Initial Geologic Provenience Studies of Stone and Metal Artefacts from Rakhigarhi

Initial Geologic Provenience Studies of Stone and Metal Artefacts from Rakhigarhi Amarendra Nath1, Randall Law2 and Tejas Garge3 . Former Director, Archaeological Survey of India, New Delhi, India (Email: [email protected]) 2. Department of Anthropology, University of Wisconsin, Madison, USA (Email: [email protected]) 3. Archaeological Survey of India, Aurangabad Circle, Bibi ka Makbara, Aurangabad, Maharashtra – 431004, India (Email: [email protected]) 1 Received: 30 August 2014; Accepted: 24 September 2014; Revised: 17 October 2014 Heritage: Journal of Multidisciplinary Studies in Archaeology 2 (2014): 74‐100 Abstract: Rakhigarhi was the largest settlement in what Gregory Possehl defined as the “eastern domain” of the Indus Civilization. However, there are no rock or mineral resources whatsoever within 50 km of the ancient city. All stone or metal object recovered during excavations at the site – from the tiniest bead to the largest grindingstone – were made of raw material that had to have been imported from distance sources. Geologic provenience studies can provide compelling evidence that an exchange network once existed between the ancient inhabitants of Rakhigarhi and those in the region where the raw material the artefact is composed of originated. In this paper, we present the initial results of the several ongoing geologic source provenience studies – instrumental neutron activation analysis of steatite and agate production debris, Pb isotope analysis of lead and silver objects and a large scale comparative study of grindingstones. Keywords: Rakhigarhi, Geologic Provenience Studies, Steatite, Agate, Lead, Silver, Grindingstone Introduction Rakhigarhi was one of largest Harappan settlements. Located in what Gregory Possehl (1992) defined as the “eastern domain” of the Indus Civilization, the ancient city could reasonably be considered the “provincial capital” of that domain. The site was surrounded by network of subsidiary settlements in a densely populated zone that provided its residents with ample agricultural surplus, forest products, and other resources from the local alluvial plain (Garge 2011). However, there are no rock or mineral resources whatsoever within 50 km of Rakhigarhi. Excavations there have revealed, among other things, a broad range of stone and metal artefacts made from raw materials that had to have come from sources outside of this local zone. The study of these objects allows us to reconstruct non‐local resource acquisition patterns at Nath et al. 2014: 74‐100 Rakhigarhi and, by doing so, examine its residents’ connections with other “domains” and/or resources areas across the Harappan realm and beyond. In the absence of historical accounts of trade, geologic provenience studies can provide compelling evidence that an exchange network once existed between the ancient inhabitants of a region where a stone or metal artefact entered the archaeological record and those in the region where the raw material the artefact is composed of originated. The soundest provenience determinations are ones based on analyses of artefacts composed of unadulterated rock or mineral rather than processed metal, which could contain metal from multiple sources as well as various alloys and additives. With unadulterated stone, “specific types of raw materials can be related to an objective geologic reality that is derived from natural (as opposed to cultural) processes” (Odess 1998: 419).There have been several major broad‐scale studies (Fentress 1976; Lahiri 1992; Ratnagar 2004) of Harappan trade networks that examined multiple varieties of stone and metal to construct models of proto‐historic resource access and exchange. Randall Law (2011) built upon these earlier studies by, using a variety of analytical methods, conducting large‐scale direct comparisons of artefacts from Harappan sites to samples collected from potential stone and metal sources throughout the greater Indus region. In 2007, he was invited by the Rakhigarhi’s excavator, Dr. Amarendra Nath, to begin geologic provenience studies of that site’s stone and metal artefact assemblage1.The report herein is a summary of the progress and preliminary findings of the geologic provenience analyses undertaken to date for this project. Rakhigarhi and Its Local Physiographic Environment The site of Rakhigarhi was excavated for three seasons – 1997‐98 (Nath 1998), 1998‐99 (Nath 1999) and 1999‐2000 (Nath 2001). Multiple mounds were identified and numbered as RGR 1 through RGR 7. Of these, RGR 1, RGR 2 and RGR 6 were revealed to have, in addition to Harappan Period occupation, a formative stage followed by an Early Harappan settlement stage. RGR 7 is a necropolis. The settlement lies on the flood plain of Sarasvati‐Drishadvati basin. The most up‐to‐ date picture of the local physiographic environment around Rakhigarhi was compiled by Tejas Garge (2011). In the past, the Drishadvati (or Chautang) River flowed through the modern districts of Karnal, Jind and Hissar before meeting the Sarasvati near Suratgarh in Rajasthan. Topographically, this area is a, monotonous upland terrain that is part of the alluvial of Satluj‐Yamuna plain; the western portion of which gradually transitions into the Thar Desert. Prominent features are aeolian sand deposits of variable shapes and thicknesses overlying the Pleistocene alluvium. The patches of older alluvium are either exposed or occur at shallow depth beneath a veneer of sand in tals or topographic depressions enclosed by fossilized dunes known as tibbas. Silted‐ up river channels with continuous or