TO DISPEL any lingering pastoral illusions about present-day agriculture, take a trip to "hog heaven". This ten-mile stretch of countryside north of Ames, Iowa, produces almost a tenth of America's pork. But there is not an animal in sight. In massive metal sheds, up to 4,000 sows at a time are reared for slaughter, their diets carefully monitored, their waste regularly siphoned away, their keepers showered and begowned, like surgeons, to avoid infecting the herd.

Or pay a visit to the farms of the Mato Grosso in Brazil's south-west. Here one man's soyabeans can stretch for 60,000 hectares, and a farmer can take a day to drive his combine down one length of a field and back again. Or wander round a giant greenhouse south of Amsterdam, where 280,000 rose bushes are tended by computer-controlled systems and just the occasional green thumb.

Since the second world war, agriculture has become bigger, more intensive and more specialised, both in the developed and much of the developing world. It is one of the world's largest industries, employing 1.3 billion people and producing $1.3 trillion-worth of goods a year. World output of food per head has gone up by some 25% over the past 40 years, even though land use has grown by only 10% and world population has increased by 90%. Food prices in real terms have fallen by two-fifths, so that rich consumers in America, for example, now spend only 14% of their household income on food.

A success story, then? Not altogether. Modern agriculture is now being shaped by many of the same technologies transforming other industries, but is also subject to unique political and economic constraints. It is expected to produce an abundance of cheap food, but at the same time take account of environmental concerns and look after rural landscapes, the welfare of farmyard animals and the health of consumers. Farmers are supposed to respond to market forces, yet find themselves so insulated in some countries and so marginalised in others that they can scarcely manoeuvre. The current debate on genetically modified foods is merely a symptom of far wider tensions.

Agribusiness -- a term coined by Ray Goldberg at Harvard Business School in the late 1950s -- used to be an orderly chain of companies and institutions stretching from the supply of inputs, such as seeds, fertiliser and machinery, to food processors and retailers. Each link of the chain did most of its business with its neighbours on either side and had little contact with the rest. Family farmers were central to the system. They decided what to grow and where to sell it, taking the risk and reaping a fair portion of the reward. In 1950, the world's agribusiness was worth $420 billion, and farmers added more than a third of the value. By 2028, the market could be worth $10 trillion, and Dr Goldberg reckons that farmers' share of that will fall to 10%.

Global competition and new technologies are forging new relationships within, and between, different layers of agribusiness, transforming the industry from a chain to a complex web. Farming in rich countries is slowly moving from a bulk-commodity industry to "boutique agriculture" producing highly specialised goods. As consumers become better informed and more demanding, companies are consolidating in pursuit of new efficiencies and economies of scale. The trend is most obvious at either end of the food chain: seeds and supermarkets.

Seeds of trouble

Since the development of hybrid crops in the 1920s, commercial farmers have bought their seeds from private firms, paying a premium for such desirable traits as consistent quality and high yield. Until the arrival of biotechnology, the world's $20 billion commercial seed market was highly fragmented, but in the past four years it has been transformed by a series of large deals, most of them done by "life sciences" companies such as Aventis and DuPont.

Their grand plan was to take the technologies that have transformed drug discovery and apply them to agriculture, creating new, "transgenic" plant varieties with traits that could not easily be engineered through conventional breeding. Such firms controlled multi-million-dollar research budgets and held a number of promising patents, but they lacked access to the seeds to which to apply their specialist knowledge, and to the farmers who would plant them. So the life-sciences firms started buying seed companies in bulk, at extravagant premiums. The most prominent example was Monsanto, which in the space of a few years made $8.5 billion-worth of acquisitions and joint ventures, giving it a dominant presence in developed markets such as America and a foothold in such up-and-coming markets as India and Brazil. But smaller firms piled in too, such as Seminis, a Mexican company that advanced from a bit player to a world leader in vegetable seeds in the 1990s.

By 1998, this strategy looked like a winner. Vast swathes of the world's prime grain regions were converting to the new seed varieties. Last year 33% of America's maize and 55% of its cotton crop, and 90% of Argentina's soyabeans, came from genetically engineered varieties, according to Clive James at the International Service for the Acquisition of Agri-biotech Applications. And companies such as Monsanto, which relied heavily on ageing agrochemicals, were enjoying bumper sales, because many of the transgenic crops had been engineered for herbicide resistance.

Despite the premium farmers had to pay for such genetically modified seed and despite having to promise not to replant it from their harvest, they, too, did well out of the new technology. According to a study by Greg Traxler and Jose Falck-Zepeda, of Auburn University in Alabama, gains from planting Bt cotton, genetically engineered to produce insecticides called Bacillus thuringiensis toxins, amounted to $200m in 1997, of which 42% went to farmers, 35% to Monsanto (which held the gene patent) and 7% to consumers.

But that is not the end of the story. Companies have taken each other to court over patents and licensing. Pressure groups such as Greenpeace, and public-sector agricultural institutions such as the United Nations' and World Bank's Consultative Group on International Agricultural Research, are worried that if intellectual-property rights are held by companies, access to technologies and products will become harder and costlier for the poor farmers they try to help. The big biotech firms came a cropper last year as consumers in Europe and Japan mounted an international boycott of genetically modified foods. And even American consumers, who until recently accepted such foods without much thought, are now starting to ask questions.

Resistance on the high street has already registered on Wall Street. Firms that dreamed of combining pharmaceuticals and crops watched the value of their thriving drug businesses being dragged down by their agricultural interests. In the past six months Novartis and AstraZeneca have merged and hived off their seed and agrichemical divisions into a new subsidiary called Syngenta. Monsanto, which announced plans to merge with a pharmaceutical company, Pharmacia & Upjohn, in December last year, is widely expected to sell its agricultural division, as is American Home Products. Further consolidation is likely.

