0227Books 2/20/09 1:38 PM Page 1171

BOOKS ETAL.

ing the translation of Origin of Species into translations involved an attempt to recast exist- And so, finally, we come to Haeckel. German, and (more intriguingly) diachroni- ing German terms in a newer, more up-to-date Gliboff’s key insight here is that Haeckel cally, as scientists reworked older words such mode that encompassed selection yet tamed originally read Bronn’s translation of Darwin, as “perfection” and “type” to lend them new Darwin’s emphasis on unpredictability to meet not Darwin in the original. Gliboff shows meanings. Gliboff’s own clear, crisp prose is the more rigorous requirements of a German Haeckel as both echoing and responding to key to the success of this analysis, as he deftly academic scientist’s understanding of a “law” Bronn’s concerns, rather than either reflect- leads his reader through dense philosophical of organic nature. Simultaneously, Bronn ing directly on Darwin’s original writing or and terminological thickets with nary a thorn sought to translate Darwin’s ideas about selec- reaching directly back to the Romantic scratch. This is some of the best close reading tion into a language without an exact equiva- embryologists. (Although Gliboff acknowl- I have seen. It also represents a profound chal- lent for the term, and for an academic audience edges the centrality of monism to Haeckel’s lenge to our standard picture of 19th-century lacking the gentlemanly traditions of breeding thought, he focuses on the working evolution- German biology. pigeons and dogs so central to Darwin’s expo- ary theorist, not the popular ideologue.) Like The old story, crudely put, is that Haeckel’s sition. The selection metaphor was further Bronn himself, Haeckel made further amend- version of evolution was a Darwinism in name fraught with an anthropomorphism foreign to ments both terminological and intellectual, only, best understood as an update on early- Germans, who were not brought up on British and Gliboff rereads Haeckel’s research pro- 19th-century idealistic morphologists such natural-theological assumptions about a per- gram as one not dominated by a typological as Carl0403NewsFocus.qxp F. Kielmeyer and 3/30/09 J. F. Meckel 5:23 thatPM Pagesonified 29 God who had created a perfectly and linear-recapitulationist mindset but retained their teleology, their typological adapted nature. Bronn’s translation, though it rather as continuing to wrestle with the need emphasis on form, and their linear recapitula- altered key ideas to make Darwin comprehen- to account for variability and unpredictable tionism. This story, emphasizing the long per- sible to a German academic audience, was not change in terms of mechanistic laws of sistence of a German transcendental approach a conservative throwback. It represented the nature—among which Haeckel included, at to nature, has been deeply entrenched in the dynamic engagement of a leading paleontolo- the top of his list, ORIGINSnatural selection. Haeckel’s history of biology. gist who had also long been working on many DarwinismIntroduction thus shows continuity with early- Gliboff challenges this history right from of the questions Darwin claimed as his own— 19th-century concerns, mediated through Stockholm. “When we started, 2 Darwin’s Inspiration, Darwin’s Legacy the beginning.the The search ascription profile was of simplebigger, linear a critical yet generous equal, who saw himself Bronn. But those concerns were always more Andrew Sugden recapitulationisma magnolia to the [flower],” views of Romanticshe as Contentsmoving science forward through the modi- flexible than has been acknowledged, and embryologists,recalls. he notes, But 30 owes years much ago, to she a carica- fications he made to Darwin’s flawed theory. their articulation changed over time. Of ture developedand byothers Karl discovered Ernst von tinyBaer in a Bronn’s death in 1862 afforded him little course Haeckel’s Darwinism was not polemical context,ancient then flowers adopted by uncriticallysieving by chance to steer the conversation further. Darwin’sArticles own, but it was not an aberration or through sand and clay sedi- influential historiansments. With such this as E.technique, S. a distortion 3 ofOn some the Origintrue theory, of Life any on more Earth Russell and theyStephen have nowJay collectedGould. hun- than any otherCarl post-Darwinian Zimmer additions or Gliboff’s freshdreds reading of of millimeter-size the orig- adjustments were.Science It was 9 January science 2009 moving 323: 198-199 on. inal sources flowers,interprets some Kielmeyer preserved in Gliboff’s overall picture of scientific and Meckelthree as fardimensions, less rigidly from Por- advance, in 5 contrastOn the to OriginRichards’s of Art emphasis and Symbolism on typological intugal their and orientation other locations and with charisma andMichael passion, Balter is one of scientists much more Cretaceousattentive to deposits nature’s 