However, there is no simple relationship between yields and amounts of land used for production—often, yield intensification leads to agricultural expansion and conversion of natural habitats in a phenomenon known as the Jevon’s paradox. Further, alternative systems are not always less productive or profitable. In fact, our results suggest that further investment in research into diversification techniques has the potential to improve productivity of sustainable agricultural methods to equal conventional yields. Encouragingly, the few long-term studies that have been conducted have demonstrated that diversification techniques enhance yields while enhancing ecosystem services, profitability and stability. Leifeld’s last point that organic agriculture will lead to more greenhouse gas emissions suffer from various fallacies that we have already addressed, including that increasing yield is needed to relieve hunger, that alternative to conventional agriculture cannot produce comparable yields and that yield is the sole determinant of the amount of land in production. Further, he fails to take a large-scale, systems perspective in his assertion that organic farms create more greenhouse gas emissions per unit of yield, because he considers only the gas fluxes from soils. When the entire system is considered,grow bucket organic farming systems have on average higher energy efficiency and lower greenhouse gas emissions per farm product than conventional farms.

The lower-energy efficiency in conventional farms is primarily owing to nitrogen fertilizers, which are highly energy intensive to produce. Despite the fact that research into refining organic and diversified farming techniques has been severely underfunded relative to conventional, these alternative forms of agriculture outperform conventional systems in terms of promoting environmental and social sustainability. We believe it is time to invest in rigorous, systems-based research oriented at eliminating yield gaps between organic and conventional agriculture , and elucidating the links between farming techniques, ecosystem services and livelihoods.Models of ancient agricultural strategies typically are built around three variables: population size, caloric needs, and agricultural yields. Although archaeologists generally are satisfied once they have identified fixed numbers that will let them calculate how much agricultural land would have been required around a given site, comparative anthropological and agricultural studies show that changes in acreage comprise only one component of agricultural dynamics. In this paper, I utilize nineteenth century documents and regional archaeobotanical data to construct parameters for the three variables of population size, caloric needs, and agricultural yields for the site of Kaundinyapura, India, a town of the Early Historic period in central India . Archaeological research at Kaundinyapura shows a widespread distribution of trade goods in the Early Historic period indicating that all, or nearly all, households possessed a standard repertoire of domestic-use items . Lacking mineral or other natural resources, Kaundinyapura’s residents probably traded perishable products, but the production of “surplus” for exchange was an extension of the methods used to address basic food needs.In general, crop productivity also is affected by nutrients and pests . These factors are interdependent, since the variety and quantity of rapidly reproducing pests can increase or decrease depending on rainfall, humidity, and other climatic factors .

A complex feedback loop thus affects crop yields apart from deliberate choices about what to plant, since loss of agricultural yields through pests and disease impacts both the active cultivation of crops and the period of storage between harvest and use of the stored grains as food or seed stock. We can now turn to the two other components of the agricultural equation: population size and caloric needs. Although it is difficult to pinpoint the exact size of ancient populations based on archaeological remains, population size can be estimated using a variety of ethnographic and archaeological comparisons that take into account living space, density of individuals per household, and number of estimated 42 ECONOMIC BOTANY [VOL. 60 households per hectare in villages, towns, or cities. Comparative analysis with other Asian cases suggests that Kaundinyapura’s population would have ranged from 350–1,000 individuals, with some fluctuation on an annual or seasonal basis given the region’s monsoon climate . Food requirements, the third component of the agricultural equation, encompass matters of taste and preference in addition to quantifiable aspects such as total caloric value and the presence of proteins, minerals, carbohydrates, and minerals essential to good human health. Calorie requirements for adults vary according to the size and occupation of the individual and there is no single global standard for required caloric intake. While numerous idealistic guidelines exist for nutrition in South Asia, more accurate data are provided by studies of actual farmers, such as that by Edmundson and Edmundson in the Deccan dry-farming region showing a value of 210 kJ per kg of body weight, which they propose as “close to the minimum energy requirement necessary for subsistence farming.” Under the assumption that caloric values for cultigens have not altered, we can apply those values to the crops known from Early Historicsites in the Deccan region .

