B OOKS ET AL. model of the solar system. Add to that the EVOLUTION Unfortunately, nearly all refer to the benefits cryptic titles of the tales and one becomes of knowing the phylogeny for a particular downright dizzy. Dawkins has written Seeing the Forest group and say little about those benefits that another wonderful book, but the manuscript can only come from assembling the entire would have benefited from a good, firm edit. for the Trees Tree. In addition to revealing common pat- Had the text been reduced by a third and the terns or coordinated evolution among clades, marginal notes either incorporated in the Scott J. Steppan having the whole Tree should lead to more main text or eliminated, The Ancestor’s Tale important but as yet unanticipated insights. could have become a true classic. It should ssembling the Tree of Life presents a For example, would Wegener have imagined also be noted that Dawkins is an entertaining preliminary view of one of the grand continental drift if he had only a collection of author precisely because he is not afraid to Aenterprises of modern science, resolv- road maps and no global map to work with? express opinions. However one might feel ing the phylogeny of all life. Imagine a vast We biologists need our own globe. about the particulars, at times these opinions tree whose myriad branches lead to millions Following the introductory section, 26 become downright caustic, and they trivialize of leaves. Each leaf, itself composed of innu- chapters by authorities on major branches the tales in which they appear. Conservative merable parts, represents an individual (clades) summarize the state of our phylo- readers might risk an aneurysm. species in the history of genetic knowledge. These begin at the base The taxonomic scope of The Assembling the life, and the tree stands of the Tree, where contributors highlight, Ancestor’s Tale is strongly affected by the Tree of Life billions of years tall. for example, the recent recognition that the author’s understandable decision to fol- Joel Cracraft and Revealing that tree is earliest branchings split life into three low human ancestry. We are animals, so Michael J. Donoghue, the shared vision of the domains: the bacteria, archaea, and eukary- this is fundamentally a book about ani- Eds. world’s systematists, otes. The chapters then proceed up the Tree mals (i.e., metazoans). In the one chapter but for now it remains a through smaller branches and less inclusive Oxford University Press, dream. We do not know on plants, Dawkins comments, “I ended a New York, 2004. 592 pp. groups (e.g., green plants, animals, and previous tale by remarking what delight it $59.95, £36.50. ISBN 0- what the whole “Tree of arthropods) to consider such “crown” is to be a zoologist at such a time. I could 19-517234-5. Life” looks like. We can groups as flowering plants, annelid worms, have said the same about being a botanist. only see parts of it, and and birds. In each chapter, the authors sum- What a pleasure it would be to demon- our situation is worse marize the constituent subgroups and typi- strate Deep Green [(2)] to Joseph Hooker— than that of the proverbial three blind men cally describe supporting evidence, regions in the company of his close friend Charles trying to describe an elephant. Thousands of of uncertain relationship, and definitive Darwin. I almost weep to think about it.” us work on particular branches, which are morphological features. Afterward, Don- Nonetheless, in practice Dawkins seems to hidden from one another in a mist. This vol- oghue and three other leading evolutionary follow another great philosopher of science, ume, the product of a 2002 symposium by biologists (Edward Wilson, David Wake, Tom Weller, who said, “The evolution of the same name held at the American and David Hillis) offer short summary per- plants is an important chapter in the history of Museum of Natural History (AMNH) in spectives. In the final chapter, the editors tie life. However, it’s a pretty dull chapter, so New York, seeks to blow away the mist and everything together by assembling a 138- we’ll skip it.” (3) Furthermore, to my aston- reveal the structure of the whole Tree and, in taxon synoptic tree. ishment, the Archaea—a third of all life—are doing so, galvanize the systematics commu- Taken individually, the chapters are use- not allotted a single tale. nity toward unifying its goals. ful summaries of our current understand- But so be it. Dawkins is an enormously A complete Tree of Life (hereafter ing, but they seem like disconnected limbs. talented author, and The Ancestor’s Tale is “Tree”) holds enormous promise for many Nonetheless, the Tree will start to assemble expansive, current, and authoritative. There fields of science, but the task of revealing it itself—an emergent property of the discon- are, of course, technical errors and dubious is an enormous undertaking—one that nected parts—in the minds of those readers assertions to be found. Few texts of such requires more data than the Human Genome who take the time to read far enough. In that scope are without them. These flaws, how- Project (just one leaf on the Tree) and orders indirect way, the editors have met their goal. ever, are mostly minor, and the book avoids of magnitude more computation. Even small In addition, even the most broadly trained many pitfalls that have trapped other parts are difficult; as Michael Whiting notes, comparative biologists will discover unap- authors. It would be an excellent choice for “A child can tell a beetle from a wasp an undergraduate honors seminar in zool- from a butterfly, but even the entomo- ogy and could serve a graduate student well logically erudite is left pondering in preparation for oral exams. It is also which two insects are most closely entertaining, witty, and—at least in the related.” The volume, edited by lead- “Pilgrimage to the Dawn of Life” version— ing systematists Joel Cracraft beautifully illustrated. If it still leaves room (AMNH) and Michael Donoghue for the botanists and microbiologists of the (Yale University), begins with three world to present their perspective on the tree chapters that explain why assembling of life, who am I to complain? the Tree is important to science and society. Most of the reasons offered References and Notes will be familiar to biologists, as the 1. R. Dawkins, The Selfish Gene (Oxford Univ. Press, revolution in systematics has pene- Oxford, 1976). 2. “Deep Green” refers to a project to reconstruct the trated many different fields. phylogeny of green plants and to a hyperbolic visuali- zation of that phylogenetic tree, available at http://ucjeps.berkeley.edu/map2.html. The reviewer is in the Department of 3. T. Weller, Science Made Stupid (Houghton Mifflin, Biological Science, Florida State University, Boston, 1985). Tallahassee, FL 32306–1100, USA. E-mail:

CREDIT: DAVID SACKS/TAXI/GETTY IMAGES CREDIT: SACKS/TAXI/GETTY DAVID 10.1126/science.1105582 [email protected] Iconic metaphor.

www.sciencemag.org SCIENCE VOL 307 4 FEBRUARY 2005 677 Published by AAAS B OOKS ET AL. treatment of the debate over the BROWSINGS effects of long-branch attraction on The Elements. Earth, Air, Fire, and analyses of the relationship Water. Art Wolfe,text by Craig Childs. between the fly orders Strepsiptera Sasquatch, Seattle, WA, 2004. 176 and Diptera (2). A more pervasive pp. $45. ISBN 1-57061-405-9. and subtle, yet profound, conse- In ancient Greece, philosophers quence of this methodological bias held that the physical world around is omission from most chapters of them was derived from four ele- fundamental aspects of evolution- ments: earth, air, fire, and water. ary history, like timing of events Wolfe presents a collection of his and rates of diversification. The color landscape and nature photo- emphasis in the volume is entirely graphs that explores the diversity of on the sequence of branching. forms these substances take. There (Branch lengths are not important are sheer rock walls of the Karakoram in parsimony analysis, and their Range, Pakistan, and loose sediments estimates are generally unreliable. of the Colorado River’s subaerial In contrast, they are integral to delta plain, Baja California (left); model-based methods.) As a result, morning mists and midday clouds; the tempo and mode of evolution volcanic eruptions and flaming (3) are lost, and we cannot see forests; Hawaiian breakers and whether the Tree looks like a Antarctic ice. The four elements are spreading oak, a willow, or a bam- also used as background for portraits boo grove—we have little sense of of wildlife, flowers, and trees. Many its gestalt. The lack of resolution in of the images are carefully composed some parts of the Tree is therefore to capture patterns of light and con- attributed to a lack of data rather trast. In four short essays, Childs than to the much more interesting offers his impressions of the effects possibility of rapid diversification. the elements have on humans and Branch lengths—as indicators of the natural world. time or amount of evolution—are important to almost every aspect of comparative biology, and the vol- preciated diversity in less familiar groups its revelation of the patterns among diverse ume would have benefited from the more and the kind of fascinating organisms that clades. Many authors cite the explosion of nuanced vision their consideration would inspired many of us to become biologists. molecular data as the reason for the revolu- have offered. These benefits would have been even easier tion in phylogenetics, especially for the The summary chapters praise the to appreciate if the material was presented field’s transformation since the previous progress and promise more to come. I in the more dynamic and immersive experi- symposium that attempted to view phyloge- would have preferred a more critical analy- ence of the volume’s Web analog, the Tree netics across all of life (1), held in 1988. The sis of the overall state of this resource- of Life project (http://tolweb.org/tree/). (It most publicized cases of conflict between limited field. Where are the biggest holes? is a shame that updatable, peer-reviewed molecules and morphology are not repre- Should we focus on broad taxonomic cover- Web pages still lack the professional status sentative of that revolution: The tidal surge age of a few universal genes, overlapping of static book chapters.) of molecular data seems to have confirmed sets of many genes, or perhaps new initia- Most authors have taken their charge numerous old hypotheses while rejecting a tives to train morphologists? But in the end, very seriously and have written unbiased, few but, most importantly, resolving many the big picture emerges from the details, synthetic, and useful accounts. Particularly branches that morphological evidence did and we gain a better appreciation of how the readable chapters include those on not. One is struck by the great reliance on a branches fit together and where some of the Holometabola (insects characterized by single gene—the small subunit (SSU) of the bigger questions remain. The vision complete metamorphosis), land plants, and ribosomal DNA, also known as 18S—for Cracraft and Donoghue articulate in their chordates (vertebrates, hagfish, lancelets, most resolution deep in the Tree, even introduction does emerge from the mist, and tunicates). A minority of the contribu- within phyla. Elsewhere, despite frequent incomplete though it may appear. tors have yielded to provincialism, focusing accolades to molecular data, the recognition Assembling the Tree of Life should also on their own work or dismissing informa- of many clades (especially among chor- meet the editors’ larger goal. It will help the tion (e.g., molecular) that they distrust. The dates) continues to rely on morphology. systematic community aspire toward a most extreme position appears in the mam- The other broad impression the volume common goal, identify priorities for future mal chapter, whose authors eschew the leaves is that of an imbalance toward coordinated work, and mobilize our summary format in favor of lecturing on authors who favor parsimony for phyloge- resources. their preferred systematic procedures. Only netic analysis over model-based or sta- a handful of conflicting conclusions appear; tistical methods such as likelihood. References and Notes

one is the description of the Holometabola Individually, this imbalance is not very 1. B. Fernholm, K. Bremer, H. Jörnvall, Eds., The Hierarchy THE ELEMENTS of Life: Molecules and Morphology in Phylogenetic as a group whose monophyly is either rou- important because all chapters include Analysis (Nobel Symposium 70, Elsevier, Amsterdam, tinely supported by both morphology and authoritative authors. The reviews of find- 1989). 2. J. P. Huelsenbeck, Syst. Biol. 46, 69 (1997). molecules (Whiting) or never supported in ings by other researchers are generally fair, 3. G. G. Simpson, Tempo and Mode in Evolution any molecular data (Rainer Willmann). although occasionally conflicting model- (Columbia Univ. Press, New York, 1944).

The volume’s principal utility stems from based results are brushed aside—as in the 10.1126/science.1106586 CREDIT:ART WOLFE FROM

678 4 FEBRUARY 2005 VOL 307 SCIENCE www.sciencemag.org Published by AAAS