70 mil- building and innovatingScience 6 February incrementally, 2009 323: work- 709-711 lion to 120 million years old. variability thanThis has fossil been seen diversity ing with what their predecessors have handed before. Bothshows for these that early-19th-angiosperms were them and sculpting8 On the it intoOrigin something of Photosynthesis new yet century naturaliststhriving, and with for several their groups understandable Mitch to those Leslie around them. His sen- Science 6 March 2009 323: 1286-1287 intellectual heirs,well-established, Gliboff argues, by 100 mil- sitive reading allows Out us of to the see past. post-1859 the critical issuelion was years to understandago. In some, the page 10 German evolutionists asTiny rational Amborella actorssits rather nature’s manifoldflower parts variety are whorled while like than irrationally10 On stucktheat Origin the in bottom some of ofFlowering early-19th-the Plants those of modern flowers; in angiosperm family tree. seeking out underlying strict natu- century momentElizabeth with Pennisi unmodern commit- others they are spiraled, con- Science 3 April 2009 324: 28-31 ral laws to accountsidered for by it. some researchers ments. By challenging the very foundations of This provides a new starting as the more primitive arrangement. Some fromthe standard one 14of the narrativeAlexander nonflowering of vonGerman seed Humboldt plants morphol- and the General Physics point for analyzing Darwin’s first “We are realizing that this flower fossils have prescribed numbers of orogy, gymnosperms, this careful,of thewhosecompelling Earth heyday account was 200 does at petals, another modern feature, whereas in million years ago. Modern gymnosperms translator, the prominent paleon- least as much Stephenas Richards’s T. Jackson to undermine the others the petal count varies. huge diversity is probably include conifers, ginkgoes, and the cycads, tologist H. G. Bronn—a figure lit- association of Science19th-century 1 May 2009German 324: Darwin- 596-597 tle attended toIn in 1998, the Chinese standard geologist Ge Sun of withism withtheir astout dangerously trunks and exceptional large fronds. view of Jilin University in Changchun, China, came the result of one innova- Before angiosperms came along, these story but the lynchpin of Gli- nature. But16 theMaking two books German offer very Evolution: different Translation and Tragedy

(W. KOEHLER, GERA-UNTERMHAUS, 1905) GERA-UNTERMHAUS, KOEHLER, (W. across what seemed to be a much older plants were much more diverse and boff’s. Intriguinglyflower. Theand fossil, plausibly, called Archaefructus, was tion piled on top of includedreads. Is cycadlikescientificLynn K.species,progress Nyhart such a matter as the of per- Gliboff arguesan aquatic that Bronn’s plant that use looked to be 144 mil- another innovation.” extinctsonal anguish Bennettitales, Scienceand triumph, 27and February many or of woody2009 intellectualREPORTS 323: 1170-1171 of terms likelion “vervollkommnet” years old. By 2002, Sun and David plantschugging called along? Gnetales,Our concept of of whichit should be Dilcherpresence of ofthe two Florida lineages Museum that co-occur of Natural in the ad- 6.2% divergence). In contrast,—Peter sites Crane, located a fewpopulation representatives, expansion (23 )including are found the in the

WANDERBILDER (perfect) as translations for Dar- capacious18 enough Darwin’s to include Originality both. Historyjacent refugia.(FLMNH) In all species,in Gainesville average net had nucle- outside (south of)University the refugia of are Chicago genetically jointunstable firs, area survive for H. today albomarginatus (see familyand H. tree, faber, win’s “improved”otide differencesor “favored” across localities (22) reflects more similar to each other, although to a lesser as well as in thePeter Bahia J. refugiumBowler area for H. faber were not aboutdescribed dragging an entire Darwin plant, from roots to flow- p. 31).References Also common andScience Notes in 9 theJanuary Jurassic 2009 were 323: 223-226 ers,high entombed geographic on a structure slab of rock within unearthed refugia (2.6 in to extentThese infossilsH. faber often(0.1 spark to 1.6%). debate Signatures because of seedand ferns,H. semilineatus a group now. The long lack gone; of signature their of backward into a German teleo- 1. E. Haeckel, Generelle Morphologie der Organismen Liaoning in northeastern China. page specimens16 tend to be imperfectly preserved most(Georg famous Reimer, member Berlin, 1866). is Caytonia, which logical view of nature (as has 22 The Red Queen and the Court Jester: Species Diversity Fig.In one 2. sense,Genetic Archaefructus diversity in putativewasn’t muchA and leave room for interpretation.B To help seems2. The to reviewer haveC previously precarpel-like served as a structures. press reader for both been claimedto by refugiallook those at. (stable) who“It’s versushavea flowering unstable plant areas before remedy that, Friis