However, the tree image is misleading in several respects. There is, for instance, no relation between the thickness of a branch and the size of the corresponding population, the thick trunk at the bottom may represent fewer organisms, or less total organic mass, than the twig at the top. (Think about the human twig ...) The way branches split may also be misleading: it implies a kind of long-term continuity of species, even when new ones appear, because on a tree the new branches grow gradually out of the old ones. Darwin thought that speciation, the formation of new species, is gener­ally gradual, but he may have been wrong. The theory of 'punctuated equilibrium' of Stephen Jay Gould and Niles Eldredge maintains the contrary: speciation is sudden. In fact there are excel­lent mathematical reasons for expecting speciation to have elements of both, sometimes sudden, sometimes gradual.

Another problem with the Tree of Life image is that many of its branches are missing, many species go unrepresented in the fossil record. The most misleading feature of all is the way humans get placed right at the top. For psychological reasons we equate height with importance (as in the phrase 'your royal highness'), and we rather like the idea that we're the most important creature on the planet. However, the height of a species in the Tree of Life indicates when it flourished, so every modern organism, be it a cockroach, a bee, a tapeworm, or a cow, is just as exalted as we are.

Gould, in Wonderful Life, objected to the 'tree' image for other reasons, and he based his objections on a remarkable series of fos­sils preserved in a layer of rock known as the Burgess Shale. These fossils, which date from the start of the Cambrian era, are the remains of soft-bodied creatures living on mud-banks at the base of an algal reef, which became trapped under a mudslide. Very few fos­sils of soft-bodied creatures exist, because normally only the harder parts survive fossilization. However, the significance of the Burgess Shale fossils went unrecognized from their discovery by Charles Walcott in 1909, until Harry Whittington took a closer look at them in 1971. The organisms were all squashed flat, and it was virtually impossible to recognize what shape they'd been while alive. Then Simon Conway Morris teased the squished layers apart, and reconstructed the original forms using a computer, and the strange secret of the Burgess Shale was revealed to the world.

Until that point, palaeontologists had classified the Burgess Shale organisms into various conventional types, worms, arthro­pods, whatever. But now it became clear that most of those assignments were mistaken. We knew, for example, just four con­ventional types of arthropod: trilobites (now extinct), chelicerates (spiders, scorpions), crustaceans (crabs, shrimp), and uniramians (insects and others). The Burgess Shale contains representatives of all of these, but it also contains twenty other radically different types. In that one mudslide, preserved in layers of shale like pressed flowers in the pages of a book, we find more diversity than in the whole of life today.

Musing on this amazing discovery, Gould realized that most branches of the Tree of Life that grew from the Burgess beasts must have 'snapped off' by way of extinction. Long ago, 20 of those 24 arthropod body plans disappeared from the face of the Earth. The Grim Reaper was pruning the Tree of Life, and being heavy-handed with the shears. So Gould suggested that a better image than a tree would be something like scrubland. Here and there 'bushes' of species sprouted from the primal ground level. Most, however, ceased to grow, and were pruned to a standstill hundreds of millions of years ago. Other bushes grew to tall shrubs before stopping ... and one tall tree made it right up to the present day. Or maybe we've reconstructed it incorrectly, amalgamating several dif­ferent trees into one.

This new image changes our view of human evolution. One ani_T mal in the Burgess Shale, named Pikaia, is a chordate. This is the group that evolved into all of today's animals that have a spinal cord, including fishes, amphibians, reptiles, birds, and mammals. Pikaia is our distant ancestor. Another creature in the Burgess Shale, Nectocaris, has an arthropod-like front end but a chordate back, and it has left no surviving progeny. Yet they both shared the same environment, and neither is more obviously 'fit' to survive than the other. Indeed, if one had been less evolutionarily fit, it would almost certainly have died out long before the fossils were formed. So what determined which branch survived and which didn't? Gould's suggestion was: chance.

The Burgess Shale formed on a major geological boundary: at the end of the Precambrian era and the start of the Palaeozoic. The early part of the Palaeozoic is known as the Cambrian period, and it is a time of enormous biological diversity, the 'Cambrian explo­sion'. The Earth's creatures were recovering from the mass extinction of the Ediacarans, and evolution took the opportunity to play new games, because for a while it didn't matter much if it played them badly. The 'selection pressure' on new body-plans was small because life hadn't fully recovered from the big die-back. In these circumstances, said Gould, what survives and what does not is mostly a matter of luck, mudslide or no mudslide, dry climate or wet. If you were to re-run evolution past this point, it's quite likely that totally different organisms would survive, different branches of the Tree of Life would be snipped off.

Second time round, it could easily be our branch that got pruned.

This vision of evolution as a 'contingent' process, one with a lot of random chance involved, has a certain appeal. It is a very strong way to make the point that humans are not the pinnacle of creation, not the purpose of the whole enterprise. How could we be, if a few random glitches could have swept us from the board altogether? However, Gould rather overplayed his hand (and he backed off a bit in subsequent writings). One minor problem is that more recent reconstructions of the Burgess Shale beasts suggests that their diversity may have been somewhat overrated, though they were still very diverse.

But the main hole in the argument is convergence. Evolution settles on solutions to problems of survival, and often the range of solutions is small. Our present world is littered with examples of 'convergent evolution', in which creatures have very similar forms but very different evolutionary histories. The shark and the dol­phin, for instance, have the same streamlined shape, pointed snout, and triangular dorsal fin. But the shark is a fish and the dolphin is a mammal.

We can divide features of organisms into two broad classes: universals and parochials. Universals are general solutions to survival problems, methods that are widely applicable and which evolved independently on several occasions. Wings, for instance, are universals for flight: they evolved separately in insects, birds, bats, even flying fish. Parochials happen by accident, and there's no reason for them to be repeated. Our foodway crosses our airway, leading to lots of coughs and splutters when 'something goes down the wrong way'. This isn't a universal: we have it because it so happened that our distant ancestor who first crawled out of the ocean had it. It's not even a terribly sensible arrangement, it just works well enough for its flaws not to count against us when combined with everything else that makes us human. Its deficiencies were tolerated from the first fish-out-of-water, through amphibians and dinosaurs, to mod­ern birds, and from amphibians through mammal-like reptiles to mammals like us. Because evolution can't easily 'un-evolve' major features of body-plans, we're stuck with it.