Warfare, Drought, and Agriculture: Coping with Conflict and Food Insecurity

Much of the research on ancient warfare has focused primarily on identifying the causes of ancient hostilities. While researchers are careful to discern between ultimate and proximate causes, it is often the case that environmental causes, such as prolonged droughts, are credited with heralding the onset of inter-group violence across the ancient world. Less examined are the effects that such conditions have on the practices of everyday life, such as how foragers and farmers reconstitute their subsistence practices in the wake of water deficit and social conflict, or how gender and other social roles are renegotiated in a contentious climate mediated by conflict.

Archaeological research in the Late Prehistoric Eastern Woodlands of North America has revealed situations where the intensity of inter-group violence resulted in the abandonment of homesteads, villages, and even entire regions. Palisade walls, catastrophically burned villages, and violence-related skeletal trauma are the most obvious archaeological indicators of these ancient hostilities.  And although such evidence informs us about the scale and intensity of these conflicts, it tells us little about what it was like for ancient people to live with war on a daily basis. Did chronic warfare compromise or constrain subsistence pursuits and other basic practices related to the social and economic reproduction of households and communities? How did farmer-foragers alter food collection and production strategies to contend with warfare-associated risks?

Our research focuses on the impacts of chronic and intensive warfare on the subsistence pursuits of 13th century Mississippian groups in the Central Illinois River Valley (CIRV) of west-central Illinois, located in the midwestern portion of the USA. In addition to chronic violence, villagers living in this region and at this time were contending with a long period of drought, which undoubtedly contributed to the scale and intensity of warfare. Our investigation of these issues examines the decisions that farmer-foragers made in order to limit their exposure to violence while attempting to produce and procure sufficient foods in a water-deficient environment. We examine how people altered their daily, seasonal, and annual foraging and farming practices to sustain their communities and their very lives. In doing so, we focus on three lines of evidence to address the effects of conflict and drought on local farming communities—ancient plant data, faunal remains, and Carbon and Nitrogen isotopic data extracted from maize kernels. These lines of evidence are used to address three primary issues that we interpret as responses to chronic violence and drought: diachronic changes in diet, seasonal/habitat changes in food production/procurement activities, and changes in agricultural strategies.

Background

The late Prehistoric Southeastern and Midwestern United States witnessed the pan-regional development of a cultural tradition known as Mississippian, beginning around AD 1000. Mississippian societies are generally characterized by their regional political hierarchies, earthen platform mounds, and the interregional exchange of symbolic and ritual paraphernalia; most had agricultural economies and hereditary leadership positions of some form.  However, Mississippian societies varied considerably in terms of geographic scale and political complexity. Some large polities like Cahokia in southwestern Illinois were comprised of multiple large settlements inhabited by thousands of people. Most other Mississippian groups such as those in the Central Illinois River Valley (CIRV) were smaller in scale and less complexly organized than Cahokia with towns and villages inhabited by hundreds not thousands of people.

In general, the Mississippian period is credited with heralding the transition from foraging and small-scale food production to intensive maize agriculture, although there is a great deal of regional variation throughout the broader region in terms of the relative importance of maize and the degree of cultivation intensity. Regardless of the relative importance of plant cultivation, Mississippian people foraged extensively for myriad nuts, fruits, and fresh greens. Foraged plants ran the t from primary/secondary staples (e.g., hickory nuts) to supplementary resources (e.g., fresh seasonal greens). A diverse range of wild edible plants would have been necessary to provide nutritional balance to a diet based on maize, which is deficient in niacin and essential amino acids lysine and isoleucine. Protein was procured through hunting and fishing, activities that were important throughout the prehistory of the region as there were no native animal domesticates except for the dog.  While the relative proportion of terrestrial versus aquatic fauna (and the species targeted) varied by region and in relation to settlement location, the most common animal prey included white-tailed deer, wild turkey, migratory waterfowl, turtles, and various species of fish.

The Mississippian period occupation of the CIRV dates from around AD 1100 and ends with the abandonment of the region around AD 1450. Early Mississippian settlements in the region consisted primarily of small and widely dispersed homesteads. Different localized settlements appear to have been ceremonially and politically associated with a number of small temple and mortuary complexes. By AD 1200 this early Mississippian settlement pattern of dispersed communities was replaced by fortified and nucleated towns linked with smaller outlying settlements. Most towns were built on defensible bluff-edge locations on the western side of the Illinois River. These bluffs generally range from 38 to 45 meters in height, and would have allowed a broad expanse of the floodplains to be visually monitored, providing views that would have hindered attempts by enemy forces to launch surprise offensives. Despite the high risks associated with attacking fortified villages situated in these locales, several such villages met their ends as a result of enemy attacks in which they were set ablaze either during large-scale direct assaults or after sudden, strategic abandonments.