and her colleagues have Thesebooks groups’ at the manuscriptand perceived the stage. Role relevance of Biotic toand Abiotic Factors Through Time paid attentionthere into were theBronn Brazilian flowers,” at all). Atlantic DilcherA rainforest. painter, notes. It too. lackedHaeckel’sbegun oil landscapeto examine of flowershighlands using in Java, synchro- from flower evolution Michael and their J. Benton relationships to petals(Top) and Species-specific sepals, but stability it did maps; have an tron radiation to generate a 3D image of angiosperms haveScience ping-ponged 6 February10.1126/science.1169621 between 2009 323: 728-732 CREDIT: ERNST HAECKEL/FROM ERNST CREDIT: Instead, Gliboff asserts, Bronn’s Wanderbilder (1905). enclosedmodeled carpel. refugia When in black. Kevin (A )NixonH. and their inner structures, allowing the fossil to camps, depending on how the evolutionaryPernambuco refugium colleaguesalbomarginatus at Cornell,(B) H. University semilineatus compared, remain intact while Friis peers inside it trees were26 constructed. Evidence for Ecological Speciation and Its Alternative (C) H. faber. Note the absencewww.sciencemag.org of large SCIENCE VOL 323 27 FEBRUARY 2009 1171 its stabletraits regionswith those in the same southern traits portion in 173 living from many angles (Science, 7 December In the mid-1980s,Dolph SchluterPeter Crane,Bahia now refugium at plants,of the Archaefructus forest (south ofcame the Bahia out as and a sister to *2007, p. 1546). “We can get fantastic reso- the University * ofScience Chicago 6 February in Illinois, 2009 pro- 323: 737-741 livingSão angiosperms Paulo refugia) and relative closer to to the the com- lution,” says Friis. “It’s really exciting.” But posed a solution, the anthophyte hypothesis. moncentral ancestor and northernthan even areas. Amborella Asterisks. so far, the flowers Friis finds are Using several lines of evidence and noting 30 The BacterialSão Paulo Speciesrefugium Challenge: Making Sense of Genetic denoteArchaefructus refugia inferred’s distinction beyond was the short- too diverse to trace back to a that both Bennettitales and and Ecological Diversity lived,current however. ranges Within of the target months, species. better dat- particular5.4% ancestor. “From Gnetales organize their male ingSymbols of the indicatesediments localities in which sampled it was for found these fossils, we cannot andChristophe female organs Fraser, togetherEric J. Alm, Martin F. Polz, et al. yieldedmolecular younger analysis. dates, Scalebar, putting 400 km. this first say what is the basic Sciencein what 6 February could be 2009 con- 323: 741-746 flower(Bottom) squarely The 50% with majority-rule other early con- fossil form,” she says. strued as a preflower, sensus Bayesian phylogenetic trees, flower parts, about 125 million years old. 7% 36 Stabilityhe considered Predicts Geneticthem, Diversity in the Brazilian Atlantic rooted with sequences from the oth- 7.8% Also,er two a 2009 congeneric phylogenetic species studied analysis of Before flowers Forestalong Hotspot with angiosperms, 67 taxa by Doyle and Peter Endress of the Although they have yet as comprising a single (root not shown). Thick internodes de- page 36 5.3 – Ana Carolina Carnaval, Michael J. Hickerson, Célio F. B. Haddad, et al. Universitynote clades of with Zurich, posterior Switzerland, probability placed to find the oldest fossil Larger than 5.8%life. Although merely Scienceangiosperm 6 February entity4% 2009 called 323: 785-789 thegreater fossil than in with 90%. water Percentages liliesindicate rather than at flowers, researchers 2.2 millimeters in diameter, this 3D anthophytes. For the next Around the world, governments turn to AAAS as an objective, multidisciplinary scientific authority totheTamura-Nei base of the corrected angiosperms, distances between although this assume that the ances- fossil flower shows that grasses date decade, most family trees

CREDITS (TOP TO BOTTOM): COURTESY OF STEPHEN MCCABE, UC SANTA CRUZ; PHOTO BY JENNIFER SVITKO, COURTESY OF WILLIAM L. CREPET, CORNELL UNIVERSITY CORNELL CREPET, L. WILLIAM OF COURTESY SVITKO, JENNIFER BY PHOTO CRUZ; SANTA UC MCCABE, STEPHEN OF COURTESY BOTTOM): TO (TOP CREDITS conclusionclades (20 ).is contested. tral angiosperm evolved back to 94 million years ago. based on morphology sup- Science funding educate public officials and judicial figures on today’s most pressing issues. Our goal is to promote 5.6% Cover image: ©Tui De Roy/Minden Pictures/FLPA Climate regulation informed policy decisions that benefit society. And this is just one of the ways that AAAS is committed to Inset: George Richmond/Bridgeman Art Library, London (Superstock) www.sciencemag.org SCIENCE VOL 324 3 APRIL 2009 29 Human rights advancing science to support a healthy and prosperous world. Join us. Together we can make a difference. aaas.org/plusyou/policy

© 2009 by The American Association for the Advancement of Science. All rights reserved.