intermittent levee undulations occupy a relatively lower position. The general gradient of the terrain is from north‐east to south‐west and then west. 75 ISSN 2347 – 5463 Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 In terms of this study it is important to note that the bedrock of this area is deeply buried and is wholly concealed under alluvial and aeolian deposits. There are no accessible sources of stone or metal whatsoever available anywhere within the local area. The nearest outcrops of rock are located in southern Haryana approximately 50 km (Tosham area) to 75 km (Kaliana Hills) south of Rakhigarhi. The main part of the Aravalli Range beings some 125 km to the south while the foothills of the Himalayas begin to rise around 175 km to the northeast. Therefore, all stone or metal object recovered during excavations at Rakhigarhi – from the tiniest bead to the largest grindingstone – was made of raw material that had to have been imported to the site from one of these sources or even more distant ones. Observations, Analyses and Results Steatite Artefacts Steatite –a rock composed mainly of the mineral talc – was one of the most important types of stone in the Harappan corpus of raw materials. It was used to create common items, such as disc beads, as well as important objects like stamp seals and inscribed tablets. Although steatite occurrences are found in many parts of South Asia (Law 2002), geologic provenience analyses of artefacts from sites across the Indus Civilization (Law 2011: Chapter 7) indicate that Harappans mainly acquired this type of stone from sources located in what today is northern Pakistan, primarily those in the Hazara District (Figure 1). Deposits in some other areas, such as northern Rajasthan, were also exploited but to a much more limited degree. Steatite objects from Rakhigarhi were examined and/or analyzed in an attempt to determine what sources residents at that site utilized. Raw steatite exhibits a wide range of colors and patterns. However, the vast majority of objects made from this stone at Rakhigarhi and other Indus Civilization sites appear solid white due to having been heated at high temperatures and/or covered with an applied white surface. At sites like Harappa and Mohenjo‐Daro, thousands of raw steatite debris fragments and unfinished steatite objects have also been recovered. In contrast, only a handful of such artefacts were observed among the Rakhigarhi materials. Perhaps more will come to light as studies of the collection continue. The examples examined are composed of types of raw steatite that are identical in appearance to types documented by Randall Law at Harappa (see Law 2011: Figure 7.4). Below are three examples. Figure 2 compares two fragments – one from Rakhigarhi [A] and one from Harappa [B] – of khaki‐colored steatite. In Figure 3, an unfinished stamp seal from Rakhigarhi [A] carved from an olive‐green and black‐ banded type of steatite is compared to a large debris fragment from Harappa [B] composed of identical looking stone. Figure 4 shows stamp seal RGR 7230 [A] and a detail of its side [B] where the surface has partially worn away to reveal the black steatite from which the object was carved. A sawn piece of black steatite manufacturing debris from Harappa [C] is shown for comparison. 76 Nath et al. 2014: 74‐100 Figure 1: Steatite sources of the Greater Indus region and Harappan steatite trade networks 77 ISSN 2347 – 5463 Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 2: Fragments of khaki‐colored steatite debris from [A] Rakhigarhi (RGR 573.B) and [B] Harappa Figure 3: [A] An unfinished stamp seal (RGR 1731) from Rakhigarhi carved from an olive‐green and black‐banded steatite. [B] A debris fragment of the same material from Harappa. Figure 4: [A] SealRGR 7230 from Rakhigarhi.[B] The side of the seal where the surface has partially worn away revealing the black steatite beneath. [C] A sawn black steatite debris fragment from Harappa. Visual examinations, while informative, are generally not sufficient for identifying the geologic sources of steatite artefacts (the reason being that identical macroscopic types of stone can and often do occur at different steatite deposits). That task is best undertaken using instrumental neutron activation analysis (INAA). This highly 78 Nath et al. 2014: 74‐100 accurate and precise method for quantifying the major, minor and trace element compositions of materials has been employed by researchers around the world in efforts to identify the geologic sources of a wide range of archaeological stone. In brief, INAA involves the irradiation (or activation) of elements within materials by exposing them to a neutron flux. Following varying periods of decay, the gamma ray emissions they produce are detected and counted. After the results are screened of elements that failed to be detected in all samples or had high count‐rate standard deviations, the data are evaluated using canonical discriminant analysis (CDA). During CDA, linear combinations of variables called discriminant functions are generated that produce a maximum degree of separation (discrimination) between various defined groups of cases, which in this study are the individual sets of samples that Law personally collected from different geologic sources in India and Pakistan. Artefacts are plotted as ungrouped cases and assigned to the group (geologic deposit) whose center (or centroid) in multidimensional space they are nearest. Steatite samples from four artefacts were subjected to INAA. Two of the four were pieces of manufacturing debris, one of which (RGR 573.B) is pictured in Figure 2 A. The other (not pictured) was