The modern state of Maharashtra is a semi-arid region of the tropics, which is characterized by scanty and extremely variable annual rainfall (average of 500 to 1150 mm). This region frequently experiences ill-timed monsoons, insufficient rainfall, and crop failure. Problems with timing or amount of rainfall are responsible for the majority of food shortage, famine, and farmer suicides in the state today. As rain-fed agriculture in Maharashtra still accounts for 84% of all production, responses to this low volume of rainfall and the high level of uncertainty generally involve diversifying the subsistence base through double cropping, diversifying staple seed banks, diversified planting strategies, and crop rotation. In addition to subsistence farming, a wide variety of other occupations are performed part or full-time by farming communities including herding, stock raising, gathering, cash cropping, and wage labor.

During the Chalcolithic period (2200-700 B.C.), and continuing to the present day, this region was populated by settled agriculturalists, settled stock raising communities, nomadic pastoralists, hunter-gatherers, fishing communities, craftspeople, and laborers. These communities maintain diverse economies in order to maximize scant resources and minimize uncertainty, irregularity, and scarcity (Panja, 1995). During the Early Jorwe phase of the Chalcolithic, paleoclimate records indicate a period of increasingly arid and dry climate culminated around 1500 B.C. Adaptive diversity allowed Chalcolithic peoples to establish a diverse array of communities during this phase. A broad range of settlement types and sizes began to flourish- from large regional centers (Daimabad) to ephemeral herding camps (Walki) (Shinde, 2002). These communities relied on drought resistant barley, stock raising, hunting, lacustrine resources, and wild foods.

Nevasa, Daimabad and Inamgaon were three settled villages occupied during this phase of the Chalcolithic period. Daimabad and Inamgaon represented major population centers for the Godavari and Bhima River Basins (Dhavalikar, 1988). Nevasa represented a smaller, peripheral village site located just across the river from Daimabad (Dhavalikar, 1988). Despite the success of adaptive diversity as a strategy in the Early Jorwe, and the development of these dynamic and growing settlements around the region, after 1000 B.C. the majority of the settlements in this area were abandoned. Only the site of Inamgaon persisted- though the remains from the Late Jorwe (1000-700 B.C.) have been described as degenerate, in comparison with the earlier period. Although the Chalcolithic period has been the subject of considerable research over the last 50 years, questions remain regarding the nature of the decline of these agrarian villages. My dissertation used skeletal material from Daimabad, Nevasa, and Inamgaon to test hypotheses derived from the archaeological record for settlement and subsistence practices across this transition. I looked at the effects of climate change and subsistence transition on population statistics and health for these three sites.

There are two competing hypotheses that have been developed to explain the decline of Chalcolithic villages. The first hypothesis, developed by the excavators of Inamgaon, suggested that increasing aridity and associated climate changes caused a reduction in arable land after 1100 B.C. (Dhavalikar, 1988). This ecological shift precipitated dramatic changes in settlement and subsistence strategies whereby settled agricultural populations dissolved into mobile sheep-and goat pastoralist camps (Dhavalikar, 1988; Thomas, 1988). This hypothesis, hereafter referred to as the Subsistence Transition Model, was based on archaeological evidence for change in material culture (artifact typology), residential structures, and analysis of food refuse deposits after 1000 B.C. at Inamgaon.

However, the nature of the evidence for subsistence and settlement practices in the Late Jorwe period at Inamgaon is somewhat contradictory and this period may not represent a significant departure from the Early Jorwe in terms of climate or subsistence practices (Lukacs et al., 2000, 2001, 2006; Panja, 1995, 1998-99). Based on alternative interpretations of the archaeological record and dental anthropological studies, some people have suggested that Chalcolithic peoples practiced a mixed agro-pastoral economy throughout the occupation of these sites. This Adaptive Diversity Model proposes subtle shifts in species preferences through time, rather than wholesale abandonment of agriculture.

