VACLAV SMIL 273 How restrictive are natural factors? Photosynthetic productivity depends on the availability of solar radiation, atmospheric carbon dioxide, plant nutrients, land, water, and sufficient biodiversity. Crop yields are almost never limited by the incoming solar radiation. Similarly, the current atmospheric concentration of CO2 (about 350 parts per million) is adequate for sustained high biomass yields, and increasing concentrations of the gas will tend to enhance the photosynthetic efficiency of nearly all well-watered and well-fertilized crops.37 Thus the four critical natural determinants of future crop productivity are the availability of land, nutrients, and water, and the protection of adequate biodiversity.
Extreme carrying capacity estimates go far outside the broad, fourfold range bracketed by the estimates just cited. They have been defined by true believers in the antipodal camps of catastrophist and cornucopian futures. A generation ago Ehrlich (1968) wrote that "the battle to feed all humanity is over" and that "hundreds of millions of people are going to starve to death" during the 1970s.2Ehrlich's global population maximum would have to be well below the 1970 total of about 3.7 billion people. In contrast, Simon (1981) maintained that food has no long-run, physi-cal limit. These extremes leave us either with the prospect of eliminating about half of humanity in order to return the worldwide count to a sup-portable level or with visions of crop harvests surpassing the mass of the planet itself.3 As Sauvy (1990: 774) noted crisply, "Lack of precision in data and in method of analysis allows shortcuts toward reaching an ob-jective predetermined by prejudice, shaped largely either by faith in progress or by conservative skepticism." Unfortunately, less extreme estimates have been hardly more impressive. Because the question of the ultimate support capacity cannot have a single correct answer, assessing the value of past estimates must be done by looking at their assumptions. Too many of them are overly simplistic, and even the more elaborate ones are usually difficult to defend. In general, the capacity predictions assume too much-as well as too little. Most notably, they almost completely ignore the demand side of the question.