a 4mm sliver of brown steatite from a bag of unnumbered surface materials. We designated this fragment RGR‐s1. A third sample was taken from a battered block of steatite (RGR 3607) that, in all probability, was a blank or roughout for a stamp seal (Figure 5). Loose material from a damaged area of the block was collected for analysis. The final artifact sampled (RGR 6304) was a fragment of a unicorn stamp seal (Figure 6 A). A small amount of the grayish‐green steatite exposed in its broken section (Figure 6 B) was carefully removed for analysis. All of the artefacts are from the mound at Rakhigarhi designated RGR‐2. The two debris fragments are surface finds while the seal blank and seal fragment are from Mature Harappan levels. Preparation of the Rakhigarhi steatite samples for INAA took place at the Laboratory for Archaeological Chemistry, Department of Anthropology, University of Wisconsin‐ Madison. Analysis was conducted at the University of Wisconsin’s Nuclear Reactor (UWNR) research facility by the team supervised by lab director Robert Agasie. The elemental data generated for the four artifact samples, which are listed in Table 1, were compared using CDA to a database of geologic samples collected from 37 steatite sources across India and Pakistan (Law 2011: Appendix 7.3). Those sources are identified by deposit and region on Figure 1 of this report. Figure 7 shows the 442 geologic samples from the 37 sources plotted using the first and second discriminant functions generated by CDA. For reference, the 141 steatite artefacts from Harappa that have been analyzed to date are plotted (using red triangle symbols on the figure) as ungrouped cases in relation to the sources. The four artefacts analyzed from Rakhigarhi are similarly plotted (using red diamond symbols) above the Harappa samples. The predicted group (source) membership for three of the artefacts (RGR‐s1, RGR‐6304 and RGR 3607) was one of the steatite deposits located in the Hazara District, Khyber Pakhtunkhwa Province or KPP (formerly known as the North‐West 79 ISSN 2347 – 5463 Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Frontier Province or NWFP), Pakistan. The fourth artifact (RGR 573.B)was assigned membership among a group of geologic samples from the Daradar steatite deposit in the Safed Koh Range of the Kurram Agency, Federally Administered Tribal Areas (FATA), Pakistan. Figure 5: Seal blank or roughout, artifact #3607 Figure 6: [A] Unicorn seal fragment #6304. [B] Detail of the grayish‐green steatite of the sealʹs interior. Although only four steatite artefacts from Rakhigarhi were analyzed, the results strongly suggest that residents of the site were part of the same extensive acquisition/distribution network for this raw material as their fellow Harappans at other Indus Civilization cities. Stone from sources in northern Pakistan makes up 95% of such artefacts analyzed from Harappa and approximately two‐thirds of those tested from Mohenjo‐Daro and Dholavira. The percentage from Rakhigarhi presently stands 80 Nath et al. 2014: 74‐100 at 100% but that is almost certain to change when a larger sample from the site is tested. Steatite from sources in the Alwar and Jhunjhunu districts of northern Rajasthan has been identified at Harappa and Mohenjo‐Daro (Law 2011) as well as at Mitathal in south Haryana (Prabhakar et al 2010). Given Rakhigarhi s relative proximity to that region, it is highly probable that stone from deposits there will eventually be detected in its assemblage. Table 1: INAA data for four steatite artifacts from Rakhigarhi Al Co Cr Eu Fe La Mn Na Sc V Zn RGR 573.B 2090 0.842 0.565 0.0308 2810 0.010 2.160 320.0 0.008 1.480 10.90 RGR‐s1 1798 0.786 1.683 0.0447 2224 0.643 14.93 302.3 0.069 3.459 24.88 RGR 6304 1657 0.631 2.054 0.6621 2886 0.179 5.227 233.8 0.074 1.326 20.01 RGR 3607 1844 1.357 2.616 2.0980 1406 0.436 11.86 364.2 0.123 4.627 27.36 Data listed in parts per million (ppm) Figure 7: Steatite artefacts from Rakhigarhi and Harappa are plotted as ungrouped cases in relation to the 442 samples (37 sources) in the geologic data set At this point it is not possible say how steatite from northern Pakistan came to Rakhigarhi in Haryana. The paths of the trade routes drawn on Figure 1, including the highlighted one extending from Harappa to Rakhigarhi, are entirely conjectural. However, given Harappa s geographic position in relation to the northern Pakistan deposits and the fact that some 95% of the steatite at the site seems to have been derived from sources in that region, it is not unreasonable to speculate that the city might have been the center where this stone was first gathered and then distributed to 81 ISSN 2347 – 5463 Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 consumers across the Indus Civilization. Thus, we have drawn the trade routes to reflect this possibility. One may ask, why would residents of Rakhigarhi and other Indus sites use steatite from northern Pakistan when there were other, often closer, sources available? The reason is that stone from deposits in the north, especially those in the Hazara District, possessed physical qualities that Harappans sought (i.e., it becomes pure white when heat‐treated). Agate‐carnelian Artefacts Ornaments (mostly beads) made from the translucent reddish‐orange variety of agate known as carnelian are almost as ubiquitous at Harappan settlements as steatite objects. Geologic provenience studies of carnelian artefacts from a half dozen sites across the Indus Civilization (Law 2011: Chapter 8) appear to confirm what researchers have long suspected – i.e., that this variety of stone was mainly derived from deposits located