Feathered or furry favourites

However frequent revolutions are in any science, the punctuations are still the most revealing parts of the story. One such is occurring in dino-bird palaeontology, and it is a triple-decker: the relationship of birds with dinosaurs is again under consideration; the magic algorithm - cladistics - for automatically generating family trees of extinct animals directly from the characteristics of their bones is being tested; and the very nature of the scientific process will, for many in the field, be brought into question.

Until now, two theories for the origins of birds have alternated in ascendancy since the mid-19th century. Birds came from dinosaurs. Or they did not.

In Dinosaurs of the Air , however, Gregory Paul intriguingly suggests other alternatives and examines the possibility that although birds came from dinosaurs, some "dinosaurs" were also flightless birds. These have long been known to have existed at the time of the dinosaurs, and one or two dino-birds have been widely accepted as such. But what has been frightening the field since Paul started suggesting the idea in the 1980s, is how far into the ranks of traditional dinosaurs this phenomenon extended.

His 1988 book, Predatory Dinosaurs of the World , attracted unparalleled scorn from the established field, but went on to become a favourite among those dinosaur amateurs lucky enough to find it, because of its visionary ideas, fabulous graphics and clarity of expression. His biggest new idea was that although the first bird, Archaeopteryx , which appeared halfway through the time of the dinosaurs, did indeed come from dinosaur stock, many predatory dinosaurs (theropods) from the second half of their reign (the Cretaceous), had themselves descended from Archaeopteryx . In other words, Velociraptor mimics, Tyrannosaurus rex and a host of others might well have been flightless versions of birds that were power fliers but too primitive to have lost their teeth, hand claws and long bony tails.

This time, Paul again offers new names for groups of these dino-birds and reprieves an approach for defining groups by a few carefully chosen features. Many will be surprised that this custom was ever dropped.

He illustrates his exploration of neo-flightlessness with copious reconstructions and diagrams by the dinosaur artist the others acknowledge as the best among them: Paul himself. Though paintings by one or two others have more artistic beauty, he is a wizard with a pencil and is in a league of his own as a reconstructor and draughtsman of the beasts because of his native common sense and his ability to master the basics of the range of subjects required. To these have to be added the paradoxical benefit of not being formally schooled in palaeontology and his dogged perfectionism.

A number of portraits and diagrams from Paul's first book make a welcome return, supplemented by such a host of new ones that it seems hard to imagine that any aspect of the osteology has been left out. Tricky bits such as the internal bones of the snout, and the subtleties of vertebrae and leg joints, are covered not just in great detail but, with that other Paul trademark, copious comparisons (helping others to develop their own ideas). Who else would offer nine versions of other people's reconstructions of Archaeopteryx 's skull down the years? We are therefore not surprised to find photos of five of the Archaeopteryx skeletons in the book, most in slab and counterslab, as well as close-ups of the head, not to mention the best of the other new fossils of dino-birds and their Triassic companions.

Paul's favourite fossil now, we might guess, is Sinornithosaurus , the earliest-known relative of the Velociraptor clan (famous from Jurassic Park ). When, in 1988, Paul said Velociraptor was his "favourite" dinosaur, the professors threw a wobbly, and they threw another when he said that it and others might be feathered or furry. Exquisite Chinese fossils have since revealed that many were. Paul seems to have skilfully infiltrated himself into the profession now, though, on the "Birds Are Dinosaurs" side. He shows little mercy in deriding the palaeo-ornithologists, the self-styled Band (Birds Are Not Dinosaurs) group, who have staked their reputations on Velociraptor and Archaeopteryx being unrelated, and now find it impossible to recant, even though, as the reconstructions on page 175 show, they were as alike as cat and dog.

This book gathers into one volume practically all the relevant hard evidence to date (except eggs) and deserves fame not only for its thoroughness but for its crucial position in the history of the science. However, it is by no means the last word on the subject. It is the greatest tribute to a researcher to have his corpus picked over and cannibalised by others, and if while he is still alive, so much the better. Paul's current offering certainly leaves the field in an even more refractory state than he wanted.

He usefully compares certain early birds with aircraft such as Spad, Spitfire and Messerschmitt 262. However, one of the fighter pilot's maxims was: "Never follow them down." It seems to me that having beaten up the "Bandits" thoroughly by pointing out the similarity between the first birds and the dinosaurs he terms avepectorans, and drawing our attention to the Bandits' faulty logic, Paul has shot out most of his victims' engines and is now diving his jet steeply to finish them off, only to see the wings fall off his own argument, since he also claims all these avepectorans are descendants of Archaeopteryx . Not only are these no longer candidates for ancestry of the first bird, but it surely follows by inductive logic that the slightly less bird-like theropods of the Jurassic might also have come from small arboreal feathered creatures themselves, according to the principle Paul has so thoroughly demonstrated for the Cretaceous (few fossils are available from that part of the Jurassic to settle the issue). Such small "early birds" might have been strict gliders that would not have flown in flocks or far over water, so Paul's view that we should have found them alongside the copious pterosaur and insect fossils we have from lakes does not hold.

He does not acknowledge the significance of the chameleon-like but plumed Longisquama for bird origins, and thinks that since more bird-like Triassic dinosaurs predated it, it cannot be a bird ancestor. But of course it does not have to be; if its plumes were homologous to feathers it still gives a suggestion of what an early glider ancestral to dino-birds might have been like. He also does not see the very small size of the first dinosaurs as evidence for their ancestors in turn having been gliders (evidence is nothing but an observation that is better predicted or explained by the theory it supports). The book is devoted to Gerhard Heilmann (like Paul a hugely influential professional not employed as a palaeontologist), who famously came within an ace of declaring dinosaurs bird ancestors - one of this book's main messages - but "messed it up" at the last minute. Paul - by saying "it is intriguing that some Triassic archosauriforms" (such as Longisquama ) "seem to have converged with later avepods, including birds, in certain aspects of their morphology, especially in the head and neck, and they even had a superficially feather-like integument" - is giving evidence for the evolution of birds not from dinosaurs, having already unwittingly failed to demonstrate the opposite. He stresses that the apparent convergence in the Cretaceous was in fact inheritance of features. By continuing "but these experiments were ultimately sterile", he may well be following Heilmann more closely than he realises in just missing the truth. Worse, though he defeats the Bandits in the Cretaceous, he brings closer their eventual triumph in the Triassic.

Dinosaurs of the Air will encourage even more intense battles over dino-birds in the future. Paul's ideas will not emerge entirely unscathed at the end, but this book is probably the most heroic contribution to the struggle yet. The final parts of the book provide, among other things: details of the giant birds that evolved after the dinosaurs; a list of 500 characters and the types expressing them, useful for his devotees and cladists alike; and 70 pages refuting John Ruben's ideas on breathing and metabolism. The last alone would provide ample material for many essays on comparative metabolics. "Paul (2002a)" might well become the most-used reference in the field. His theories are, at last, beginning to be acknowledged by the "experts", who will not be able to ignore this book, and who must now break their silence to the media on the "heresy" of neo-flightlessness.

John Jackson is a knowledge engineer.