium extends slightly more than 1 km from the present channel. It is predominantly clayey and silty with occasional gravel lenses, while the Vel alluvium is predominantly sandy and gravelly. None of the Bhima alluvium has been dated at this locality, but in general the dates for the alluvium in western Maharashtra belong to the Late Pleistocene. Prior to our study, only one Early Holocene date had been obtained for the Bhima Basin. This was from Akoni (Kale and Rajaguru 1986), on a small tributary of the Bhima in the downstream region. Therefore while we cannot be categorial that the Bhima and Vel alluvium at Talegaon are of different ages, it is suggestive that all the dates on the Vel alluvium are Holocene, while all dates on the Bhima alluvium are Pleistocene. The only other Early Holocene date is also from the alluvium of a Bhima tributary. This different response of the fluvial system to climatic change is due to the different amount of rainfall in the catcment zones and also the different sizes of the basins. The Vel River has less rainfall and a smaller catchment than the Bhima and would respond to relative aridity by becomip'g more ephemeral and would have little capacity to deposit sediment in the absence of sufficient discharge. During relatively humid phases the abundant relict sedident~weatheredmaterial would be locally deposited and reworked. The Bhima River, on the other hand would be able to erode material from its basin during relatively humid periods and would aggrade during relatively arid periods when the sediment to discharge ratio increased. While this complex fluvial response to climatic change complicates straightforward climatic interpretations, it means that the fluvial record missing in one part of the system is present in another part of the system. This has provided us a glimpse into the Early Holocene story of western Maharashtra which was missing earlier. Important Implications of the Study It is only in the last few years that the importance of the Tertiary in the present landscape has been realized. HLGs are part of the regolithic cover on the lower dedunational surfaces (700 m-540 m). This regolith has supplied the Quaternary fluval systems which are characterized by local reworking of the Tertiary material. The Vel appears to have responded differently than the Bhima to changing climate during the Late Quaternary.
Man and Environment XXIV ( I ) - 1999 While the Bhima and Ghod were aggrading during the phase of Late Quaternary arid climate, all the units on the Vel which could be dated belong to the Holocene. The only other Holocene date from the Bhima basin also is from a small tributary, the Nandi, at Akoni (Kale and Rajaguru 1986). The microliths from the Vel have been dated to 7 and 10 kyr, providing new evidence for human adaptations during this time period. This is just a begining, and as yet little beyond the presence of a microlithic industry during that time has been established. However further archaeological work will be a priority as it is during this time that man transformed his subsistence from hunting gathering to plant and animal domestication.
Kale, V.S. and S.N. Rajaguru 1988. Morphology and Denudational Chronology of the Coastal and Upland River Basins of Western Deccan Trappean Landscape (India): a Collation, Zeitschrqt fur Geomorplzologie 32(3): 3 1 1-327 Mishra S., S.N. Rajaguru, S. Naik, S.N. Ghate and A. Kshirsagar 1998. Climatic Change during the Pleistocene/Holocene Transition in Upland Western Maharashtra, Western India, in Water Environment and Society in Times of Climatic Change. (A.S. Issar and N. Brown Eds.), pp. 323333. Netherlands: Kluwer Academic Publishers. Mishra, S. 1992. The Age of the Acheulian in India, Current Anthropology 33: 118-124,
The Chalcolithic site of Talegaon belongs to a period for which no archaeological evidence was known earlier. It also provides some corroborating evidence for increased aridity during that period.
Mishra, S. 1995. Prehistoric and Quaternary Studies at Nevasa: the Last Forty Years in Quaternary Environments and Geoarchaeology of India (S. Wadia, R. Korisettar and V.S. Kale Eds.) pp. 334332. Bangalore: Geological Society of India.
References
Mishra, S., T.R.Venkatesan, S.N. Rajaguru and B.L.K. Somay ajulu. 1 995. Earliest Acheulian Industry from Peninsular India. Current Anthropology 36: 847-85 1.
Adhav, Utpala 1998. Talegaon-Dhamdhere yethil Vel Nadichya parisaracha Bho-Puratattva shashtriya Abhyas. M.A. Dissertation. Pune: Deccan College. Adhav, Utpala and Sushama G. Deo 1998. TalegaonDhamdhere : Ek Bhu-puratattva Vishleshan, Samshodhak 66(4): 3- 17 (in Marathi). Corvinus, G. 1983. The Pravara River System Vol2: The Excavations of the Acheulian Site of Chirki-onPravara, India. Tubingen: Archaeologica Ventoria Institute fiir Urgeschichte. Dikshit, K.R. 1970. Polycyclic Landscape and the Surfaces of Erosion in the Deccan Trap Country with Special Reference to Upland Maharashtra, National Geographic Journal of India 16(3-4): 236-252. Foote, R.B. 1876. The Geological Features of the South Mahratta Country and Adjacent Districts, Memoirs of the Geological Survey of India 12: 1-268. Joshi, R.V., R.S Pappu, A.R Marathe and R.P. Pandey 1980. Excavation at Wadoli-Wagholi, a Middle Palaeolithic Site of the High Level Gravel on the Godavari River, Bulletin of the Deccan College Research Institute 39: 49-89. Kale, V.S. and S.N. Rajaguru 1987. Late Quaternary Alluvial History of the North Western Deccan Upland Region, Nature 325: 612-61 4.
Mishra, S., V.S.Kale and A. Kshirsagarl993. High Level Gravel Occurrence Near Atner, on the River Tapti Bulletin of the Deccan College Post-Graduate and Research Institute 53: 253-259. Rajaguru, S.N. and V.S. Kale 1985. Changes in the Fluvial Regime of Western Maharashtra Upland Rivers During Late Quaternary, Journal of tlze Geological Society of India 26(1): 16-27. Rajaguru, S.N., V.S. Kale and G.L. Badam 1993. Quaternary Fluvial Systems in Upland Maharashtra,. Current Science 64(11-12): 81 7-822. Sadakata, N., H. Maemoku, S.N. Rajaguru, S. Mishra and K. Fujiwara 1995. Late Quaternary Environmental Change in the Pravara Basin, Northwestern Deccan Upland, India, in Proceedings of the International Symposium on Palaeoenvironmental Change in Tropical-Subtropical Monsoon Asia (Ed.), pp. 43-56. Special Publication no. 24 of the Research Center for Regional Geography, Hiroshima University. Stuiver, M., and P.J. Reimer 1993. Extended I4CDatabase and Revised Calib Radiocarbon Calibration Program, Radiocarbon 35: 215-230.
