An international team of marine ecologists recently completed an exhaustive historical study of coastal ecosystems, ranging from coral reefs and tropical seagrass beds to river estuaries and continental shelves. Their findings were disturbing. In every case, fish numbers had declined precipitously with the onset of modern methods of industrial fishing. As the researchers concluded: "Everywhere, the magnitude of losses was enormous in terms of biomass and abundance of large animals that are now effectively absent."
The situation has become especially critical in the past few decades. Stocks of Atlantic cod have reached historic lows, while haddock and other species have been declared commercially extinct. Thriving food webs that were stable for millions of years have in the past 20 been radically altered, and almost three-quarters of the world's commercially important marine fish stocks are now fully fished, overexploited or depleted.
This is just one illustration of the trouble facing the global ecosystem. Biologists estimate that the rate of species extinction worldwide is at least a thousand times greater now than it was before human beings walked the earth, and that one-quarter of all species could be obliterated in 50 years.
But does it really matter to us? The political scientist Bj0rn Lomborg, in The Skeptical Environmentalist, has argued that much of what environmentalists have said is overstated - that fears of ecosystem collapse are irrational and largely the result of scare tactics. On a strict cost-benefit analysis, he says, the consequences of species extinction, like those of global warming, are not serious enough to warrant the expense of trying to stop them. We are better off trying to adapt - by seeking other sources of fish to eat, for example. And many others think the extinction of species is of interest and concern only to nature lovers.
Any ecosystem, however, is a staggeringly complex network in which many species interact with one another in delicate and all but unfathomable patterns. Indeed, it is our inability to understand how these living networks hang together - and consequently, how they might fall apart - that has seriously undermined efforts to assess the vulnerability of the global ecosystem.
But in the past few years, researchers have discovered that ecological networks are not unique in their complexity. In their basic architecture and pattern of assembly, ecosystems turn out to be in many ways identical to other complex networks such as the internet, and even to our webs of social acquaintances.
What emerges from this new science is anything but reassuring. The biological world turns out to be a remarkably small one, with the predator-prey links between species arranged in such a way that no species is more than a handful of steps away from any other. More than anyone suspected, the global ecosystem is an intimately connected whole, and we should indeed be very worried about what we are doing to it.
Most of us have run into a friend of a friend far away from home and felt that the world is somehow smaller than we thought. We usually put such encounters down to coincidence even though they happen with disconcerting frequency. Recent scientific work suggests that this "small world" phenomenon is by no means limited to social relations.
In the social setting, the "small world" experience is closely linked to the notion of "six degrees of separation" - the idea that each of us is linked to everyone else on the planet by a chain of no more than six intermediary acquaintances. Amazingly, this seems to be roughly true. In the 1960s, the American social psychologist Stanley Milgram sent letters to random people living in Nebraska and Kansas, asking each to forward the letter to a stockbroker friend of his living in Boston. He stipulated that they were to send the letter only to someone they knew personally and whom they thought might be socially "closer" to this man. Even though the US then had a population of around 200 million, most of the letters made it to the stockbroker in just five or six mailings.
Researchers have found similarly small worlds in many other settings. The worldwide web is a network of more than one billion sites connected by hypertext links. Take two sites at random, and it needs only about 19 clicks to get from one to the other. Other studies have come upon a similar architecture in the layout of the world's electrical power grids, in the patterns of neural connections in the mammalian brain, and in the web of chemical reactions within the living cell. The world's ecosystems - or more precisely, the food webs that underlie them - appear to share this "small world" character.
How many species-to-species links does it take to link any two organisms in some chain of cause and effect? In the ecological setting, two species are linked if one feeds upon the other, be it a fox eating a rabbit or a beetle munching an oak leaf. Last year, a Spanish physicist, Ricard Sole, and an ecologist, Jose Montoya, studied Silwood Park, an ecosystem in the UK for which researchers know the fairly complete food web. They found the number of degrees of separation to be only two or three. The tapestry of life is made of a truly dense cloth.
Silwood Park does not represent the global ecosystem; it is certainly more than two steps from a woodpecker in Illinois to a shrimp in the South China Sea. Even so, whales and many species of fish populate the oceans as a whole, and numerous birds migrate between the continents. Bacteria, algae, tiny spiders and other creatures fly round the world in storm systems. These organisms provide links that tie the biological world together. For the global ecosystem, the number of degrees of separation may not be two, but it is probably not much higher than ten.
This discovery is not comforting. It suggests that the extinction of one species will affect not only everything that the species eats, competes with, or is eaten by, but will send out fingers of influence which, in a few steps, will reach most other species in the entire system. It suggests that any belief in our capacity to control the effects of ecological destruction is badly misplaced. That lesson becomes clearer as one delves more closely into the small world phenomenon and into exactly how large networks - such as the human social network - can be so remarkably small.
As first suggested by the American sociologist Mark Granovetter in the 1970s, the answer can be seen by making a distinction between "strong" and "weak" social ties. Strong ties bind us to family members and good friends, or to colleagues at work. These links form the threads of a dense fabric of social structure, and are socially most important to us. But these are not the ties that make for a small world.
Each of us also has "weak" links to people we see rarely, or may never see again. Think of some of your friends from the past - long-lost college mates, say. Or someone you met when travelling. Perhaps you went to Japan and briefly made friends with a fellow tourist from Australia. Your links to this person, or to those friends now out of touch, are weak social links.
