Dinosaurs, the big dumb oafs of the past, are getting a new image from contemporary paleontologists. Adrian Desmond, an English expert now at Harvard, and colleagues on both sides of the Atlantic have done surface-to-volume measurements, comparative studies of limb anatomy, facial structure and brain size. Their conclusion is that dinosaurs truly belong to a (taxonomic) class by themselves; they must have been warm-blooded like mammals rather than cold-blooded like lizards. Otherwise they never would have had enough get-up-and-go to move swiftly on two feet as huge predators like Tyrannosaurus did, or browse, elephant-style, like some of the superherbivores. You need an internally maintained temperature to achieve high metabolic efficiency and get energy fast to muscles and limbs. Lacking such an arrangement, as Desmond comments about a swift species of ostrich-like dinosaurs, would be ""tantamount to building the Concorde, then equipping it with renovated Sopwith Camel engines!"" A few latter-day dinosaurs even experimented with feathers as insulation. This, along with other arguments, leads Desmond to suppose that birds are actually descended from dinosaurs. He admits there are difficulties with these revolutionary theories--but maybe fewer than exist as things stand now. He presents a persuasive case, and though at times the way is heavy with nomenclature and details, there are also delightful excursions with eccentric 19th-century fossil collectors.