in the Gujarat region. It seems, however, that the agate deposits Harappans primarily exploited occur in northern Gujarat rather than, as was widely assumed, the more famous source at Ratanpur in the southeastern part of that state. Agate‐carnelian artefacts from Rakhigarhi were examined in order to determine if the same acquisition pattern would be evident at that site, which of all Indus cities was the most far removed from the Gujarat sources. Due to time limitations, only cursory visual examinations of finished agate‐carnelian ornaments were conducted. Randall Law’s overall impression was that carnelian beads at Rakhigarhi are basically identical to other such artefacts at Harappan sites, both in terms of their styles and the material from which they were made (that is to say, Gujarati carnelian). However, he observed a number of examples made from carnelian exhibiting a hue and/or patterning that is somewhat unlike that he encountered during his explorations of sources in Gujarat. While this does not rule out Gujarat as the source area for those particular artefacts, it does open the possibility that raw material from occurrences in alternate regions was sometimes exploited. It is clear though that agate‐carnelian ornaments were being manufactured at Rakhigarhi. Artefacts representing all stages of bead production were observed in the collection. Among these materials are nodule fragments, which indicate that at least some carnelian was transported to the site in raw, unmodified form. Five carnelian artefacts (nodule fragments and non‐diagnostic bead making debris), all of which were recovered from the surface or near surface levels on mound RGR‐2, were selected for INAA (Figure 8). Using CDA, the elemental data generated (Table2) were compared to a database of geologic samples from three agate sources in Gujarat and one in eastern Iran (Figure 9). The Gujarat sources include the deposit at Khandek in eastern Kachchh (some 70 km from the Harappan city of Dholavira), the extensive agate beds on Mardak Bet in the Little Rann of Kachchh (105 km from Dholavira), and the famous agate mines of Ratanpur (390 km from Dholavira). The source in Iran is represented by proxy using carnelian nodule fragments recovered from the proto‐ historic site of Shahr‐i‐Sokhta. All five of the Rakhigarhi artefacts were assigned to one 82 Nath et al. 2014: 74‐100 Figure 8: Agate‐carnelian nodule fragments and flakes from Rakhigarhi Table 2: INAA data for six carnelian artifacts from Rakhigarhi Al Co Cr Eu Fe La Na Sb Sc V RGR 234 1485 0.290 0.597 0.0470 644.2 0.0254 83.49 0.0819 0.004 1.228 RGR 573 1557 0.837 0.783 0.0461 571.4 0.4638 69.60 0.3380 0.036 1.153 RGR 7220 A 1807 0.055 0.322 0.0226 443.0 0.0408 136.10 0.4227 0.023 1.581 RGR 7220 B 1626 0.532 0.404 0.0662 1060.0 0.0560 63.98 0.3177 0.051 1.014 RGR‐a1 1642 0.068 2.915 0.0270 659.5 0.5868 257.70 0.0555 0.571 1.641 Data listed in parts per million (ppm) Figure 9: CDA comparison of Rakhigarhi carnelian artefacts to Gujarati and Iranian agate sources 83 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 10: Greater Indus Valley region agate occurrences and carnelian acquisition networks 84 Nath et al. 2014: 74‐100 of the Gujarat sources – three to the Khandek deposit (RGR 234, RGR 572 and RGR 7220 B), one to the Mardak Bet deposit (RGR‐a1) and one to the Ratanpur mines (RGR 7220 A). Although they represent an extremely small sample of agate‐carnelian bead production debris at the city, the five provenience assignments do correspond to the general acquisition pattern for this raw material detected at multiple Harappan sites (Figure 10), which is – most of the carnelian utilized came from sources in northern Gujarat while only minor amounts originated in the more famous (but more distant) mines of Ratanpur. These results should be treated cautiously, however, as there are many agate‐carnelian sources in the Greater Indus region and beyond that remain to be characterized (some of these are identified on Figure 10). Note that in Figure 9 two samples – RGR 573 and RGR‐a1 – plot somewhat away from the datapoints representing the deposits to which they were assigned (Khandek and Mardak Bet respectively). It is possible that those particular artefacts were derived from a source(s) not represented in the geologic database. Lead and Silver Artefacts Lead objects and silver ornaments, while generally not abundant at Harappan settlements, have been recovered at numerous sites across the Indus Civilization. Occurrences of lead, many of which are viably argentiferous (meaning that they contain an extractable quantity of the precious metal silver), are found in numerous parts of the Greater Indus Valley region as well as areas outside of it with which Harappans are known to have had long‐distance contacts. Archaeologists have long employed lead (Pb) isotope analysis in efforts to identify the geologic sources of artefacts composed of or containing that metal. Using this method, a large‐scale provenience study was initiated in 2002 and has now grown to include samples from over three‐dozen geologic sources and artefacts from eight Harappan sites (Law and Burton 2006; Law and Burton 2008; Law 2011: Chapter 12). Those sources and sites are identified onFigure 11. Although this study remains ongoing, a picture of lead and silver acquisition networks during the Harappan Period (identified using lines and arrows on Figure 11) is beginning to emerge. It appears that peoples across the Indus Civilization primarily acquired these metals from deposits in southern Balochistan. However, residents of Harappa and Dholavira also utilized lead and/or silver nearer sources (deposits in Jammu and Kashmir, and northern Gujarat respectively). Five artefacts from the Rakhigarhi collection were isotopically assayed in order to determine if a similar acquisition pattern existed there. The ancient city lies approximately 950 km from the southern Balochistan source area but there are rich lead and lead‐silver deposits some 200 to 400 km north of the site in the Himalayas. The first object sampled was a plano‐convex (or bun‐shaped) lead ingot (Figure 12 A& B) inscribed with Harappan characters on both sides (Figure 12 C & D). A small prism‐ shaped piece of lead (Figure 13) was also assayed. The remaining three artefacts sampled were silver ornaments – two hoops and a small disc (Figure 14). All of the objects were recovered from excavations on mound RGR‐2, save the prism‐shaped piece of lead, which was from mound RGR‐4. 