Index Introduction Background Geological Context Burial Description Craniometric Analysis and Population Affinity Discussion Conclusion Acknowledgments References

Introduction

The practices of India’s adivasi or ‘tribal’ populations have been the subject of ethnographic analyses since the time of British colonization (Postans, 1848; Shortt, 1865; Sinha, 1958; Fuchs, 1977; Singh, 1994; Ratnagar, 2003). More recently, several genetic studies have focused on the relationship of the adivasi to existing caste populations and hypothesized relationships to the original Late Pleistocene inhabitants of South Asia (Cordaux et al., 2003; Kivisild et al., 2003; Chaubey et al., 2007; Reich et al., 2009). Virtually no archeological sites directly related to tribal communities are known, due primarily to their often semi-nomadic and somewhat remote habitation patterns as well as their dwindling numbers (Ratnagar, 2003).

In this study, we report on a rare burial found along the lower Narmada River in Rampura, Gujarat, western India. Both morphometric and cultural analyses suggest that this was a woman of tribal affinity. We detail the geological context of the burial, its characteristics, and the condition of the skeleton. We also provide a paleopathological analysis and assess whether cause of death can be determined. Based on these data, we interpret the specimen in the context of known burial practices of local adivasis. The results provide rare first-hand osteoarcheological evidence of these practices among the modern tribal populations of western India.

Background

The term ‘tribal’ refers to a level of Indian social organization that is separate from the stratified class-based society that characterizes the majority organized-religious communities such as the Hindus, Muslims, and Christians. These groups are often spoken of as the indigenous inhabitants of India, despite having a varied and often unclear or unknown population history (Guha, 1999). The term ‘Scheduled Tribe’ is the sanctioned English term used to refer to these communities by the Indian government, with the term adivasi (translated as ‘original inhabitants’) also used in formal documents. Both words have potentially negative connotations. Referring to someone as a ‘tribal’ is conventionally perceived as pejorative today, a throwback to British colonial rule. The word adivasi was coined in the 1930s from the words adi meaning ‘beginning’ or ‘early time’ and vasi meaning ‘resident of.’ It was used by the members of tribal communities to refer to themselves in contradistinction to all outsiders, and was appropriated in the 1930s by Indian political activists as they opposed British rule (Hardiman, 1987).

Adivasi carries the connotation of the original inhabitants of India whose land was taken over by outsiders and has been translated as ‘aboriginal,’ thereby suggesting that tribals are the indigenous inhabitants of the subcontinent. However, this assumption is unfounded—as Ratnagar (2003: 19) argues, for example, “In no way can we prove … that the Mahars are the aboriginal settlers of Maharashtra.” Some caste community Indians also take exception to the notion that they are not similarly indigenous owners of the land. To avoid the loaded implications of the term adivasi, the government has recently also started referring to them as vanavasi, which translates as ‘people of the forest.’

For the purposes of this paper we will use the terms ‘tribal’ and adivasi to refer to the community to which we believe the specimen to belong. It is consistent both with the official language of the Indian government as well as the background literature on which this study is based, although it does not automatically imply an autochthonous role for these communities to the exclusion of members of the organized caste society.

Geological Context

In 2006 during a survey of the lower Narmada valley, the acetabulum of a human pelvis was found eroding out of river terrace at the village of Rampura (22°06′N, 73°31′E) in western Gujarat. The locality of the site is on the westernmost extent of the Narmada River (Figure 1). The Narmada River, the largest river of peninsular India, flows in an ENE–WSW direction for about 1312 km across central and western India before meeting the Gulf of Cambay in Gujarat. The alluvial plain sediment in which the specimen was found is comprised of paleosols which were dated in nearby river basins from 125 to 40 ka ago (Chamyal et al., 2003). Early Holocene uplift in the region (Maurya et al., 2000) has resulted in the exposure of Late Pleistocene sediments which form the cliff section all along the river basin.

View Details

Figure 1. Map of India and Gujarat showing burial locality.

The specimen was found in Quaternary fluvial deposits at a 16 m high cliff section at Rampura, within an incised cliff. It was situated about 9 m from the present-day river level. The deposits of the section can be broadly divided into a lower half of various gravelly horizons and an upper silty sand horizon. The clasts of the gravel bed consist of basalt derived from the Deccan Trappean upland. They are poorly sorted and the matrix is mainly coarse to medium sand. The horizon in which the burial was found is characterized by a regular cyclic sequence of sand and medium-sized gravel.

