Scott J. Steppan, "Seeing the Forest for the Trees" (2005)

EVOLUTION:
Seeing the Forest for the Trees, "Science" February 4, 2005.

A review by Scott J. Steppan

Assembling the Tree of Life, Joel Cracraft and Michael J. Donoghue, Eds. Oxford University Press, New York, 2004. 592 pp. $59.95, Ł36.50. ISBN 0-19-517234-5.

Assembling the Tree of Life presents a preliminary view of one of the grand enterprises of modern science, resolving the phylogeny of all life. Imagine a vast tree whose myriad branches lead to millions of leaves. Each leaf, itself composed of innumerable parts, represents an individual species in the history of life, and the tree stands billions of years tall. Revealing that tree is the shared vision of the world's systematists, but for now it remains a dream. We do not know what the whole "Tree of Life" looks like. We can only see parts of it, and our situation is worse than that of the proverbial three blind men trying to describe an elephant. Thousands of us work on particular branches, which are hidden from one another in a mist. This volume, the product of a 2002 symposium by the same name held at the American Museum of Natural History (AMNH) in New York, seeks to blow away the mist and reveal the structure of the whole Tree and, in doing so, galvanize the systematics community toward unifying its goals.

A complete Tree of Life (hereafter "Tree") holds enormous promise for many fields of science, but the task of revealing it is an enormous undertaking--one that requires more data than the Human Genome Project (just one leaf on the Tree) and orders of magnitude more computation. Even small parts are difficult; as Michael Whiting notes, "A child can tell a beetle from a wasp from a butterfly, but even the entomologically erudite is left pondering which two insects are most closely related." The volume, edited by leading systematists Joel Cracraft (AMNH) and Michael Donoghue (Yale University), begins with three chapters that explain why assembling the Tree is important to science and society. Most of the reasons offered will be familiar to biologists, as the revolution in systematics has penetrated many different fields. Unfortunately, nearly all refer to the benefits of knowing the phylogeny for a particular group and say little about those benefits that can only come from assembling the entire Tree. In addition to revealing common patterns or coordinated evolution among clades, having the whole Tree should lead to more important but as yet unanticipated insights. For example, would Wegener have imagined continental drift if he had only a collection of road maps and no global map to work with? We biologists need our own globe.

Following the introductory section, 26 chapters by authorities on major branches (clades) summarize the state of our phylogenetic knowledge. These begin at the base of the Tree, where contributors highlight, for example, the recent recognition that the earliest branchings split life into three domains: the bacteria, archaea, and eukaryotes. The chapters then proceed up the Tree through smaller branches and less inclusive groups (e.g., green plants, animals, and arthropods) to consider such "crown" groups as flowering plants, annelid worms, and birds. In each chapter, the authors summarize the constituent subgroups and typically describe supporting evidence, regions of uncertain relationship, and definitive morphological features. Afterward, Donoghue and three other leading evolutionary biologists (Edward Wilson, David Wake, and David Hillis) offer short summary perspectives. In the final chapter, the editors tie everything together by assembling a 138-taxon synoptic tree.

Taken individually, the chapters are useful summaries of our current understanding, but they seem like disconnected limbs. Nonetheless, the Tree will start to assemble itself--an emergent property of the disconnected parts--in the minds of those readers who take the time to read far enough. In that indirect way, the editors have met their goal. In addition, even the most broadly trained comparative biologists will discover unappreciated diversity in less familiar groups and the kind of fascinating organisms that inspired many of us to become biologists. These benefits would have been even easier to appreciate if the material was presented in the more dynamic and immersive experience of the volume's Web analog, the Tree of Life project (http://tolweb.org/tree/). (It is a shame that updatable, peer-reviewed Web pages still lack the professional status of static book chapters.)

Most authors have taken their charge very seriously and have written unbiased, synthetic, and useful accounts. Particularly readable chapters include those on Holometabola (insects characterized by complete metamorphosis), land plants, and chordates (vertebrates, hagfish, lancelets, and tunicates). A minority of the contributors have yielded to provincialism, focusing on their own work or dismissing information (e.g., molecular) that they distrust. The most extreme position appears in the mammal chapter, whose authors eschew the summary format in favor of lecturing on their preferred systematic procedures. Only a handful of conflicting conclusions appear; one is the description of the Holometabola as a group whose monophyly is either routinely supported by both morphology and molecules (Whiting) or never supported in any molecular data (Rainer Willmann).

The volume's principal utility stems from its revelation of the patterns among diverse clades. Many authors cite the explosion of molecular data as the reason for the revolution in phylogenetics, especially for the field's transformation since the previous symposium that attempted to view phylogenetics across all of life (1), held in 1988. The most publicized cases of conflict between molecules and morphology are not representative of that revolution: The tidal surge of molecular data seems to have confirmed numerous old hypotheses while rejecting a few but, most importantly, resolving many branches that morphological evidence did not. One is struck by the great reliance on a single gene--the small subunit (SSU) of the ribosomal DNA, also known as 18S--for most resolution deep in the Tree, even within phyla. Elsewhere, despite frequent accolades to molecular data, the recognition of many clades (especially among chordates) continues to rely on morphology.

The other broad impression the volume leaves is that of an imbalance toward authors who favor parsimony for phylogenetic analysis over model-based or statistical methods such as likelihood. Individually, this imbalance is not very important because all chapters include authoritative authors. The reviews of findings by other researchers are generally fair, although occasionally conflicting model-based results are brushed aside--as in the treatment of the debate over the effects of long-branch attraction on analyses of the relationship between the fly orders Strepsiptera and Diptera (2). A more pervasive and subtle, yet profound, consequence of this methodological bias is omission from most chapters of fundamental aspects of evolutionary history, like timing of events and rates of diversification. The emphasis in the volume is entirely on the sequence of branching. (Branch lengths are not important in parsimony analysis, and their estimates are generally unreliable. In contrast, they are integral to model-based methods.) As a result, the tempo and mode of evolution (3) are lost, and we cannot see whether the Tree looks like a spreading oak, a willow, or a bamboo grove--we have little sense of its gestalt. The lack of resolution in some parts of the Tree is therefore attributed to a lack of data rather than to the much more interesting possibility of rapid diversification. Branch lengths--as indicators of time or amount of evolution--are important to almost every aspect of comparative biology, and the volume would have benefited from the more nuanced vision their consideration would have offered.

The summary chapters praise the progress and promise more to come. I would have preferred a more critical analysis of the overall state of this resource- limited field. Where are the biggest holes? Should we focus on broad taxonomic coverage of a few universal genes, overlapping sets of many genes, or perhaps new initiatives to train morphologists? But in the end, the big picture emerges from the details, and we gain a better appreciation of how the branches fit together and where some of the bigger questions remain. The vision Cracraft and Donoghue articulate in their introduction does emerge from the mist, incomplete though it may appear. Assembling the Tree of Life should also meet the editors' larger goal. It will help the systematic community aspire toward a common goal, identify priorities for future coordinated work, and mobilize our resources.

References and Notes

B. Fernholm, K. Bremer, H. Jörnvall, Eds., The Hierarchy of Life: Molecules and Morphology in Phylogenetic Analysis (Nobel Symposium 70, Elsevier, Amsterdam, 1989).

J. P. Huelsenbeck, Syst. Biol. 46, 69 (1997). 
G. G. Simpson, Tempo and Mode in Evolution (Columbia Univ. Press, New York, 1944).