85 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 11: Isotopically assayed lead ore and lead‐silver deposits of the Greater Indus region, archaeological sites from which lead or silver artifacts have been analyzed, and lead and/or silver acquisition networks 86 Nath et al. 2014: 74‐100 Figure 12: Side [A] and top [B] views of a lead ingot inscribed with Harappan characters. Detail images of the top [C] and bottom [D] inscriptions. Figure 13: Lead piece, RGR 424 87 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 14: Three silver ornaments In order to extract lead from the artefacts for isotopic analysis, a solution was prepared that consisted of ultrapure water and 0.05% dissolved EDTA, which is a hexadentate chelating agent that forms coordinate bonds with lead atoms. Each artifact was immersed in the solution for five minutes, which is usually sufficient to extract lead atoms in concentrations from 100 ppb to as much as 100 ppm – orders of magnitude more than required for isotopic analysis. At that point, the lead‐enriched solutions were poured into sample vials for return to Madison. The artefacts were rinsed in ultrapure water, allowed to dry, and then returned to their place of storage. The brief immersion time in the sampling solution did not result in any macroscopic alteration whatsoever of the artefacts. Table 3: Pb isotope data for two lead and three silver artifacts from Rakhigarhi Pb207 / Pb206 Pb207 / Pb204 Artifact Pb208 / Pb207 Inscribed lead ingot 2.4677 0.84747 15.682 RGR‐424, lead lump 2.4393 0.86481 15.667 RGR‐4386, smaller silver hoop 2.4666 0.84675 15.679 RGR‐4098, larger silver hoop 2.4574 0.85137 15.650 RGR‐3949, silver disc 2.4539 0.85594 15.676 Once the artifact sample solutions were back in Madison, they were prepared and then sent on to the Keck Isotope Laboratory at the University of California‐Santa Cruz where they were analyzed by Dr. Emily Peterman on a Thermo Scientific NEPTUNE multiple‐collector inductively‐coupled‐plasma magnetic‐sector mass‐spectrometer (MC‐ICP‐MS). The results, which are listed in Table 3, were compared to the database of Pb isotope values for South Asian lead and lead‐silver ore deposits. In Figure 15, the values for the five artefacts are plotted in relation to select ore fields using the ratios 208Pb/207Pb (y axis) and 207Pb/206Pb (x axis). Three of the artefacts – the inscribed lead ingot and both silver hoops – plot squarely within the lead isotope field defined by geologic samples from the southern Balochistan sources. The silver disc and the prism‐ shaped lead piece fall somewhat away from that field in a part of the plot that is not, at 88 Nath et al. 2014: 74‐100 present, represented by a lead source. These two artefacts could either be made of metal derived from a deposit (or deposits) not in the database or composed of metal from two or more sources. The latter possibility might have given the objects isotopic characteristics that cause them to plot along a mixing line between a southern Balochistan deposit and an unknown source (s). Figure 15: Lead and silver artefacts from Rakhigarhi compared to South Asian lead and lead‐silver sources Grindingstones Most of our time with the Rakhigarhi collection was spent examining querns, mortars, mullers, pestles and whetstones, which we will collectively refer to here as grindingstones. These constitute an especially important and informative category of stone artifact. Grindingstones were essential utilitarian tools that would have been used daily for processing foods as well as for performing numerous kinds of craft activities. Moreover, they were among the bulkiest (that is to say, heaviest and largest) stone objects acquired by Harappans. Providing a steady supply of grindingstones to 89 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 the urban population dwelling at Rakhigarhi, where there are no local stone resources whatsoever, would have necessarily required a significant expenditure of time and energy. Law had previously conducted a study at the site of Harappa in which the entire assemblage of grindingstones was compared to samples collected from potential geologic sources in and around the upper Indus Basin (Law 2011: Chapter 5). The general source area for around 70% of the artefacts was determinable based on an assortment of qualitative physical criteria including rock sub‐variety, color, texture, grain size, patterning, visible inclusions, degree of silicification and toughness. An identical study was initiated for the Rakhigarhi grindingstones. In total, we were able to examine 665 examples of querns, mortars, mullers, pestles and whetstones recovered during excavations at the site and/or from surveys of the local area around it. A rough spatial breakdown (by mound) of where the artefacts came from is provided