The burial was found in a horizon that had been previously dated to between 40 and 70 ka, providing a maximum age range for the specimen. Two samples from the skeleton were dated using accelerated mass spectrometry (AMS) dates by Stafford Research, Inc. in Lafayette, Colorado, USA: a section of the upper M3 and a portion of the right distal radius. These confirm an age of 165 ± 15 radiocarbon years, or between 324 and 60 years ago. In other words, the individual died some time between 1685 and 1950, with a mean year of death at 1817. An ESR date was also performed by Rainer Grün at Australia National University on the lower M3, with dosage rates calculated from soil samples taken from horizons adjacent to the specimen. This method returned an age of younger than 500 years BP (before present).

Burial Description

The specimen consists of a single individual whose head was positioned face up, and pointing due north (Figure 2). Since the Narmada River at this locality runs northeast–southwest, the body was not aligned with the flow of the water. Given the fully articulated and complete state of preservation of the skeleton (including a preserved hyoid bone), its provenance in poorly sorted matrix comprised of medium-sized gravels, the position of the individual (prone on its back with hands resting on pelvis), and its orientation due north, we consider this an intentional burial. There were no cultural remains associated with the specimen, which as we will discuss later is characteristic of some tribal community burials. Overall, intentional human interments along rivers are rare in contemporary India for cultural reasons: the vast majority of Hindus, who comprise 80% of India’s population, cremate their dead; Muslims bury their dead facing Mecca which would have been almost due west of our position; and many, but not all tribals or adivasis in this region, also cremate their dead (Bendann, 1969; Goswami, 2009). Hence we initially considered the possibility that this burial was prehistoric or that of an outsider.

View Details

Figure 2. Burial in situ.

The individual was discovered with its left side exposed due to riverbank erosion. As a result, recent postdepositional processes had caused the loss of the left femur, patella, tibia, and fibula, and of portions of the right femur, radius, and ulna. The entire set of right metacarpals and carpals was also missing prior to discovery. Damage to the specimen by curious onlookers after its discovery but before its extraction resulted in the partial destruction of the sacrum, lumbar vertebrae, and parts of the pelvis.

The specimen consists of a complete skull with some postdepositional distortion whereby the occipital portion of the skull has been pushed anteriorly, making the maximum cranial length appear shorter than the natural length (Figure 3). The coronal, parietal, and lambdoid sutures are still patent, and there is an extrasutural bone near lambda. The dentition is entirely intact, and the third molars are not quite fully erupted to occlusion. Full craniometric and odontometric measurements are given in Table 1. A partial list of measurements for the postcrania are reported in Table 2.

View Details

Figure 3. Rampura skull lateral/frontal views.

Based on the near complete eruption of the third molars and the unfused basilar suture, the individual is estimated to be between the ages of 15 and 21, following the combined standards of Bass (2005) and Buikstra and Ubelaker (1994). The pelvic remains indicate that the individual is a female following Steele and Bramblett (1988), and White and Folkens (1991). Her stature is 155 cm based on a measurement of the articulated skeleton in situ.

Due to poor security at the site, the excavation of the burial was performed as a salvage operation in one 17-hour day. Therefore, some information was inevitably lost. Priority was given to (a) recording the geological context of the specimen, and (b) careful removal of each bone. No further work was possible in terms of identifying the outline of a burial pit, or surveying the area for other burials. Recording was achieved primarily through photographs and drawings.

The specimen’s anterior upper dentition (incisors and canines) and all but the molars in the lower dentition exhibit a pink discoloration (see Figure 3). This is internal to the tooth structure and not on the enamel, so is not due to diet. Previous research into the cause of pink teeth (Beeley and Harvey, 1973; van Wyk, 1988a, 1989; Watts and Addy, 2001) have isolated the source of this discoloration to some kind of hemoglobin which has diffused into the dentin (Kirkham et al., 1977; van Wyk, 1988b; Fish, 2005). While pink teeth are not correlated exclusively with one cause of death, it is most commonly found in conjunction with death by drowning (Clark and Law, 1984; Brøndum and Simonsen, 1987; van Wyk, 1987; Pollanen, 1998). In death by asphyxiation or drowning, there is an accumulation of blood in the tooth pulp due to either a rapid increase in venous pressure to the head, or a pooling of blood in the head and neck region due to the victim lying in a head-down position after death (Clark and Law, 1984; Borrman et al., 1994). Studies also show that the discoloration is often confined to the incisors, canines, and premolars (Kirkham et al., 1977), as was found here.

