Each hyrax midden from Petra yielded remains of up to 100 plant species, and the habitat prevailing when each midden's owner was alive could be calibrated by comparing pollen proportions in the midden with those in modern habitats. From the middens, the following trajectory was reconstructed for the degradation of Petra's environment. Petra lies in an area of dry Mediterranean climate not unlike that of the wooded mountains behind my home in Los Angeles. The original vegetation would have been a woodland dominated by oak and pistachio trees. By Roman and Byzantine times, most of the trees had been felled, and the surroundings had been degraded to an open steppe, as expressed in the fact that only eighteen per cent of midden pollen came from trees, the rest from low plants. (For comparison, trees contribute forty to eighty-five per cent of the pollen in modern Mediterranean forests, eighteen per cent in forest steppes.) By 900 AD, a few centuries after Byzantine control of the Petra area ended, two-thirds of the remaining trees had disappeared. Even shrubs, herbs, and grasses had declined, converting the environment into the desert that we see now. Surviving trees today have their lower branches pruned off by goats and are scattered on goat-proof cliffs or in groves protected from goats.

Juxtaposing these data from hyrax middens with archaeological and literary data yields the following interpretation. Deforestation from Neolithic to Imperial times was driven by the clearing of land for agriculture, browsing by sheep and goats, gathering of firewood, and wood needs for house construction. Even Neolithic houses not only were supported By massive timbers but also consumed up to thirteen tons of firewood per house to make the plaster for the walls and floor. The Imperial population explosion quickened the pace of forest destruction and overgrazing. Elaborate systems of channels, pipes, and cisterns were needed to collect and store water for the orchards and city.

After Byzantine authority collapsed, orchards were abandoned and the population crashed, but land degradation continued as the remaining inhabitants became dependent on intensive grazing. The insatiable goats began to eat their way through the shrubs, herbs, and grasses. The Ottoman government decimated surviving woodlands before the First World War, to obtain the wood needed for the Hejaz Railway. I and many other movie-goers thrilled at the sight of Arab guerrillas led by Lawrence of Arabia (a.k.a. Peter O'Toole) blowing up that railway in widescreen technicolour, without realizing that we were watching the last act in the destruction of Petra's forests. Petra's ravaged landscape today is a metaphor for what happened to the rest of the cradle of Western civilization. The modern surrounds of Petra could no more feed a city that commanded the world's main trade routes than the modern surrounds of Persepolis could feed the capital of a superpower such as the Persian Empire once was. The ruins of those cities, and of Athens and Rome, are monuments to states that destroyed their means of survival. Nor are Western civilizations the only literate societies that committed ecological suicide. The collapse of Classic Maya civilization in Central America, and of Harappan civilization in India's Indus Valley, are other obvious candidates for eco-disasters due to an expanding human population overwhelming its environment. While courses in the history of civilization often dwell on kings and barbarian invasions, deforestation and erosion may in the long run have been more important shapers of human history. These are some of the recent discoveries making the supposed past Golden Age of environmentalism look increasingly mythical. Let's now go back to the larger issues I raised at the outset. Firstly, how can these discoveries of past environmental damage be reconciled with accounts of conservationist practices by so many modern pre-industrial peoples? Obviously, not all species have been exterminated, and not all habitats have been destroyed, so the Golden Age could not have been all black. I suggest the following answer to this paradox. It is still true that small, long-established, egalitarian societies tend to evolve conservationist practices, because they have had plenty of time to get to know their local environment and to perceive their own self-interest. Instead, damage is likely to occur when people suddenly colonize an unfamiliar environment (like the first Maoris and Easter Islanders); or when people advance along a new frontier (like the first Indians to reach America), so that they can just move beyond the frontier when they have damaged the region behind; or when people acquire a new technology whose destructive power they have not had time to appreciate (like modern New Guineans, now devastating pigeon populations with shotguns). Damage is also likely in centralized states that concentrate wealth in the hands of rulers, who are out of touch with their environment. Some species and habitats are more susceptible to damage than others—such as flightless birds that had never seen humans (like moas and elephant birds), or the dry, fragile, unforgiving environments in which both Western civilization and Anasazi civilization arose. Secondly, are there any practical lessons that we can learn from these recent archaeological discoveries? Archaeology is often regarded as a socially irrelevant academic discipline that becomes a prime target for budget cuts whenever money gets tight. In fact, archaeological research is one of the best bargains available to government planners. All over the world, we are launching developments that have great potential for doing irreversible damage, and that are really just more powerful versions of ideas put into operation by past societies. We cannot afford the experiment of developing five counties in five different ways and seeing which four counties get ruined. Instead, it will cost us much less in the long run if we hire archaeologists to find out what happened the last time, than if we go on making the same mistakes again. Here is just one example. The American Southwest has over 100,000 square miles of pinyon and juniper woodland that we are exploiting more and more for firewood. Unfortunately, the US Forest Service has little data available to help it calculate sustainable yields and recovery rates in that woodland. Yet the Anasazi already tried the experiment and miscalculated, with the result that the woodland still has not recovered in Chaco Canyon after over 800 years. Paying some archaeologists to reconstruct Anasazi firewood consumption would be cheaper than committing the same mistake and ruining 100,000 square miles of the US, as we may now be doing. Finally, let's face the touchiest question. Today, environmentalists view people who exterminate species and destroy habitats as morally bad. Industrial societies have jumped at any excuse to denigrate pre-industrial peoples, in order to justify killing them and appropriating their land. Are the purported new finds about moas and Chaco Canyon vegetation just pseudo-scientific racism that in effect is saying, Maoris and Indians dp not deserve fair treatment because they were bad? What has to be remembered is that it has always been hard for humans to know the rate at which they can safely harvest biological resources indefinitely, without depleting them. A significant decline in resources may not be easy to distinguish from a normal year-to-year fluctuation. It is even harder to assess the rate at which new resources are being produced. By the time that the signs of decline are clear enough to convince everybody, it may be too late to save the species or habitat. Thus, pre-industrial peoples who could not sustain their resources were guilty not of moral sins, but of failures to solve a really difficult ecological problem. Those failures were tragic, because they caused a collapse in lifestyle for the people themselves.