Table 1. Population genetic summary metrics used in model validation. n, q, and average Da values of the former were obtained not only from the total Sample size; S, number of segregating sites. The diversity parameter q and mean number of samples, but also from all possible combinations of spatially Da across localities are given per base pair (bp). Hs test (23) is used to detect contiguous localities distributed within the geographic extension of the unstable population expansion. BA, Bahia; SP, São Paulo refugia. Because predicted area. Parentheses encompass minimum and maximum values from subsamples. refugia were often larger than predicted unstable (recently colonized) areas, n, S, P values in bold highlight statistical significance at 0.05 probability level.

n S q Mean D Hs Mantel’s corr. coef. Species Area a (min.; max.) (min.; max.) (min.; max.) (min.; max.) (P value) (P value) H. albomarginatus Stable (BA) 36 207 0.076 0.062 –20.546 0.499 (970 bp) (13; 23) (81; 155) (0.034; 0.072) (0.020; 0.082) (0.141) (0.001) Unstable 27 22 0.003 0.001 –11.498 –0.140 (south of BA) (0.004) (0.580) H. semilineatus Stable (BA) 28 71 0.031 0.036 –17.778 0.054 (718 bp) (6; 13) (14; 58) (0.009; 0.034) (0.007; 0.041) (0.029) (0.460) Unstable 15 9 0.003 0.004 0.114 0.436 (south of BA) (0.357) (0.248) H. faber Stable (BA) 28 94 0.018 0.026 –38.111 0.803 (771 bp) (13; 23) (42; 80) (0.012; 0.022) (0.001; 0.044) (0.003) (0.0003) Stable (SP) 15 48 0.023 0.028 –5.981 0.305 (0.115) (0.221) Unstable 18 40 0.015 0.016 –13.255 0.0001 (south of SP) (0.014) (0.456)

www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 787 0109NewsFocus.qxp 1/5/09 6:38 PM Page 198

NEWSFOCUS

and proteins. Later, RNA-based life may have evolved the ability to assemble amino acids into Introduction OnOn thethe OriginOrigin ofof proteins. It’s a small step, biochemically, for DNA to evolve from RNA. In modern cells, RNA is remarkably versa- tile. It can sense the levels of various com- Darwin’s Inspiration, Darwin’s Legacy Life on Earth pounds inside a cell and switch genes on and off to adjust these concentrations, for example. The day has passed delightfully. Delight itself, however, is a weak term to express in some warm little pond, with all sorts of It can also join together amino acids, the build- ammonia and phosphoric salts, light, heat, elec- ing blocks of proteins. Thus, the first cells the feeling of a naturalist who, for the first time, has wandered by himself in a tricity, etc., present, that a protein compound might have tapped RNA for all the tasks on Brazilian forest. The elegance of the grasses, the novelty of the parasitical plants, was chemically formed ready to undergo still which life depends. the beauty of the flowers, the glossy green of the foliage, but above all the general more complex changes, at the present day such For 60 years, researchers have been honing matter would be instantly devoured or theories about the sources of the amino acids luxuriance of the vegetation, filled me with admiration. A most paradoxical mixture absorbed, which would not have been the case and RNA’s building blocks. Over time, they of sound and silence pervades the shady parts of the wood. The noise from the before living creatures were formed.” have had to refine their ideas to take into Scientists today who study the origin of life account an ever-clearer understanding of what insects is so loud, that it may be heard even in a vessel anchored several hundred do not share Darwin’s pessimism about our early Earth was like. yards from the shore; yet within the recesses of the forest a universal silence appears ability to reconstruct those early moments. In an iconic experiment in 1953, Stanley to reign. To a person fond of natural history, such a day as this brings with it a “Now is a good time to be doing this research, Miller, then at the University of Chicago, because the prospects for success are greater ignited a spark that zapped through a chamber deeper pleasure than he can ever hope to experience again. than they have ever been,” says John filled with ammonia, methane, and other —Charles Darwin, The Voyage of the Beagle, Feb 29th [1832] Sutherland, a chemist at the University of Man- gases. The spark created a goo rich in amino chester in the United Kingdom. He and others acids, and, based on his results, Miller sug- are addressing each of the steps involved in the gested that lightning on the early Earth could transition to life: where the raw materials came have created many compounds that would ike many other scientists raised in temperate latitudes, Charles examines the extent to which biotic and abiotic factors have shaped from, how complex organic molecules such as later be assembled into living things. Darwin was enthralled by his first glimpse of the tropical rain species diversity in the fossil record. Dolph Schluter reviews how RNA formed, and how the first cells arose. In By the 1990s, however, the accumulated AN AMAZON OF WORDS FLOWED FROM doing so, they are inching their way toward evidence indicated that the early Earth was forest. His Beagle diary entry conveyed those immediate and research on speciation has shifted in focus from morphological L Charles Darwin’s pen. His books covered the making life from scratch. “When I was in grad- dominated by carbon dioxide, with a pinch of thrilling first impressions, but the encounter with the Brazilian Atlan- evolution to reproductive isolation, tracing the links between