decisive roles in the European conquests of Native Americans, Australians, South Africans, and Pacific islanders. In short, plant and animal domestication meant much more food and hence much denser human populations. The resulting food surpluses, and (in some areas) the animal- based means of transporting those surpluses, were a prerequisite for the development of settled, politically centralized, socially stratified, economically complex, technologically innovative societies. Hence the availability of domestic plants and animals ultimately explains why empires, literacy, and steel weapons developed earliest in Eurasia and later, or not at all, on other continents. The military uses of horses and camels, and the killing power of animal-derived germs, complete the list of major links between food production and conquest that we shall be exploring. CHAPTER5 history's haves and have-nots MUCH OF HUMAN HISTORY HAS CONSISTED OF UNEQUAL conflicts between the haves and the have-nots: between peoples with farmer power and those without it, or between those who acquired it at different times. It should come as no surprise that food production never arose in large areas of the globe, for ecological reasons that still make it difficult or impossible there today. For instance, neither farming nor herding developed in prehistoric times in North America's Arctic, while the sole element of food production to arise in Eurasia's Arctic was reindeer herding. Nor could food production spring up spontaneously in deserts remote from sources of water for irrigation, such as central Australia and parts of the western United States. Instead, what cries out for explanation is the failure of food production to appear, until modern times, in some ecologically very suitable areas that are among the world's richest centers of agriculture and herding today. Foremost among these puzzling areas, where indigenous peoples were still hunter-gatherers when European colonists arrived, were California and the other Pacific states of the United States, the Argentine pampas, southwestern and southeastern Australia, and much of the Cape region of South Africa. Had we surveyed the world in 4000 b.c., thousands of years after the rise of food production in its oldest sites of origin, we would have been 94 *GUNS,GERMS, AND STEEL surprised too at several other modern breadbaskets that were still then without it—including all the rest of the United States, England and much of France, Indonesia, and all of subequatorial Africa. When we trace food production back to its beginnings, the earliest sites provide another surprise. Far from being modern breadbaskets, they include areas ranking today as somewhat dry or ecologically degraded: Iraq and Iran, Mexico, the Andes, parts of China, and Africa's Sahel zone. Why did food production develop first in these seemingly rather marginal lands, and only later in today's most fertile farmlands and pastures? Geographic differences in the means by which food production arose are also puzzling. In a few places it developed independently, as a result of local people domesticating local plants and animals. In most other places it was instead imported, in the form of crops and livestock that had been domesticated elsewhere. Since those areas of nonindependent origins were suitable for prehistoric food production as soon as domesticates had arrived, why did the peoples of those areas not become farmers and herders without outside assistance, by domesticating local plants and animals? Among those regions where food production did spring up independently, why did the times at which it appeared vary so greatly—for example, thousands of years earlier in eastern Asia than in the eastern United States and never in eastern Australia? Among those regions into which it was imported in the prehistoric era, why did the date of arrival also vary so greatly—for example, thousands of years earlier in southwestern Europe than in the southwestern United States? Again among those regions where it was imported, why in some areas (such as the southwestern United States) did local hunter-gatherers themselves adopt crops and livestock from neighbors and survive as farmers, while in other areas (such as Indonesia and much of subequatorial Africa) the importation of food production involved a cataclysmic replacement of the region's original hunter-gatherers by invading food producers? All these questions involve developments that determined which peoples became history's have-nots, and which became its haves. Before we can hope to answer these questions, we need to figure out how to identify areas where food production originated, when it arose there, and where and when a given crop or animal was first domesticated. The most unequivocal evidence comes from identification of plant and HISTORY'Shaves and have-nots * 9 5 animal remains at archaeological sites. Most domesticated plant and animal species differ morphologically from their wild ancestors: for example, in the smaller size of domestic cattle and sheep, the larger size of domestic chickens and apples, the thinner and smoother seed coats of domestic peas, and the corkscrew-twisted rather than scimitar-shaped horns of domestic goats. Hence remains of domesticated plants and animals at a dated archaeological site can be recognized and provide strong evidence of food production at that place and time, whereas finding the remains only of wild species at a site fails to provide evidence of food production and is compatible with hunting-gathering. Naturally, food producers, especially early ones, continued to gather some wild plants and hunt wild animals, so the food remains at their sites often include wild species as well as domesticated ones. Archaeologists date food production by radiocarbon dating of carbon-containing materials at the site. This method is based on the slow decay of radioactive carbon 14, a very minor component of carbon, the ubiquitous building block of life, into the nonradioactive isotope nitrogen 14. Carbon 14 is continually being generated in the atmosphere by cosmic rays. Plants take up atmospheric carbon, which has a known and approximately constant ratio of carbon 14 to the prevalent isotope carbon 12 (a ratio of about one to a million). That plant carbon goes on to form the body of the herbivorous animals that eat the plants, and of the carnivorous animals that eat those herbivorous animals. Once the plant or animal dies, though, half of its carbon 14 content decays into carbon 12 every 5,700 years, until after about 40,000 years the carbon 14 content is very low and difficult to measure or to distinguish from contamination with small amounts of modern materials containing carbon 14. Hence the age of material from an archaeological site can be calculated from the material's carbon 14/carbon 12 ratio. Radiocarbon is plagued by numerous technical problems, of which two deserve mention here. One is that radiocarbon dating until the 1980s required relatively large amounts of carbon (a few grams), much more than the amount in small seeds or bones. Hence scientists instead often had to resort to dating material recovered nearby at the same site and believed to be 'associated with' the food remains —that is, to have been deposited simultaneously by the people who left the food. A typical choice of 'associated' material is charcoal from fires. But archaeological sites are not always neatly sealed time capsules of 96 • GUNS, GERMS,and steel materials all deposited on the same day. Materials deposited at different times can get mixed together, as worms and rodents and other agents churn up the ground. Charcoal residues from a fire can thereby end up close to the remains of a plant or animal that died and was eaten thousands of years earlier or later. Increasingly today, archaeologists are circumventing this problem by a new technique termed accelerator mass spec-trometry, which permits radiocarbon dating of tiny samples and thus lets one directly date a single small seed, small bone, or other food residue. In some cases big differences have been found between recent radiocarbon dates based on the direct new methods (which have their own problems) and those based on the indirect older ones. Among the resulting controversies remaining unresolved, perhaps the most important for