Both warfare and drought appear to have been particularly intense in the CIRV in comparison to many other portions of the late Prehistoric Eastern Woodlands. In terms of violence, analyses of the human skeletal remains have revealed a higher adult rate of violence-related skeletal trauma than any other Mississippian site in the entire Eastern Woodlands. Violence-related mortuary patterns include a mass grave with 15 plus interments and numerous other individuals with embedded arrow points, scalp marks, and blunt force cranial trauma. By AD 1300 much of the regional populace had relocated to fortified towns, indicating that it was no longer safe to live outside the protective boundaries of palisade walls. In addition to the risks associated with warfare, recent models of rainfall patterns based on the Palmer Drought Severity Index (PDSI) reveal that this region also suffered a prolonged period of drought between AD 1100 and 1245, bracketed by wet periods before and after this extreme water deficit. Thus, between AD 1200 and 1250, villagers had to contend with both: (1) the risks of violent attack and death when leaving their palisaded villages to hunt, fish, and forage; and (2) deteriorating growing conditions for their maize crops, which they likely planted directly outside the palisade where the farmers, which were women, could be protected by archers stationed on the walls. This focus on infields situated close to the village would leave little space for crop rotation, with negative consequences for soil fertility. After AD 1250, drought conditions ended but regional violence continued to plague the populace.

We assess how people dealt with problems of warfare and drought through a diachronic analysis of food remains at four consecutively occupied sites which span both the onset of regional hostilities and the cessation of drought conditions.  The earliest site, Lamb, is a small farmstead that dates to AD 1100-1150 and marks the beginning of the Mississippian occupation in the region. The Lamb occupation was settled around the onset of drought conditions but it predates the outbreak of regional hostilities. Cooper, a small village, was occupied between AD 1200 and 1235, a time of continued drought, and brackets the earliest outbreak of regional hostilities. The palisaded Orendorf village was settled around AD 1200 until the incineration of its final occupation around AD 1250, the same time when wet conditions returned to the valley. The final site, Myer-Dickson, was occupied from AD 1250 to 1300, and represents a community in flux—residents were still faced with the same hostile environment, but drought had ceased to be an issue.

Diachronic Changes in Subsistence

Analysis of the plant remains from flotation samples from these four sites reveal a clear shift away from wild, foraged plants.  There are clear, statistical declines in nuts (a staple food that rivals maize in its abundance), fruits, wild seeds, and fresh greens. The plant diet also became less diverse in general, with maize as the primary foodstuff. This finding suggests that plant foragers, which were women, significantly curtailed food collection excursions that would have taken them away from the protected limits of their villages. However, maize abundance did not increase at all during the sequence, suggesting that villagers were unable to offset the decline in wild plant foods by growing more maize.

Analysis of the animal bones reveal a sharp decline in fish with the onset of warfare, accompanied by an increase in mammals in the diet. Metric analysis of fish vertebrae indicate that people stopped procuring fish of variable sizes and instead emphasized larger-bodied fish, suggesting a decline in net-based fishing techniques. With respect to mammals, there was a clear shift towards larger prey (deer and elk) over smaller prey that can be caught in traps and snares. It bears noting that smaller fish and mammals tend to be taken with nets and traps/snares, respectively, which are often checked and maintained by women and children as they forage for plant foods. Thus, this shift away from fish and small mammals and towards large prey suggest that women and children’s protein foraging activities became more limited in favor of cooperative male hunts. Although larger deer and elk became the primary meats, there was nevertheless a decrease in the size of their bone fragments; this pattern indicates more intensive processing for grease and marrow, which is often associated with food shortages and/or nutrient deficiencies.

Diachronic Changes in Seasonality and Habitat

Analysis of habitat preferences of the wild plants reveals a clear shift away from wetlands exploitation, with a primary focus on woodlands (nuts) and disturbance plants. A seasonal assessment of wild plant collection also demonstrates a shift away from spring and early summer foraging towards a concentration on the late summer and fall seasons—foragers focused primarily on nuts, with incidental collection of other plants in the context of mast exploitation. This seasonal shift towards foraging primarily in the fall is bolstered by the faunal data. Deer and elk hunting season occurs in the fall months, and the faunal assemblage becomes heavily skewed towards these prey. The change in fishing strategies also supports a seasonal change towards fall-based animal foraging. We see a clear increase in large-to-small fish over the sequence. During most of the year, these larger fish species live in deep lake or river waters and are difficult to capture. Lakes and rivers flood annually in the spring and summer, creating new streams and connecting the rivers, lakes, and sloughs. Following floodwater recession in the fall, bottomland lakes and sloughs are cut off from the main river channels, leaving shallow pockets in which large fishes get trapped and are easily captured by humans. Thus, although regional residents deemphasized fishing as a primary subsistence pursuit, they narrowed their focus towards large species that could be quickly exploited during a narrow seasonal window in the fall. This shift away from spring and summer foraging suggests that the regional inhabitants oriented their spring and summer months around planting and protecting their maize crops.