in Table4. We estimate that together they represent approximately one‐quarter to one‐third of all objects in the Rakhigarhi grindingstone collection. Table 4: A rough spatial breakdown of the 665 grindingstones examined Location Number of Objects Percentage of Total Examined Mound RGR‐1 18 2.71 Mound RGR‐2 237 35.64 Mound RGR‐4 80 12.03 Mound RGR‐5 5 0.75 Mound RGR‐6 228 34.29 RGR‐mound not specified 83 12.48 Other local mounds 14 2.11 Total 665 100 The geologic source area was determinable for 555 of the 665 grindingstones examined, or just over 83% of the total (Table 5). In descending order, the types of stone identified are: quartzite from the Kaliana Hills in southern Haryana; water‐worn cobbles of various kinds from the Himalayan foothills region; Pab sandstone from the Sulaiman range in Pakistan; and Mathura sandstone from southeastern Uttar Pradesh. One‐hundred ten grindingstones, or around 17% of those examined, could not be confidently assigned to a known geologic formation and, therefore, were designated as source unknown. Nearly three‐quarters of the grindingstones examined are made from stone derived from the Kaliana Hills, which are a series of small outcrops (Figure 16) in southern Haryana around 75 km south of Rakhigarhi. The rock there is actually a variety of Delhi quartzite, which mainly occurs along an extensive zone extending from Northern Rajasthan to the city of New Delhi. However, the quartzite along that main zone is gray in colour, has a highly silicified, often glassy texture and is generally unsuitable for use as grindingstone. The Delhi quartzite found in the Kaliana area outcrops 90 Nath et al. 2014: 2 74‐100 (which h are outlierrs around 50 km west of the main n Delhi quarrtzite formattion) has a tightly y packed graanular textu ure and is stiill used to make m querns, mullers, m mortars and pestless today. Mo ost importan ntly, only thee Delhi quartzite at this location l hass the highly distincctive appearrance – redd dish in colo our with thin red seams (Figure 177) – that is identiccal to the almost a 75% of the grin ndingstones in the Rakh higarhi colleection (see Figuree 18 for an example). e B Both Law an nd Garge hav ve observed d grindingstones made from this t same material m at numerous n a ancient sitess across Harryana and the t Punjab includ ding Harapp pa, where it comprises c 200% of the ov verall assemb blage. Table 5: The geologic source areaas of the 665 grindingsto ones examin ned Nu Sourcee(Material) umber of Percentagee of Total O Objects Examiined Kalian na Hills, Harryana (Delhii quartzite) 494 74.29 Himallaya foothillss (basalt and d quartzite cobbles) c 51 7.67 Sulaim man Range, Pakistan P (Paab sandstonee) 5 0.75 Uttar Pradesh P (Maathura sandsstone) 5 0.75 Sourcee unknown (various) ( 110 16.54 Total 665 100 Fifty‐o one of the grrindingstonees (or a littlee under 8% of o those exaamined) werre whole or fragmentary cobb bles that had h undoub btedly been n shaped in a dynam mic fluvial enviro onment. Som me are comp posed of blaack basalt (Figure 19 leftt image) whille most are white,, gray or pin nk varieties of quartzite (Figure 19 right r three exxamples). Su uch heavily water‐‐rounded co obbles are not n found in n the northeern Aravalliis. These sttones were obtain ned, in all lik kelihood, fro om the footh hills of the Himalayas, which w begin roughly175 r km to the northw west of Rakh higarhi. Theey are quite common in the beds off the major Figure 20), Beas, and Suttlej. rivers draining thee Himalayass such as thee Ghaggar (F F Figure 16: Deelhi Quartziite outcrop near n Kaliana, Bhiwani District, D Haryana 91 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 17: Detail of Delhi quartzite in the Kaliana Hills Figure 18: A grindingstone from Rakhigarhi composed of Kaliana Hills variety Delhi quartzite Figure 19: Water‐shaped cobbles and cobble fragments from the Rakhigarhi grindingstone collection 92 Nath et al. 2014: 2 74‐100 Figure 20: Waterr‐shaped cob bbles of various kinds in i the bed (lleft) and ban nk (right) of the Ghaggar Riveer Althou ugh the ten n grindingsstones madee from onee the remaiining two identifiable i materiial types rep present a sm mall percenttage of the examined artefacts, a theey serve to demon nstrate the long‐distan nce connectiions that Rakhigarhi R h had with regions r far beyon nd the plainss of Haryanaa. Five of the ten are com mposed of a gray‐whitee sandstone that has h a sugary y texture an nd distinctiv ve brown patches (Figu ure 21). This material, known n as Pab sandstone, deerives from the Sulaimaan Range (F Figure 22), which w rises some 550 5 km to th he west‐nortthwest of Raakhigarhi. This T very sam me rock typee makes up around 30% of th he grindingsstone assemb blage at thee site of Harrappa. The remaining posed of finee‐grained reddish sandsstone with liight khaki‐ five grrindingstonees are comp colored patches (Figure 23 A) that iss at immed diately reco ognizable ass Mathura sandsttone. This stone s occurs around 2500 km southeaast of Rakhigarhi in Utttar Pradesh and was w a popular material fo or historic peeriod sculptu ure (Figure 23 2 B) and architecture. Figu ure 21: Pab sandstone s q quern fragm ment (full vieew and detail) from Rak khigarhi 93 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Figure 22: Pab formation in the Sulaiman Range, Pakistan (left) and detail of sandstone found there (right) Figure 23: [A]Mathurasandstone fragment from Rakhigarhi and a [B] Mathura sandstone sculpture Figure 24: Saddle quern (left) and fragment (right) composed of deep red sandstone of unknown origin 94 Nath et al. 2014: 74‐100 One‐hundred ten grindingstones (or around 