Dar- gamut from barnacles to orchids, from geol- uate school, people thought investigating the nitrogen—two gases not found in Miller’s tic Forest had an enduring influence on the development of his ideas win’s ideas and current thinking. Christophe Fraser and colleagues ogy to domestication. At the same time, he origin of life was something old scientists did at flask. When scientists tried to replicate Miller’s over the following decades, with resounding echoes even today in discuss the contentious area of microbial species formation, an is- filled notebooks with his ruminations and the end of their career, when they could sit in an experiments with carbon dioxide in the mix, 21st century evolutionary science. sue that would surely have vexed Darwin horribly had the bewil- scribbled thousands of letters packed with armchair and speculate,” says Henderson their sparks seemed to make almost no amino dering diversity of microbes been known in his day. observations and speculations on nature. Yet James Cleaves of the Carnegie Institution for acids. The raw materials for life would have The collection reprinted here is a sample of the articles published Darwin dedicated only a few words of his great Science in Washington, D.C. “Now making an had to come from elsewhere, they concluded. in 2009 by Science magazine in celebration of the Darwin bicen- Finally, with a focus on conservation, a Report by Ana Carnaval et verbal flood to one of the biggest questions in artificial cell doesn’t sound like science fiction In 2008, however, lightning began to look tenary. We start with four of the essays from our “On the Origin al., who model evolutionary processes in endemic tree-frog spe- all of biology: how life began. any more. It’s a reasonable pursuit.” promising once again. Cleaves and his col- of” series, prepared by Science’s news writers; further essays in this cies in the Brazilian Atlantic Forest, the very biodiversity hotspot The only words he published in a book leagues suspected that the failed experiments series are appearing monthly in Science throughout the year. A Per- that so inspired Darwin on his South American landfall, and that is appeared near the end of On the Origin of Raw ingredients were flawed because the sparks might have pro- Species: “Probably all the organic beings which Life—or at least life as we know it—appears to duced nitrogen compounds that destroyed any spective by Stephen Jackson then considers the legacy of Alexander now reduced to a collection of small fragments scattered along the have ever lived on this earth have descended have emerged on Earth only once. Just about all newly formed amino acids. When they added von Humboldt, for whom, like Darwin, the South American tropics coast. Darwin returned to the Brazilian coast on his final homeward from some one primordial form, into which life organisms use double-stranded DNA to encode buffering chemicals that could take up these were a critical inspiration, and who died 150 years ago in the year leg, more than four years after his first landfall there. His enthusi- was first breathed,” Darwin wrote. genetic information, for example. They copy nitrogen compounds, the experiments gener- of the publication of Darwin’s Origin. (Humboldt’s Personal Nar- asm for the tropical forested landscape was undiminished. Darwin believed that life likely emerged their genes into RNA and then translate RNA ated hundreds of times more amino acids than rative of his tropical explorations was acknowledged by Darwin as spontaneously from the chemicals into proteins. The genetic code scientists had previously found. ‘far exceed[ing] in merit anything I have read’ on the subject.) A In my last walk I stopped again and again to gaze on these beau- it is made of today, such as carbon, THE YEAR OF they use to translate DNA into pro- Cleaves suspects that lightning was only book review by Lynn Nyhart explores two recent volumes on Ernst ties, and endeavoured to fix in my mind for ever, an impression nitrogen, and phosphorus. But he The English teins is identical, whether they are one of several ways in which organic com- did not publish these musings. emus or bread mold. The simplest pounds built up on Earth. Meteorites that fall to Haeckel’s work, his interpretations of Darwin and his contributions which at the time I knew sooner or later must fail … they will leave, DARWINnaturalist had to evolutionary thought. like a tale heard in childhood, a picture full of indistinct, but most The English naturalist had built explanation for this shared biology Earth contain amino acids and organic carbon beautiful figures. his argument for evolution, in built his argu- is that all living things inherited it molecules such as formaldehyde. Hydro- In the first of four Review articles reprinted here, Peter Bowler ana- —Charles Darwin, The Voyage of the Beagle, August 1836 large part, on the processes he ment for evo- from a common ancestor— thermal vents spew out other compounds that could observe around him. He did namely, DNA-based microbes that could have been incorporated into the first life lyzes the originality of Darwin’s contribution to the understanding of not think it would be possible to lution, in large lived more than 3.5 billion years forms. Raw materials were not an issue, he the diversity and diversification of the living world. Michael Benton Andrew Sugden, Deputy Editor see life originating now because part, on the ago. That common ancestor