The proportion of different foods in the diet can be estimated from anthropological observations such as that by Pushpamma et al. showing that grains constituted 75– 85% of the calories and protein in the diet. On the basis of modern ethnographic studies, Dhavalikar and Possehl also estimated a crop mix for ancient populations at the site of Inamgaon consisting of 70% sorghum, 20% wheat, and 10% rice. However, the proportion of the 2006] SMITH: ANCIENT AGRICULTURALISTS AND WILD PLANTS 43 diet provided by one or two grain “staples” may in itself be variable. Other studies of grain consumption in contemporary India indicate that the percentage of different foods in the diet varies by income level and geographic location. Ramnath, Vijayaraghavan, and Swaminathan noted that for the wealthiest income group in their survey of central and south Indian households, the category of grain provided 56 11 % of calories, while the poorestincome group had 75 15 % of calories from cereals. Similar social distinctions in food preference can be traced through the literary record of the premodern period . For comparative purposes, the calculations below are made at three levels of consumption and project the total agricultural output required for staple grains and pulses at 60%, 75%, and 90% of dietary intake. Table 2 identifies the crop area required using three examples of a single-cultigen standard: wheat, sorghum, and gram . High and low yield values are taken from colonial-era documents for central India: wheat ; sorghum ; gram . In addition to the acreage needed to meet the minimum human requirement, additional factors of waste, spoilage, insect damage,dutch bucket for tomatoes and other losses between harvest and consumption must be taken into account . Y. Pomeranz provides an estimate of 30% for these losses on average; lacking more precise data for crop losses in the Early Historic period, the calculations in Table 2 use the 30% figure as a proxy added on to the basic human calorie requirements. The resultant total quantities of staple grains and pulses required for a site the size of Kaundinyapura in the Early Historic period thus take into account two maxima and two minima , as well as variability in the ratio of basic grain and pulse crops. The figures in Table 2 can be translated to best-year and worst-year acreage minima and maxima for Kaundinyapura, which is located adjacent to a river, providing a 300-meter wide strip of alluvial land. At the highest estimated population level of 1,000 inhabitants, and assuming the use of all alluvial land in the vicinity of the site, best-year cultivation of staple crops when calculated at the rate of 75% of the diet would have required a relatively small radius from the site for wheat , sorghum , and gram . However, worst year cultivation of these same crops would have required the exploitation of a considerably larger radius for wheat , sorghum , and gram . The regular zone of cultivation was likely to have been somewhere between the best-year/worst-year land requirements. Long-term corrections probably included labor investment to increase acreage, but it would have been difficult to quickly clear new lands for cultivation in years of shortfall. Instead, agriculturalists would have utilized short-term labor allocations for other strategies, including crop substitution and increased reliance on wild foods.

Crop substitution includes the practice of double-cropping, attested in the archaeological record of central India in the form of winter crops and summer crops . Colonial documents show that by double-cropping, farmers can compensate for short falls in one season by planting a different crop mixture in the following season. Some plants, such as millets, are particularly hardy and will produce a crop even under drought conditions . Although millets may be less palatable and thus not a preferred food, their caloric value is the same as wheat or rice . Similarly, the plant known as lakh is not a preferred food since it results in paralysis in some consumers, but it is noted in colonial documents as a food eaten in place of desired alternatives such as the pulses Cajanus indicus and Cicer arietinum when the latter were unavailable . The presence of Lathyrus sativus in the archaeobotanical record of a relatively large number of Early Historic sites in Table 1 indicates that this plant was known and used in the premodern period as well. Wild foods comprise another important potential component of agricultural systems, although the role of wild resources is often underappreciated for ancient populations. Investigators may discount the presence of noncultigens in archaeological deposits as “weeds,” may lack sufficient comparative collections to positively identify a range of wild flora, or may unwittingly emphasize cultigens in their reports on sites of time periods where agriculture is known to have been practiced. The visibility of wild resources may further be lowered if primary processing in antiquity took place away from the kinds of domestic contexts where archaeobotanical samples are often taken, or if the resource was completely consumed without leaving archaeological remains. However, there are several reasons why wild resources factored into ancient agricultural decision-making. As Clement and B. Smith have recently discussed, the dichotomy between wild and domesticated plants is an overstated one given the many ways in which cultigens can be manipulated, helped along, or fostered without full domestication. In the case of India, it appears that wild and domesticated strains may have been utilized at the same time, with the transition to a fully “domesticated” form difficult to distinguish . Vishnu-Mittre also noted that the archaeobotanical collection from sites such as Navdatoli and Maheshwar “includes the seeds of some weeds which may have been gathered for food”; these include what he identified as 2006] SMITH: ANCIENT AGRICULTURALISTS AND WILD PLANTS 45 Pisum arvense, Lathyrus sphaericus, Vicia sativa, and Vicia tetrasperma. Similarly, at least in modern times, the ber is “not truly wild in India and is seen profusely around former villages or native settlements” . Several of the plants noted in Table 1 may therefore indicate transitional stages, or plants that were tolerated or assisted but not deliberately cultivated. Use of wild flora can also be extrapolated for Early Historic sites on the basis of recovered faunal materials and paleodental records, which show that “the economy was based on a combination of agriculture and cattle pastoralism, augmented by hunting and the exploitation of aquatic and avian resources” . Nineteenth-century documents from central India indicate that many wild plants were used as “famine foods,” a practice echoed in Early Historic texts as well. The Sirupanattrupadai, a South Indian poem of the early centuries BC/AD, paints the scenario of “. . . the wife of the drummer with a lean and slender waist and bangled wrists whom cruel hunger gnawed did saltless cook the herb her sharp nails plucked from refuse heaps, and made a meal of it with poor relations, having closed the door ashamed to be so seen by prying folk” . Another ethnohistorically known famine food represented in the archaeological record is Acacia nilotica, the bark of which is edible when ground and whose seeds can be eaten roasted or raw . Even when they are well supplied by domesticates, however, agricultural groups regularly use a number of wild foods as Pieroni has shown for twentieth-century Italy and Tardío et al. have illustrated for modern Spain.