16% of those examined) could not be assigned to any geologicformation and so were designated source unknown. While most of these artefacts were composed of nondescript stone, some were made from very distinctive types of rock, such as the deep red variety of sandstone pictured in Figure 24. Given Rakhigarhi s relative proximity to the Gangetic Basin is quite possible (though by no means certain) that many of the latter types were derived from sources on the western margins of that area. Figure 25: Hematite cobbles/nodules of unknown origin Among the artefacts designated source unknown are a dozen roughly spherical‐ shaped stones that are metallic‐gray in color and unusually heavy (two examples are pictured in Figure 25). A small sample was taken from a broken example and analyzed on a Rigaku Rapid II X‐ray diffractometer (XRD) at the Department of Geoscience, University of Wisconsin‐Madison. Its XRD spectrum indicates the material is hematite (iron oxide), which is a common ochre mineral. Although there are no marks (striations) visible on their surfaces that would indicated this, it is possible that these stones were powdered to make red ochre. Or they may have simply been heavy pounding tools. Hematite sometimes occurs as nodular masses and so the spherical form of these objects might be natural. On the other hand, they could be cobbles shaped by flowing water. Whatever the case may be, the geologic source of these stones is, at present, unclear. A picture of grindingstone acquisition at Rakhigarhi (Figure 26) is beginning to emerge. It can now be stated that site residents obtained the raw material for such tools from sources located roughly in each of the four cardinal directions. They relied most heavily on stone from the south in the Kaliana Hills, which were among the nearest sources to the city (outcrops in the vicinity of Tosham are slightly closer but the igneous rocks there do not seem to have been exploited for grindingstone material). A substantially smaller portion of the grindingstones used at Rakhigarhi seems to have been derived from riverbeds and/or alluvial deposits to the north of the site in the 95 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Himalayan Foothills region. Long‐distance connections to the west are demonstrated by the presence of a handful of Pab sandstone querns from the Sulaiman Range while a few Mathura sandstone artefacts confirm links toward the Gangetic Basin in the east. Figure 26: Rakhigarhi grindingstone acquisition networks Although only a portion Rakhigarhi s grindingstones have been examined, some changes in raw material acquisition patterns over time seem to be evident. The two mounds from which the largest numbers of artefacts were examined are RGR‐2 (n = 237) and RGR‐6 (n = 228). The grindingstones examined from RGR‐2 were from Harappan Period levels. RGR‐6 is entirely an Early Harappan Period mound. Over 86% of the artefacts from RGR‐6 are attributable to the Kaliana Hills while most of the remaining are from unknown sources. In contrast, just under 60% of the stone in the Harappan Period levels of RGR‐2 came from the Kaliana Hills. Nearly 20% came from the Himalayas while the rest are Pab sandstone, Mathura sandstone or from unknown sources. So at Rakhigarhi there appears to have been a shift over time toward the acquisition of grindingstone from more diverse and distant sources. A similar (but even more pronounced) shift was detected at the site of Harappa between the Early Harappan and Harappan periods. Summary of Findings and Outline of Continuing Research The Rakhigarhi stone and metal acquisition networks that have been identified thus far are summarized here and on Figure 27. INAA of steatite artefacts indicates that 96 Nath et al. 2014: 74‐100 Figure 27: Rakhigarhi stone and metal sources and acquisition networks identified in this study. Potential, but as of yet unconfirmed, copper, gold and chert source areas are also indicated. 97 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 residents used raw material derived from deposits in northern Pakistan. INAA of agate‐carnelian nodule fragments and manufacturing debris confirms that this variety of stone was being transported to the site from sources in Gujarat. Lead isotope assays of lead and silver objects suggest that these metals were obtained from deposits in southern Balochistan. Visual examinations of grindingstones have revealed that while multiple source areas were being exploited, the large majority of these objects are composed of rock occurring in southern Haryana. It is important to note that for each of these materials, save steatite, there are indications that other, presently unknown sources were also being exploited. The data generated from the analyses conducted thus far, although limited, clearly show that residents of Rakhigarhi, like Harappans dwelling at other Indus cities, were participating in extensive inter‐regional stone and metal acquisition networks. These networks are almost certain to become even more diverse as studies of the assemblage continue. One category of artifact found at the site that has not yet been examined is tools and debris composed of chert. When samples from such objects are compared to the INAA database of South Asian chert sources (Law 2011: Chapter 6), it is highly probable that a link from Rakhigarhi to the chert quarries in the Rohri Hills of Sindh will be confirmed. Similarly (and furthermore), it is our feeling that copper from the northern Aravalli Range (i.e., the Khetri copper belt) and alluvial gold from rivers debauching the Himalaya foothills were probably traded via ancient Haryana to Harappan consumers