What makes them especially important is that they connect you to people who otherwise belong to quite distinct social spheres. Your link to the Australian tourist, for example, establishes a social bridge that connects you in just two steps to every person this man knows. Not only that, but this single link connects each of your local acquaintances, in London, say, to every one of his local acquaintances in Australia. In this way, weak links act like short cuts through the social world.
Mathematics backs up this insight. In 1998, in a paper published in Nature, two mathematicians from Cornell University showed that the effect of weak ties in a social network really does explain six degrees of separation. In a large network - even one of six billion people - just a few weak links running between people from distant places will indeed make for an extremely small world, with every pair of persons linked by a short chain of intermediaries.
The small-world character of the world's ecosystems can be traced to similarly weak links - that is, to links between species that interact only occasionally. Perhaps just one bird in an English wood migrates long distances, and, en route, settles briefly in southern Spain. This is enough to link the organisms of these two food webs together by short chains of cause and effect.
But ecologists are beginning to suspect that weak links within food webs also play an important role in maintaining ecosystem stability. Their argument is subtle, but important, as it could help us to protect the world's food webs from disintegration.
If a predator eats just one other species, it will do so frequently, having no other options. Consequently, the link between these species will be strong. Conversely, if a predator feeds upon 15 different prey, it may eat each species only occasionally. It will then have relatively weak links with these species.
Suppose that, after a climate change or some human intrusion, the numbers of a predator's favourite prey have been severely depleted. What will happen? If this particular predator feeds on only this one prey - if they share a strong link, that is - then the predator must continue to seek that prey even though its numbers are vanishing, driving this species even closer to extinction. When this happens, the population of predators may then fall precipitously as well. As a paper in Nature pointed out a few years ago, this should be a general tendency: the loss of a strong link within a food web will be destabilising, tending to stir up large and dangerous fluctuations in species numbers.
But weaker links can save the day. Consider a predator with 15 different prey. If the numbers of one of these species become very low, for whatever reason, the natural response of the predator is to shift its attention to another species that is more numerous and easier to catch. As a result, the predator would continue to find food, while the prey in danger of extinction could revive its numbers. In this way, weak links between species not only make for a small ecological world but also act as natural pressure valves, playing a central role in guaranteeing the health of an ecosystem.
You might expect that all species would have roughly the same number of links with other species. Not so. Nature doesn't dole the links out equitably. Studies in Silwood Park and elsewhere reveal that a few species always play the role of superconnected hubs: they "own" a high fraction of the links in the food web, far and away more than the average species.
By simple logic, most of these links will be weak links. So these hub species provide the network with an ability to redistribute stress and prevent one species from wiping out another by uncontrolled predation or competition. And that explains why we should be so worried about extinctions.
Half the tropical forests, where two-thirds of all species find their habitat, have now been logged or burned to clear land for human development, with another one million square kilometres disappearing every five to ten years. If healthy ecosystems are small worlds characterised by a few hub species, with a preponderance of weak links providing their stability, then the global depletion of species numbers is truly alarming. As species continue to disappear, the remaining species will necessarily be linked more strongly - if only by simple arithmetic. If some predator preys on only six species where before it preyed upon ten, its links with the six will be stronger, and ecosystem stability can only suffer. As one ecologist, Kevin McCann, argues, the lesson is that, if we wish to preserve an ecosystem, or any species within it, we had best proceed "as if each species is sacred".
What's more, the consequences of removing just one of the "superconnected" species can be dramatic, as a huge number of weak stabilising links would go with it. Ecologists have long talked about "keystone" species, crucial organisms whose removal might bring the web of life tumbling down like a house of cards. A recent study has demonstrated just how crucial their preservation may be.
Suppose you begin removing species from an ecosystem. Slowly but surely, the food web should fall apart. But how? First the good news. Sole and Montoya have used a computer to mimic the loss of species from a food web and have found that real communities stand up relatively well when the species to be removed are selected at random. Now the bad news. Suppose instead that the most highly connected species get knocked out first. In this case, ecological disaster ensues quickly. Removing even 20 per cent of the most highly connected species fragments the web almost entirely, splintering it into many tiny pieces. As the web falls apart, the disintegration triggers numerous secondary extinctions as some species lose all their connections to others and become totally isolated.
The obvious answer is to take special care to preserve the highly connected "hub" species. But it is not easy to predict which species will be the hubs for any particular food web. In the past, ecologists have suspected that the hubs would tend to be large predators, but this does not seem to be true. Sole and Montoya found that they were often inconspicuous organisms in the middle of the food chain, or were sometimes basic plants at the very bottom.
Most species now going extinct are ants, beetles and other kinds of insect. Some take comfort in this, but they are wrong to do so. These species may well be linchpins of the living fabric.
What Sole and Montoya achieved on their computer, human activity is achieving in reality - the methodical dismantling of the world's ecosystems. The leaders of many governments and large corporations find it convenient to suppose that worries about the ecosystem are overstated, and anyway, that it would be demented to carry out reforms that are not politically popular. But we are disassembling the web of life that supports our existence, with little understanding of what we are doing. That is truly demented.
Mark Buchanan's Small World: uncovering nature's hidden networks has just been published by Weidenfeld & Nicolson (£18.99)