This discoloration can also occur when the body decomposes in a humid or damp environment (van Wyk, 1987; Borrman et al., 1994; Pollanen, 1998; Fish, 2005; Campobasso et al., 2006). Van Wyk’s studies into this phenomenon concluded that there is no statistical correlation between the cause of death, period after death, and the occurrence of pink teeth, and that the discoloration exists as part of a spectrum of staining patterns of both the pulp and dentine (van Wyk, 1987, 1988a, b, 1989). An investigation into the presence of pink teeth in an archeological context suggests that they are not caused by the same mechanism as forensic cases (Dye et al., 1995). These authors report the discoloration of teeth among individuals from a medieval cemetery in Chichester, England. The authors contend that the pink coloration of the teeth in their archeological cases is the result of a structural and chemical alteration of dentin which occurred while the skeletons were buried (Dye et al., 1995). However, no further analyses into the condition of pink teeth in archeological cases has been reported. Given this previous research, it is difficult to determine if the pink teeth of the Rampura specimen presented here are related to burial conditions or to the cause of death, but there is data to suggest that this condition could be the result of death by drowning, which is the circumstance under which it most commonly occurs.

Craniometric Analysis and Population Affinity

Comparative morphometric analyses of the Rampura individual were undertaken in order to assess its population affinity. As stated earlier, because of the rarity of intentional burials in this area, we considered the possibility that this was an outsider. Potential foreigners who have been known to traverse through this region include British colonialists, Portugese traders, and members of the Siddi community from East Africa. We therefore kept these groups in mind as we proceeded with the analysis. The craniometric measurements of the Rampura specimen were compared to those from a sample of individuals from the subcontinent, collected from published literature, as well as to a global sample taken from Howells (1996) and Brown (1997).

Several different configurations of the comparative data were analyzed in order to maximize the repeatability of the results. In all analyses, craniometric measurements which had been taken on the Rampura specimen following the protocol in Howells (1973, 1989) were used. The data were reduced using principal components analysis and analyzed using discriminant function analysis (DFA), the purpose of which is to assign an unknown individual relative to a known set of comparative groups (Rao, 1948). Missing data among the comparative sample were imputed using the expectation-maximization and data augmentation methods, with the software program NORM 2.3 (Schafer, 1999), a method that has precedence in the study of morphometric skeletal and dental variation (Strauss et al., 2003; Stefan, 2004; Scherer, 2007). No missing data were imputed for the Rampura specimen. Since the size of individuals is a meaningful and diagnostic tool in assessing population affinity among modern humans, the data were not adjusted for size.

In the first set of analyses, the Rampura specimen was compared to 209 individuals from several different populations using 13 variables primarily from the cranial vault and base (Table 3). Just under half of the cases (n = 92) were from archeological sites in India, Pakistan, and Sri Lanka dating from the Neolithic to the post-Harappan period. Thirty-one females from the Andaman Islands were included, taken from Howells’ data set. Because the Rampura individual dates to the period of British occupation, a sample of females from the Poundbury and Spitalfields collections were also included. Finally, craniometric data on individuals of central/south Indian tribal affinity were taken from the published literature (Turner, 1898–1899, 1900). These included Gonds, Veddahs, and Mundas. It was not possible to include some tribal populations from Gujarat to which the Rampura individual is believed to belong due to unavailability of cranial data. In these analyses, the Rampura specimen was given a 99% posterior probability and a 76.7% typicality probability of grouping with the Andaman Islanders (Table 6).

In the second set of analyses, a larger set of 29 variables was used to compare the specimen to a global sample from Howells’ data set (n = 436). Only female individuals were included in this analysis. Eight populations were represented, as seen in Table 4. In this stage of the analysis the Rampura specimen did not group with the Andaman Islanders, grouping with both the Egyptian sample (52% posterior probability) and the Atayal sample, a group of indigenous people from Taiwan Island (36% posterior probability). Finally, the analyses were conducted removing the measurements that included the occipital (GOL, OCC, OCA) due to the distortion that is present in this part of the skull. The results aligned the Rampura individual with the Andaman Islanders with a higher posterior probability than any other group, although the results were weaker (48.0%).

In the third analysis, a greater number of facial variables were used and the number of specimens from South Asia was increased, while the global sample was somewhat reduced. The final configuration was a comparative sample of 260 with individuals from six Eurasian and African populations as well as three separate Indian tribal groups (Table 5). The Rampura individual was given a 60% posterior and < 1.0% typicality probability of grouping with the Atayal sample of indigenous Taiwanese when the data were size-adjusted.

