Book review. - Paleobiology as a subject has evolved considerably in the last 50 years, becoming a reflection of all biology, with the added advantage of the time dimension and the disadvantages of an incomplete record and the inability to experiment. With progress in many areas inevitably has come a loss of confidence in others; we have long since learned to doubt out ability to arrange our fossils in neat lines that tell a story. And although paleobiology seems one of the least likely subjects to fall under the spell of post-modern relativism, we have become less and less comfortable making any kind of firm pronouncement of "fact" about ancestry and origins, once our stock-in-trade. Cladistics, our most powerful tool for analyzing relationships, reduces ancestor-descendent relationships to the status of second or third order (and consequently less reliable) derivatives. The search for ancestors has become unfashionable and the last refuge of the publicity seekers. But I suspect that it lurks unbidden and unacknowledged behind a great deal of what we do. In the odd moments when we yearn for ancestors, cladistically expressed relationships are fine but are literally too much like kissing your sister. Nowhere is this philosophically incorrect urge to identify the "actual certified UL-listed ancestor" more evident than at the two ends of a particular spectrum-the origins of humans and the origins of vertebrates (or should we say, craniates?). But it also colors investigations of the origins of tetrapods, mammals, and birds and may well, in fact, be a particular preoccupation of vertebrate paleontologists. The methodological dilemma is fundamental: what to do with the available evidence? One tactic is to assume that the fossil record is hopelessly incomplete and that comparative study of living forms is simply looking through a glass darkly. Then one creates hypothetical stages of morphology, "stages" (rather than creatures) that never existed, and phylogenies that are reliable only to the extent of their vagueness. The other approach is to work with what we've got, arranging known groups (defined as narrowly as possible) into provisional phylogenies ready to be overturned (or hopefully only fine-tuned) by the next great discoveries. That paleontologists all subtly or openly crave that "great new find" only reinforces the sense of our need for palpable ancestors, not archetypes or sisters. The problem with vertebrate origins is that the ancestors we seek lived so long ago and were so soft bodied that their remains may never be found. This leaves us with a motley collection of living tunicates, hemichordates, and cephalochordates (plus a few fossils) to sort through. As for the known invertebrate groups, it is patently fallacious to derive "any one highly differentiated animal type from another of comparable complexity" (Berrill 1955) but that has not stopped many from trying. Next to ancestors, we want mechanisms, and failing that we can live with process. The first kinds of process we discovered were trends (often quite false, of course). Comparative embryology gave us a different sense of process and has always been a mainstay of vertebrate evolutionary analyses. Comparative DNA data and hox gene distributions give us yet another insight into relationship. All of these have been central in the search for chordate and vertebrate origins. In dealing with this often dense subject, Gee adopts a breezy, light-handed, but sure style. He is fascinated equally by the characters in this odd history, the power of different narratives, in and out of fashion, to sway an audience regardless of merit, and the powerful attraction of an unsolved mystery. Vertebrate origins in the perfect mystery. The result is, as they say, a "good read." As a reference tool its value is diminished by the curious inadequacy of the index. The tension between looking for ancestors among known and hypothetical taxa marked the search for vertebrate origins from an early date. Dominating the field early were two massive works of comparative biology: Gaskell's "The Origin of Vertebrates" (1908) and Patten's "The Evolution of Vertebrates and their Kin" (1912). Both sought the ancestry of vertebrates among arthropods turned upside-down (an idea recently made less improbable on developmental grounds [De Robertis and Sasai 1996]). Behind some of this lies a fascination with the resemblances between a horseshoe crab carapace and the cranio-thoracic shields of ostracoderms and placoderms. One senses also the authors getting caught up in the pure intellectual challenge of finding morphological homologies and devising anatomical reorganizations, however tortuous the logic, once the initial premise of dorsoventral reversal had been accepted. Most recently, Jeffferies' "The Ancestry of Vertebrates" (1986), is set in the same steamroller style of massed details and extrapolations, finding ancestors among the fossil stem-echinoderm "calcichordates." The is the most comprehensive study in modern times, aimed at finding connections between two fully differentiated and diversified groups. Perhaps one of Gee's most useful contributions is to sort carefully through Jefferies' evolving arguments clearly, carefully, and impartially. Jefferies at least finds the ancestor from among vertebrates' co-deuterostomes. Quite early on, biologists like MacBride had seen that, of all invertebrates, the echinoderms shared more characters with vertebrates, a view that has been extended by biochemical and genetic results. Meanwhile, there was a growing problem of what to do with amphioxus and the other protochordates. Was amphioxus the ancestor, and if so, what were its relations? Gaskell's solution was to consign amphioxus to the status of secondary