across the Indus Civilization. If this was indeed the case, then it is very likely that these important metals first came to Rakhigarhi before being transported onwards. Geologic provenience studies of Rakhigarhi’s stone and metal artifact assemblage are ongoing or in the planning stages. Larger numbers of agate‐carnelian artefacts and unfired steatite artefacts are to be analyzed in order to confirm the acquisition patterns detected for these materials. A substantial set of chert artefacts is likewise to be subjected to INAA in order to determine if, like chert at other Harappan cities, a significant portion was being imported to the site from the Rohri Hills. The EDTA lead isotope sampling technique employed in this study is completely non‐destructive and, in addition to lead and silver, can also be used to extract lead samples from copper artefacts. In 2013 most of Rakhigarhi’s copper artefacts were sampled in this way and are being analyzed as this paper goes to press. The results will permit us to confirm or refute what has long been suspected – that the northern Aravalli Range was a major source area for Harappan copper. The study of the grindingstone assemblage will also continue until all those recovered have been examined. Doing so will permit us to fully examine changes in grindingstone acquisition patterns over time. Lastly, a complete and detailed program of mineralogical identification and inventory has been initiated. Rakhigarhi is one of the most important and unique Indus Civilization sites and a comprehensive examination of its stone and metal artifact assemblage is certain to shed new light on the issues of Harappan raw material acquisition and trade. 98 Nath et al. 2014: 74‐100 Notes In a letter (F. No. 09/09/05 ‐ RGR) dated 30 April 2007, Dr. AmarendraNath (Director, Archaeological Survey of India, retired) informed Randall Law that then Director‐General of the ASI, Ms. AnshuVaish, had given permission to Dr. Law to carry out a detailed study of the stone and metal artefacts recovered during excavations at the Indus Civilization city of Rakhigarhi, District Hissar, Haryana. The preliminary examination of artefacts housed in the Rakhigarhi section of Purana Qila, New Delhi took place over one week in late June of 2008. During this brief period, multiple types of stone and metal objects were studied and photographed. All data (on excel spreadsheets) and photographs generated were placed on the section’s main computer and a back‐up on DVD was made for the Rakhigarhi archives. Select lead and silver artefacts were briefly washed in a solution of ultra‐pure water and 0.05% EDTA. The solution was retained for lead isotope analysis. A small set of non‐diagnostic raw material debris fragments was set aside for further, more detailed characterization and/or geologic provenience studies. In October of 2008, Dr. R.S. Fonia (then the Director of Exploration and Excavation, ASI) brought this small set of samples to the University of Wisconsin‐Madison, USA for analysis. 1 References Fentress, Marcia A. 1976. Resource Access, Exchange Systems and Regional Interaction in the Indus Valley: An Investigation of Archaeological Variability at Harappa and Moenjodaro. Ph.D. Thesis. Oriental Studies, University of Pennsylvania. Garge, Tejas. 2011. Settlement Pattern of the Harappan Culture in Chautang Basin.Unpublished PhD Thesis. Pune: Deccan College. Lahiri, Nayanjot. 1992. The Archaeology of Indian Trade Routes Up to c. 200 BC: Resource Use, Resource Access and Lines of Communication. Oxford University Press, Delhi. Law, Randall W. 2002. Potential Steatite Sources for the Indus Civilization. In Indus Valley Civilization: Collection of Papers Presented in the International Colloquium on Indus Valley Civilization at Islamabad (6th ‐ 8th April 2001), edited by M. A. Halim, pp. 158‐69. Ministry of Minorities, Culture, Sports, Tourism and Youth Affairs, Government of Pakistan, Islamabad. Law, Randall W. 2011.Inter‐Regional Interaction and Urbanism in the Ancient Indus Valley: A Geologic Provenience Study of Harappaʹs Rock and Mineral Assemblage. Current Studies on the Indus Civilization, Volume VIII, Parts 1 and 2. Manohar, New Delhi. Law, Randall W. and James H. Burton. 2006. Non‐Destructive* Pb Isotope Analysis of Harappan Galena Fragments Using Ethylenediaminetetraacetic Acid and ICP‐MS. (*Practically). In Proceedings of the 34th International Symposium on Archaeometry, Zaragoza, 3‐7 May 2004, edited by J. Pérez‐Arantegui, pp. 181‐ 185. Institución Fernando el Católico, Zaragoza. Law, Randall W. and James H. Burton. 2008. Non‐Destructive Pb Isotope Sampling and Analysis of Archaeological Silver Using EDTA and ICP‐MS. American Laboratory News 40(17):14‐15. Nath, Amarendra. 1998. Rakhigarhi: A Harappan Metropolises in the Saraswati‐ Drishdwati Divide. Puratattva 28: 39‐45. 99 ISSN 2347 – 5463Heritage: Journal of Multidisciplinary Studies in Archaeology 2: 2014 Nath, Amarendra. 1999. Further Excavations at Rakhigarhi. Puratattva 29: 46‐49. Nath, Amarendra. 2001. Rakhigarhi: 1999‐2000. Puratattva 31:43‐46. Odess, Daniel. 1998. The Archaeology of Interaction: Views from Artifact Style and Material Exchange in Dorset Society. American Antiquity 63(3):417‐435. Possehl, Gregory L. 1992. The Harappan Cultural Mosaic: Ecology Revisited. South Asian Archaeology 1989. C. Jarrige. Madison, WI, Prehistory Press: 237‐244. Prabhakar, V.N., T. Garge and R.W. Law. 2010. Mitathal: New Observations based onSurface Reconnaissance and Geologic Provenance Studies. Man and Environment 35:54‐61. Ratnagar, Shereen. 2004. Trading Encounters from the Euphrates to the Indus. Delhi: Oxford University Press. 100