Some firms will continue to benefit. Local seed companies will still be doing deals with biotechnology firms. Outfits such as Ceres, which churn out the basic stuff of genetic engineering -- DNA sequences -- will prosper because their information can be used in less controversial, conventional breeding programmes. And those firms that can reassure uneasy shareholders and keep their drug and agricultural divisions together may ultimately reap the promised synergies. But that could still be a long way off.

Sales force

Much of the pressure on agrochemical and seed companies has come from retailers, whose power to influence the agribusiness chain has grown along with their size. Their industry, too, has seen widespread mergers among supermarket chains, giving them more purchasing power and increasing their global reach.

The supermarket sector is most concentrated in Europe, where, in Germany for example, five supermarket chains control almost two-thirds of the market, according to Rabobank, a Dutch bank. Last year Wal-Mart, the world's largest retailer, followed a foray into Germany with the $11 billion acquisition of Asda, a British supermarket chain. The merger of Promodes and Carrefour, France's two leading supermarkets, created the world's second-largest food retail business. In America, supermarket consolidation was slow to start, but is now gaining ground.

The bigger they get, the more leverage the retailers have over their suppliers, not just on price, but also on issues such as quality and timing of delivery. The retailers sit on a treasure trove of information about their customers, thanks to the product codes stuck on their merchandise that allow them to track their sales in real time. Food manufacturers need this detailed information to meet the retailers' requirements, especially for own-brand lines, which are popular in Europe. And increasingly, supermarkets make contracts direct with producers, cutting out the middlemen.

Those middlemen -- the ones who distribute the milk, pack the beef or collect the grain -- are also consolidating, most obviously in America. Nearly four-fifths of all the cattle going for slaughter there are handled by only four firms. Likewise, four firms crush 80% of the soyabeans to make oil, and another four produce almost 50% of the broiler chickens. Firms such as Smithfield Foods and IBP are not exactly household names, but they control a critical step from production to consumption.

At the same time these middlemen are themselves becoming producers. Most chickens in the industrialised world now come from integrated operations, and the pork industry is moving in the same direction, squeezing out small producers and concentrating the business in certain regions. Vertical integration allows companies to match technological developments upstream to consumer demand downstream. If, for example, a retailer wants pork with a certain colour of fat, company geneticists can go to work on the pig.

Integration also allows companies to control hygiene more rigorously at every point in the production process, and apply the growing list of standards imposed by governments and their new food-safety agencies. This is yet another demand flowing upstream from consumers, who are understandably nervous about the quality of their increasingly industrialised food. Last year animal feed tainted with polychlorinated biphenyls and furans, some of the nastiest persistent organic pollutants known to man, was sold to 1,700 Belgian farmers and found its way into their chickens, pigs and cattle, losing the country's farmers $600m in sales. Britain is still reeling from the effects of mad-cow disease, probably triggered by feeding tainted bone meal to cattle, which cost #3.5 billion ($5.5 billion) in lost exports and culled herds.

Sign on the dotted line

One way in which a processor can control production is through explicit contracts with farmers. Such arrangements usually mean that the grower provides land and labour, whereas the contractor supplies the seeds, chemicals or animals and makes many of the decisions. Such contracting is common in vegetable production and is growing in grain production too, driven by new technologies that are turning bulk commodities into tailor-made products. The first generation of genetically modified crops had "agronomic" properties, such as herbicide resistance, that were useful to growers but held no obvious appeal for consumers. The next generation of high-tech crops, now on their way to market, have been manipulated to produce so-called "output traits" to benefit consumers directly.

For example, biotechnology enthusiasts speak glowingly of the prospect of "nutraceuticals" -- foods with all manner of enhanced nutritional, even medicinal, properties. A few of these, such as rice enhanced with vitamin A, are moving out of the laboratory, but progress is slow because of the sheer technical complexity of creating them, as well as consumer uneasiness about the whole idea of genetic modification. Some four-legged consumers, however, are already eating crops with enhanced output traits to boost their growth or reduce the waste they produce.

Such crops are often grown under contract, and must be separated from other varieties during the trip from farm to consumer, because the benefits of their genetic enhancement will be felt only in their consumption. This is different from, say, insecticide-producing maize, whose benefits are confined to its cultivation. Such products are routinely mixed in with conventional ones in the vast handling systems of commodity shippers such as Cargill and Archer Daniels Midland. But many other types of crops, such as soyabeans for the Japanese market, are already kept separate, so segregating output-trait-modified grains should not prove much of a problem -- provided that consumers will pay a premium for the added value.

Many farmers like contracting because it can pay well and helps them to reduce their risk. For young farmers without capital, it is a way into the business; for those in poor countries, it offers access to new techniques. But others fear falling in hock to a company that has both size and, increasingly, patent protection on its side. And there is the offputting precedent of the chicken farmers, whose contracts tend to be nasty, brutish and short.

In America there is pressure for legislation to offer farmers a fighting chance in the new food web (see ). New partnerships forming along the food chain, such as that between DuPont and General Mills, worry farmers. They foresee perhaps only half a dozen "food clusters" that will control the passage of food from soil to supper.

Yet even in the industrialised world's fastest-moving markets, such as America or Australia, that degree of integration in agribusiness is still a long way off. In Europe and parts of Asia, where an understanding of the domestic subsidy system is still more important than an appreciation of consumer tastes, progress is even slower. Modern agribusiness is not yet ready for the sort of just-in-time manufacturing practised by Dell Computer, in which a signal from a customer triggers a new round of the production cycle. But that day is coming, and farmers are getting ready for it.

** NOTICE: In accordance with Title 17 U.S.C. Section 107, this material is distributed for research and educational purposes only. **