was says: “The real hurdle is how you put together the life that’s already here would already fairly complex, and many organic compounds into a living system.” prevent it from emerging. processes he scientists have wondered how it In 1871, he outlined the prob- Thiscould essay isobserve the first might have evolved from a simpler Step 1: Make RNA lem in a letter to his friend, botanist in a monthly series, with predecessor. Some now argue that An RNA molecule is a chain of linked morearound on evolutionary him.. Joseph Hooker: “But if (and Oh! roots online at blogs. membrane-bound cells with only nucleotides. Each nucleotide in turn consists ; GEORGE RICHMOND/BRIDGEMAN ART LIBRARY, LONDON (SUPERSTOCK) LONDON LIBRARY, ART RICHMOND/BRIDGEMAN GEORGE SUTLIFF/ SCIENCE ; KATHARINE BOTTOM): TO (TOP CREDITS CREDIT: KATHARINE SUTLIFF/ SCIENCE KATHARINE CREDIT: what a big if!) we could conceive sciencemag.org/origins RNA inside predated both DNA of three parts: a base (which functions as a

2 198 9 JANUARY 2009 VOL 323 SCIENCE www.sciencemag.org 3 0109NewsFocus.qxp 1/5/09 6:38 PM Page 199 0109NewsFocus.qxp0206NewsFocus.qxp 1/5/09 6:38 2/2/09 PM Page 7:13 199 PM Page 709

EVOLUTIONARY ROOTS EVOLUTIONARY ROOTSORIGINS

“letter” in a gene’s recipe), a sugar molecule, soup. “We’ve got the molecules in our RNA, producing the first protocells. “The goal “letter” in a gene’s recipe), a sugar molecule, soup. “We’ve got the molecules in our RNA, producingVenus, the first phallus, protocells. or “Thepebble? goal and a cluster of phosphorus and oxygen sights,” he says. is to have something that can replicate by itself, and a cluster ofOnOn phosphorus thethe and oxygenOriginOriginsights,” he of says.of is to have something“I don’t that know can replicate much about by itself, Art, but I know what atoms, which link one sugar to the next. For Sutherland can’t say for sure where these using just chemistry,” says Szostak. atoms, which link one sugar to the next. For Sutherland can’t say for sure where these using just chemistry,”I like,” says quipped Szostak. the humorist and art critic years, researchers have tried in vain to synthe- reactions took place on the early Earth, but he After 2 decades, he and his colleagues years, researchers have tried in vain to synthe- reactions took place on the early Earth, but he After 2 decades,Gelett he Burgess and his back colleagues in 1906. For archaeo- size RNA by producing sugars and bases, notes that they work well at the temperatures have come up with RNA molecules that can size RNA by producing sugars and bases, notes that they work well at the temperatures have come up withlogists, RNA distinguishing molecules that art canfrom nonart is still joining them together, and then adding phos- and pH levels found in ponds. If those ponds build copies of other short RNA molecules. joining them together,Art and then addingand phos- and pHSymbolism levels found in ponds. If those ponds build copies of quiteother a short challenge. RNA Takemolecules. the 6-centimeter-long phates. “It just doesn’t work,” says Sutherland. dried up temporarily, They have been able to phates. “It just doesn’t work,” says Sutherland. dried up temporarily, piece Theyof quartzite have been known able toas the Venus of This failure has led scientists to consider they would concentrate mix RNA and fatty This failure has led scientists to consider they would concentrate “Now making an Tan-Tan.mix Found RNA in Moroccoand fatty in 1999 next to a two other hypotheses about how RNA came to the nucleotides, making “Now making an acids together in such a two other hypotheses about how RNA came to the nucleotides, communicatemaking meaning, whether they be the rich troveacids oftogether stone in tools such estimated a to be be. Cleaves and others think RNA-based life conditions for life even artificial cell doesn’t way that the RNA gets be. Cleaves and others think RNA-based life conditions for lifewords even that makeartificial up our languages, cell thedoesn’t musi- betweenway 300,000 that the and RNA 500,000 gets years old, it may have evolved from organisms that used a more favorable. trapped in vesicles. The may have evolved from organisms that used a more favorable. cal sounds that convey emotion, or the dra- resemblestrapped a human in vesicles. figure The with stubby arms different genetic material—one no longer Were these Darwin’s sound like science vesicles are able to add different genetic material—one no longer Were these Darwin’smatic paintingssound that, 30,000 like years science after their and legs.vesicles Robert are Bednarik,able to add an independent found in nature. Chemists have been able to warm little ponds? “It fatty acids to their found in nature. Chemists have been able to warm little ponds?creation, “It caused the discoverers of the Chau- archaeologistfatty acids based to in their Caulfield South, use other compounds to build backbones for might just be that he fiction any more. It’s membranes and grow. use other compounds to build backbones for might just be vetthat Cave he to breakfiction down inany tears. more. It’sAustralia, membranes insists and that grow. an ancient human nucleotides (Science, 17 November 2000, wasn’t too far off,” says In July 2008, Szostak nucleotides (Science, 17 November 2000, wasn’t too far off,” Whilesays sites like Chauvet might be vivid deliberatelyIn July modified 2008, Szostak the stone to p. 1306). They’re now investigating whether Sutherland. a reasonable