Discussion

Taken together, the morphometric analyses indicate that the Rampura individual possesses a mixed African/Asian morphology, grouping with a global sample of tropically adapted and/or hunter-gatherer populations. This is consistent with previous assessments of both archaic Indian as well as modern tribal morphology, including that of Andaman Islanders (Howells, 1973, 1989; Mukerjee, 2003; Athreya, 2010), who are often classified as sharing morphology with the ‘Negrito’ populations of southeast Asia (Hanihara, 1993; Lahr, 1995). However, based on the craniometric analysis alone it is impossible to attribute the Rampura individual to a specific tribal population from Gujarat, or from the mainland subcontinent in general. Given the remote settlement patterns of Andaman Islanders and their active avoidance of contact with outside groups since colonial times, particularly with mainland India (Cooper, 2002; Mukerjee, 2003; Thangaraj et al., 2003; Venkateswar, 2004; Dhingra, 2005), there are no known historical scenarios to support the notion that this specimen is actually of Andaman origin.

The individual’s stature and overall size indicate an adaptation to a hot, humid environment (Hiernaux et al., 1975; Perry and Dominy, 2009), thereby making it unlikely that she is from an East African (such as Siddi) or European trading community. Therefore, the present interpretation of the results based on analyses of the craniofacial morphology is that this individual derives from a tribal South Asian group. However, exactly which community she belongs to remains unresolved. Further information may be obtained by a detailed study of the metric and non-metric dental traits, which have the potential to determine more accurately the Rampura specimen’s affinity (Lukacs, 1984; Lukacs and Hemphill, 1993).

An analysis of the burial itself sheds a bit more light onto the specimen’s population affinity. As stated earlier, there are a handful of organized religious communities in India who habitually bury their dead: Christians, Jews, and Muslims inter their dead in cemeteries. Most Hindus, Jains, and Sikhs cremate their dead. Among tribals, cremation is now common and has been for at least 100 years, although there are several notable exceptions, including groups who will bury their dead under specific circumstances (Crooke, 1899; Bendann, 1969; Goswami, 2009). Several of these populations are present in Gujarat today, or populated parts of Gujarat at some point in the last 300 years. These include the Raikaris, who bury individuals who died by drowning or who died suddenly, as well as the Varlis, who perform burials under similar situations (Crooke, 1899). Although there is no way to exclusively conclude that the Rampura individual died by drowning, the presence of pink teeth points to that as one possibility.

The Gonds, Kurmi, Khasis, and certain lower-caste Hindus will bury their dead if the individual is an adult who expired due to an infection or contagious disease, but also if the family is poor and can not afford firewood for a cremation (Bendann, 1969). In many of these cases, the individual is buried with their head facing downwards. Among the Bhils, the dead are usually buried face up with their heads pointing towards the south (Bahadur, 1977).

The population whose burial practices bear the greatest number of similarities to the conditions of the Rampura specimen is the Mavchi of western India. The Mavchis occupy the region between Maharashtra and Gujarat, particularly in the Dangs district of Gujarat. Among the Mavchis, the burial ground is usually a distant place away from general habitation—either on the edge of a forest or on a river bank (Goswami, 2009). In their death rituals, the dead body is laid in the ground with the head facing north. If it is a female, she is typically given a piece of gold in her mouth but all other precious items are taken away from body before death. The only things that may be left are a few personal items, including a necklace of stone (Goswami, 2009). A perfectly round flat stone that was centered between the Rampura individual’s upper and lower incisors could have been the result of the gravelly matrix in which the individual was found, but appeared to be deliberately placed.

Conclusion

Based on preliminary analyses of the morphological and cultural data of the Rampura burial, we believe this individual represents a rare case of a burial of a member of India’s tribal community, most likely a Mavchi. Further postcranial skeletal and dental analyses may confirm precisely from which mainland tribal population she derives. Given the scarcity of modern tribal burials, this case provides a rare piece of evidence on the death rituals of these communities in the South Asian bioarcheological record.

Acknowledgments

A special thanks to Dr L.S. Chamyal at MS University of Baroda for his extraordinary expertise in the field as well as for logistical and moral support. Thanks also to Dr. Deepak Maurya, Dr. Alpa Sridhar, Sanjay Makwana, and Dr. Trudy Turner. This research was supported by the National Geographic Society’s Expeditions Council, the National Science Foundation (Grant BCS-0645368), and the College of Liberal Arts at Texas A&M University.