degenerate. Goodrich, among many others, opposed this view. Consistently most popular among attempts to get around the problems faced in making a vertebrate out of any known invertebrate has been the tunicate tadpole theory-really an elaborate narrative-of Kovalesky, Garstang, and Berrill. This had the great attraction of combining comparative morphological approaches with the growing field of embryology and its hopes first of recapitulatory, and later of paedomorphic, glory. Here Garstang is the first great modern hero of the drama. His works well merit rereading today and that reading will reveal that his light-hearted "Larval Forms and Other Verses" (1951) are thoroughly science zoology. The tadpole theory circumvents the problem of comparing "fully differentiated" forms by seeking origins in the processes creating that differentiation. Berrill's recasting of the tadpole theory is criticized by Gee for a lack of internal rigor but despite these lapses, like its rival predecessors, it has the momentum and resilience of all good stories. Berrill's ideas were championed by that great enthusiast Alfred Sherwood Romer whose "dual animal" metaphor (1972) looks more and more intriguing as an analysis of the vertebrate system of functioning. A curious result of all this argumentation about arthropods and tadpoles is that for many years that most fascinating of animals, amphioxus (regrettably renamed Branchiostoma in one of zoology's periodic fits of legal rectitude), was neglected. Is it "that wretched animal without brains, without eyes, without a nose" or is MacBride right: "no one who has seen it swim will fail to realize the immeasurable superiority of the Vertebrate motor system over that of the Arachnid. The comparison of the one to a screw of a steamer and of the other to an eight-oared boat gives some idea of the difference"? The two quotations come from a wonderful collection of essays from a debate on this subject organized by the Linnean Society of London for the 1909 Darwin Jubilee (Gaskell et al, 1910). One would give a great deal to be able to time-warp back to the Linnean Society's rooms and listen to Gaskell, MacBride, Goodrich, Gadow, Lankester, Smith Woodward, and a host of others, going at it hammer and tongs, their zoology colored by their (evolutionary) politics, their invective coated in silvery Edwardian politeness. (We must be grateful to Gee for "discovering" this forgotten gem.) As Gee demonstrates in this delightful book, none of the existing theories of vertebrate origins quite worked, then or now. As the Rev. T. R. R.Stebbing remarked in the debate just cited, "When we return home and our friends gleefully enquire, 'What then has been decided as to the Origin of Vertebrates?' so far we seem to have no reply ready, except that the disputants agreed on one single point, namely, that their opponents were all in the wrong." All of which makes the search equally frustrating and fascinating. Without breaking into new territory, we will continue to chase our tails around the same circle, defined by the same groups as 100 years ago. Where must the search lead next? Some Cambrian fossils to add to creatures like Pikaia would be nice. We need also to know a great deal more about amphioxus and its fellow(?) protochordates. These forms, like so many other phylogenetically important taxa (lungfishes, for example) were one of the many casualties of modern biology's headlong rush to become (to rephrase Medawar) the art of the manipulatable. Mice and chick-not even the old amphibian standbys-became de rigueur for successful grant applications. (What a wealth of interesting information has come from the discovery that nematodes and the zebra fish are good experimental animals!) Rescue has come from a most unexpected source. Modern developmental biology, resurrected from the old "experimental embryology" by the power of molecular tools and language, has revived interest in amphioxus and thereby in the origins and relationships of vertebrates. Perhaps the best result of the last 20 years therefore is not conceptual, but the revival of study of amphioxus, conodonts, hemichordates, and urochordates, which have been unconscionably neglected given their potential importance to this subject. And, of course, patterns of hox-type genes from every kind of animal from nematodes to humans continue to add new, sometimes conflicting, results. New evidence on homologous gene sequences sheds light on one of the oldest questions: where did the vertebrate head come from? Amphioxus lacks one. Hox sequence results suggest, however, that the head in craniates is not an entirely new invention but corresponds largely to the "head" of amphioxus. The forebrain region and everything in front of the tip of the notochord may be another matter. Gee's book is the history of particular quest. For maximum benefit it should be read side-by-side with the contemporary, and very elegant, historical study by Peter Bowler. An intellectual rather than scientific history, "Life's Splendid Drama" (1996) sheds a great deal of light on the fascinating question of why people thought and argued the way they did throughout the period in which debate over vertebrate origins flourished. Gee concludes that none of the elaborate theories and comparisons of the last hundred years has yet produced either a convincing chordate ancestor or vertebrate ancestor. After reviewing claims of rival authors to possess the truth, he gracefully refuses to come down on one side or the other. Indeed, no other ending is possible, frustrating as that might be. The result is a thorough summary of a tantalizing but incomplete subject.
Keith Stewart Thomson. New School for Social Research 65 West Eleventh Street New York, New York 10011