pursuit.” reported that he had p. 1306). They’re now investigating whether Sutherland. examples ofa whatreasonable some researchers pursuit.” still make itreported look more that like he a person. had these humanmade genetic molecules, called figured out how proto- these humanmade genetic molecules, called consider a “creative explosion” that began If so, thisfigured objet out d’art how is soproto- old PNA and TNA, could have emerged on their Step 2: The cell —HENDERSON JAMES CLEAVES, cells could “eat” and PNA and TNA, could have emerged on their Step 2: The cellwhen modern humans —HENDERSONcolonized Europe JAMES CLEAVES,that it wascells created could not “eat” by our and own on the early Earth more easily than RNA. If life did start out with CARNEGIE INSTITUTION FOR SCIENCE bring in nucleotides to own on the early Earth more easily than RNA. If life did start outabout with 40,000 years ago,CARNEGIE an increasing INSTITUTION num- FOR SCIENCEown species,bring in whichnucleotides first to According to this hypothesis, RNA evolved RNA alone, that RNA build the RNA. According to this hypothesis, RNA evolved RNA alone, thatber RNA of prehistorians are tracing our sym- appearsbuild in Africa the RNA. nearly later and replaced the earlier molecule. would need to make copies of itself without All living cells depend on complicated later and replaced the earlier molecule. would need to makebolic rootscopies much of itself further without back in time—andAll living cells200,000 depend years on ago, complicated but But it could also be that RNA wasn’t put help from proteins. Online in Science this channels to draw nucleotides across their But it could also be that RNA wasn’t put help from proteins.in some Online cases, in to Science species thisancestralchannels to Homo to drawby onenucleotides of our ances- across their together the way scientists have thought. “If week (www.sciencemag.org/cgi/content/ membranes, raising the question of how a together the way scientists have thought. “If week (www.sciencemag.org/cgi/content/sapiens. Like modern humans themselves,membranes, raisingtors, the perhaps question the of how a you want to get from Boston to New York, abstract/1167856), Tracey Lincoln and Ger- primitive protocell membrane brought in these you want to get from Boston to New York, abstract/1167856),symbolic Tracey behavior Lincoln seems and Ger- to haveprimitive its origins protocelllarge-brained membrane broughtH. in these there is an obvious way to go. But if you can’t ald Joyce of the Scripps Research Institute in molecules. By experimenting with different there is an obvious way to go. But if you can’t ald Joyce of thein Scripps Africa. Research Recent Instituteexcavations in havemolecules. turned By experimentingheidelbergensis with, different get there that way, there are other ways you San Diego, California, have shown how that recipes for membranes, Szostak and his col- get there that way, there are other ways you San Diego, California,up elaborate have shownstone tools,how that beads,recipes and ochre for membranes,thought bySzostak some and his col- could go,” says Sutherland. He and his col- might have been possible. They designed a leagues have come up with protocells leaky could go,” says Sutherland. He and his col- might have beendating possible. back They 100,000 designed or more a leagues years ago, have comeanthropologists up with protocells to leaky leagues have been trying to build RNA from pair of RNA molecules that join together and enough to let nucleic acids slip inside, where leagues have been trying to build RNA from pair of RNA moleculesalthough that researchers join together are and stillenough debating to let nucleicbe the acids common slip inside, where simple organic compounds, such as formalde- assemble loose nucleotides to match their they could be assembled into RNA, but not so simple organic compounds,Since their discoverysuch as formalde- by Frenchassemble spelunkers loosewhich nucleotides of these to findsmatch reallytheir demonstratethey could be assembledancestor intoof mod- RNA, but not so hyde, that existed on Earth before life began. partner. Once the replication is complete, old porous that the large RNA could slip out. hyde, that existedin on1994, Earth the before magnificent life began. lions, partner.horses, Onceand thesymbolic replication expression. is complete, But old there’sporous wide- that the largeern humansRNA could and slip out. They find they make better progress toward and new RNA molecules separate and join Their experiments also show that these They find they rhinosmake better that seem progress to leap toward from theand walls new RNAof spreadmolecules agreement separate that and the join buildingTheir blocks experimentsNeandertals. also showThat that these producing RNA if they combine the compo- with new partners to form new RNA. In 30 vesicles survive over a 100°C range. At high producing RNAChauvet if they combine Cave in the southern compo- Francewith new have partnersof symbolism to form new precededRNA. In 30 full-blownvesicles art.survivewould over a 100°Cmean thatrange. At high nents of sugars and the components of bases hours, Lincoln and Joyce found, a population temperatures, protocells take in nucleotides nents of sugars reignedand the components as the world’s of bases oldest cavehours, paint- Lincoln and“When Joyce we found, talk about a population beads andtemperatures, art, we are protocellsart is an extremelytake in nucleotides