A peek under the canopy

A tropical rainforest is an ecosystem of infinite complexity, beauty, conflict and occasional danger. But a newcomer can find it bewilderingly lifeless - unaware that its millions of species are concealed by camouflage, the canopy, stealth and nocturnal darkness. A stranger needs a skilled, articulate guide. He has one in David Campbell.

In A Land of Ghosts , Campbell revisits botanical transects in the forests of the Serra do Divisor (Watershed Hills) that separate the remotest south-western corner of Brazil from Peru. He is a seasoned botanist with a fund of fascinating anecdotes about the flora and fauna of his beloved forests. His knowledge is augmented by his team of caboclos , the resilient people who remained in the Amazon after the rubber boom and became the only outsiders, apart from indigenous peoples, to adapt to this challenging environment. Caboclos have jeito , a versatile Portuguese word for savvy, sharp wits and know-how.

On a trail one night Campbell's torch shines into the eyes of a spiny tree rat. His woodsman Tarzan warns him that it is dead: be careful. Sure enough, they soon see a large, perfectly camouflaged surucucNo pit viper returning to eat his victim. The snake lunges at the torch beam, attracted not by its light but by its warmth, led by sensors that follow the heat trails left by potential prey. Campbell explains the science behind all this: how the snake's venom and heat-sensors work, its behaviour patterns, and how its jaws expand to take the rat.

What makes his book a joy to read is that it imparts science lucidly to a lay reader, and with a beautiful style. Campbell appreciates the minutiae, the smells and sounds, and the dynamics of the forest. Here he describes a tiny tragedy: " A bico de argulha - a jacamar - waits on a twig, stiletto-still, monitoring the insects that pass through a light fleck. Now it spots a darner laying her eggs in the flooded chalice of a bromeliad, dipping the tip of her abdomen into the still pool. After each bout of egg-laying the darner rises - using her fixed compound eyes to interpret the air for an enemy's movement - then dips again. But the jacamar darts into the shadows and attacks from darkness, clipping the insect in midair like a pair of tweezers. It carries the darner to a low bough and, with violent swats of its head, clubs her to pieces. The thuds are audible on the moss-covered branch. The darner's thorax crumples: her four clear wings twist in different directions. One wing falls to the forest floor, turning in the still air, intercepts a shaft of sunlight, and splinters it into a momentary rainbow."

The tropical forest is brought to life through countless similar observations. Campbell is best when telling us about botany: how taxonomists go about their tedious, painstaking work, and why and how they classify so many features of each species - particularly the sex organs, as decreed by Linnaeus. He imparts modern thinking about species diversity, and the importance of tree-fall light traps in perpetuating such diversity. He ponders why western Amazonia has the greatest biodiversity. Clouds hitting the Andes yield more rain, which could cause more trees to topple, with resulting evolution of new species. Another reason may be blasts of cold air from Antarctica, friagems that contribute to the gaps that lead to unparalleled diversity. Campbell reminds us how much Amazonia changed during the drier ice ages, when rainforests retreated into refugia in which vegetation evolved differently.

The book does have shortcomings. Although full of fascinating knowledge, it is no manual and has little discernible pattern. At times it is an account of the team's journey, then there are vivid portraits of the abandoned frontier and its elderly survivors, then a jumble of observations from the author's various trips onto the upper Juru . Someone unfamiliar with Brazil and its forests could be irritated by the peppering of vernacular names and Portuguese words.

There are quite a few spelling mistakes and unnecessary accents. The author's forays into history are not as strong as his natural history. He does not realise that the state of Acre's straight northern border came from a line drawn by negotiators of the Treaty of Madrid in 1750. An often-cited 17th-century missionary was Laureano de la Cruz, not Leandro; the botanist Richard Spruce was a proud Yorkshireman, not aScotsman.

But these are minor quibbles about a vibrant book, one that truly captures the wonder of the Amazon rainforest and the decay of its remote frontier settlements.

John Hemming was director, Royal Geographical Society, from 1975 to 1996.

He has written or edited nine books on Brazilian Amazonia.