Settlement and subsistence practices are distinct but inter-related cultural adaptations that can potentially buffer human populations from biological effects of environmental change or uncertainty (Larsen 2002). Within an adaptive framework, evidence from paleodemography, long bone growth, and cross-sectional properties can be useful for evaluating the biocultural stress levels of these communities through time.  Below are two abstracts for the 2007 HBA and AAPA meetings in Philadelphia, where I presented my methods and some of the results of this work. Links to these posters are available on my homepage (http://www.appstate.edu/~Robbinsgm). I am currently working on writing up the results of this work for publication.

Abstract for poster presentation for AAPA annual meeting in March 2007

Saving the Babies from the Bathwater: a new Method for estimating Fertility.

G. Robbins.  Department of Anthropology, Appalachian State University.

    This poster provides a new method of estimating Gross Reproductive Rate (GRR), crude birth rate, and life expectancy at birth for samples that are heavily biased toward representation of sub-adults under 5 years of age. Methods for estimating demographic parameters in fertility-centered demography include the Index of Juvenility (Bocquet-Appel and Massett, 1977, 1982: Bocquet-Appel, 2002), the Mean Childhood Mortality Quotient (Jackes, 1986), and other proportional hazards (Buikstra et al., 1986). These methods exclude perinates, infants, and children because they are often poorly represented in skeletal assemblages. There are skeletal assemblages, such as those from the Deccan Chalcolithic in India (3800-2700 B.P.), that are primarily comprised of individuals less than five years of age. These infants and children have a lot to offer paleodemography because of the relatively greater accuracy and precision of age estimates for these individuals.

    Using Weiss (1973) model tables and demographic data from published sources (McCaa 1998, 2002), this poster provides a quadratic equation for estimating fertility (GRR) solely from proportions of individuals less than five years of age. This method provides a statistically significant way of predicting GRR from early childhood skeletal remains (F = 21.38; p = 0.000) producing results not significantly different (adjusted R squared = 0.689; p = 0.000) from estimates provided by other methods for adult samples, such as Bocquet-Appel and Massett (1977). An estimate for GRR and archaeological estimates of settlement growth rate makes these samples more comparable with other populations through estimates of fertility and life expectancy at birth using McCaa (2002).

    This research was supported by grants from Fulbright, American Institute of Indian Studies, and the George Franklin Dales Foundation.

Abstract for HBA annual meeting March 2007

Cortical Bone Ontogeny: Activity, Nutritional Status, and Archaeology

G. Robbins.  Department of Anthropology, Appalachian State University.

    Research on the role of activity in the ontogeny of cortical bone cross-sectional parameters has demonstrated that declining % Cortical Area (%CA) during infancy, once interpreted as evidence for stress, is part of a normal pattern of growth. In modern reference populations, and some archaeological samples like Pecos Pueblo, section modulus (strength) continues to increase through apposition of mass at the periosteal surface despite the appearance of declining %CA. The majority of variation in cross-section parameters is explained by body mass and timing of the acquisition of locomotor skills (crawling and walking, for example). Long bone length is not a significant predictor of %CA in normal ontogeny in the humerus (R² = 0.0244) or the femur (R² = 0.0035).

    This project compared data from the Denver longitudinal study with a sample from Chalcolithic Inamgaon, India (3500-2700 B.P.) to re-evaluate usefulness of cross-sectional parameters for bioarchaeological research. Humeri and femora from Inamgaon (n = 137) demonstrated a significant correlation between low Z-scores for length and reduction in cortical thickness, %CA, and section modulus. Z-scores for length were a significant predictor of %CA (R² = 0.47) and section modulus (R² = 0.66). These samples differed from the reference standard in the velocity of bone loss during infancy. Individuals 6-30 months old with low Z-scores for length also demonstrated greater than expected declines in cortical thickness, %CA, and section modulus, with suppression of periosteal apposition resulting in perinatal parameters persisting to 36 months. The results support the suggestion that nutritional deficiency, small body mass, and low activity levels synergistically and significantly alter the normal pattern of development and cortical bone mass is useful for bioarchaeological research on diet, stress, and biocultural adaptation.

Results from my dissertation research are also available on this site:

Powerpoint presentation of my dissertation defense

Pdf file of the title pages and text of my dissertation (please cite my dissertation if you are using any of my methods or results for your own work).