Changes in the Agricultural regime

Given the indicators of dietary insufficiency (decline in abundance and diversity of wild plants, increased bone fragmentation) that followed the onset of warfare in the region, one might expect an increase in agricultural yields. But this is not the case, despite multiple measure of abundance. We thus assess changes in agricultural strategies using metric data on hundreds of maize kernels and Carbon/Nitrogen isotopes on a sample of 15-20 kernels from each site. Metric data reveal that, by AD 1250, maize kernels were smaller and more variable in size, suggesting that Late Mississippian farmers planted denser stands than previously, which leads to more soil nutrient loss. Not surprisingly, the combination of denser stands (signaled by more size variability) and declining soil fertility (signaled by smaller kernels) did not increase yields. While there was plenty of rich upland soil to permit extensive shifting cultivation, by AD 1250, the risks of attack while overseeing scattered plots outweighed the risks of possible food shortage and crop loss, leading farmers to intensify cultivation in infield plots for longer periods of time.

Luckily, the drought conditions that had plagued the region for 150 years had finally come to an end by AD 1250. Carbon and Nitrogen isotopes on maize kernels support the Palmer Drought Severity Index model for the timing of regional drought conditions. Together, Carbon and Nitrogen isotopes track water uptake and water use efficiency in crops, a technique that until now has only been employed in Old World case studies. In addition, Nitrogen can also provide information on relative soil fertility. Maize kernels dating prior to AD 1250 have higher Carbon and Nitrogen values than the kernels post-dating AD 1250; the decline in both Carbon and Nitrogen isotope values in the post-AD 1250 maize kernels signal a clear increase in water uptake into the maize plants. The Nitrogen decrease is also suggestive of poor soil fertility.

While the drought may have ended by AD 1250, farmers living at the Myer-Dickson site still had to contend with deteriorating soil conditions from over-cropping the same plots, which is signaled by the smaller kernels with greater size variability that they were planting. But these Myer-Dickson farmers were canny and resilient; they incorporated a completely new variety of maize into their cropping system, in addition to common beans. This smaller kerneled-maize (smaller even than the dominant variety which had shrunk over the last 150 years) could germinate faster than the dominant maize variety, producing a crop earlier in the season. Moreover, the introduction of beans would have improved the nitrogen economy of the soil, thus maintaining maize yields despite deteriorating soil conditions.

Conclusions

We interpret this combination of patterns as reflecting an intensification of peoples’ concerns over safety as the Central Valley became a more uncertain, dry, and violent place. The correlation of regional drought conditions and the subsequent onset of regional hostilities is undoubtedly linked. Shifts in the timing and location of subsistence pursuits through time suggest that people narrowed foraging forays to target a suite of resources in a manner consistent with large group excursions. The exploitation of supplemental foods became increasingly embedded within higher-yielding subsistence pursuits, such as farming, mast and fish collection, and deer/elk hunting. Evidence of increased fragmentation and decreased identifiability of large mammal bones suggest that the trade-off between deer and fish may have not have been a perfect one, resulting in a lower overall protein yield for villagers.

Likewise, the trade-off between wild plant foods and farmed produce was not a perfect one either. While there was plenty of rich upland soil to permit extensive shifting cultivation, by the Late Mississippian period, the risks of attack while overseeing scattered plots outweighed the risks of possible food shortage and crop loss, leading farmers to intensify cultivation in infield plots for longer periods of time. Moreover, the impact of drought on the farming cycle is clear in the reduction of maize kernel size and the decline in soil fertility. Nevertheless, farmers adapted to these challenging conditions by introducing a new maize variety with a shorter growing cycle and inter-cropping with nitrogen-fixing legumes. Collectively, these data provide clarity on human decision-making in a drought-plagued peace-time/war-time transition, providing a nuanced look at the people straddling a critical historical period of change. By linking the connected histories of people living in the pre-warfare and warfare periods to high resolution environmental data, we can better illuminate how practices are altered in crisis situations and determine when people reach thresholds at which they are willing to sacrifice both material resources and close-held ideologies to keep themselves and their families safe.