together instead of separately making com- of RNA molecules could grow 100 million quickly, and at lower temperatures, Szostak together insteadings. of separately Expertly makingcomposed com- in redof ochre RNA andmoleculesactually could talking grow about 100 million material technologiesquickly, and at ancientlower temperatures, part of the Homo Szostak plete sugars and bases first. times bigger. found, they build RNA molecules faster. plete sugars andblack bases charcoal,first. the vivid drawingstimes demon- bigger. for symbolic expression that certainlyfound, theypost- buildrepertoire. RNA molecules “Ignoring faster. the Symmetry in stone. Over the past few years, they have docu- Lincoln and Joyce kept their RNA mole- He speculates that regular temperature Over the paststrate few years,that the they artistic have docu- gift stretchesLincoln back anddate Joyce the kept origins their RNAof symbolic mole- thoughtHe speculates and few that specimens regular we temperature have Some stone tools mented almost an entire route from prebiotic cules in beakers. On the early Earth, however, cycles could have helped simple protocells sur- mented almost anmore entire than route 30,000 from years.prebiotic Thesecules paintings in beakers.communication, On the early Earth, potentially however, by acycles very couldwide haveof helped very early simple paleoart, protocells sur-require a mental molecules to RNA and are preparing to pub- replicating RNA might have been packed in the vive on the early Earth. They could draw in molecules to RNAare almostand are surepreparing to be mentionedto pub- replicating in any arti- RNAmargin,” might have says been archaeologist packed in the Dietrichvive on Stout the earlyexplaining Earth. They them could away, drawimage in to create. lish even more details of their success. Dis- first cells. Jack Szostak and his colleagues at nucleotides when they were warm and then use lish even more cledetails or paper of their about success. the earliest Dis- art.first But cells. what Jack ofSzostak University and his College colleagues London. at nucleotides whenor they rejecting were warm them and out then use covering these new reactions makes Suther- Harvard Medical School in Boston have been them to build RNA when the temperature covering these newdo they reactions really makes tell us Suther- about theHarvard origins Medical of SchoolThe evolution in Boston ofhave symbolism been them was to once build RNAof hand when does the not temperature serve this discipline well,” land suspect it wouldn’t have been that hard investigating how fatty acids and other mole- dropped. In Szostak’s protocells, nucleotides land suspect it wouldn’tartistic expression? have been that hard investigating howthought fatty acidsto have and been other as mole-rapid as dropped.“flicking Inon Szostak’sBednarik protocells, wrote innucleotides a 2003 analysis of the for RNA to emerge directly from an organic cules on the early Earth might have trapped are arranged along a template of RNA. Strands for RNA to emergeThe directly prehistoric from an humans organic whocules decorated on the earlya light Earth switch,” might as have archaeologist trapped Cliveare arranged Gamble alongVenus a template of Tan-Tan of RNA. in Current Strands Anthropology. of RNA tend to stick together at low tempera- Chauvet’s walls by torchlight arrived at the of the Royal Holloway, Universityof RNA of Lon- tend to stickYet together many archaeologistsat low tempera- are skeptical, tures. When the protocell warmed up again, the cave with their artistic genius already in full don, put it some years ago. Buttures. given When new the protocellarguing warmedthat the up stone’s again, resemblancethe to a heat might cause the two strands to pull apart, flower. And so, most researchers agree that evidence that symbolic behaviorheat appears might causehuman the two figure strands might to pull be apart,coincidence. Indeed, allowing the new RNA molecule to function. the origins of art cannot simply long before cave allowingpaintings, the newthe RNA debate molecule over the to Tan-Tanfunction. “figurine” is rem- Now Szostak is running experiments to be pegged to the latest discovery THE YEAR OF Gamble now says that hisNow much- Szostakiniscent is running of a experiments similar controversy to over a bring his protocells closer to life. He is devel- of ancient paintings or sculp- Recent cited comment needsbring to be his mod- protocellssmaller closer stone to life. discovered He is devel- in 1981 at the site of oping new forms of RNA that may be able to ture. Some of the earliest art DARWINexcavations ified: “It’s a dimmer opingswitch new now, formsBerekhat of RNA Ram that mayin the be Israeli-occupied able to Golan replicate longer molecules faster. For him, likely perished over the ages; a stuttering candle.” replicate longerHeights. molecules To faster. some For archaeologists, him, this the true test of his experiments will be much remains to be found; and have turned As they more preciselythe true pin- test of250,000-year-old his experiments object will resembles be a woman, whether his protocells not only grow and archaeologists don’t always up elaborate point when symbolicwhether behavior his protocellsbut others not argue only that grow it was and shaped by natural reproduce, but evolve. agree on how to interpret what is stone tools, began, scientists are reproduce,hoping they but evolve.forces, and, in any case, looks more like a “To me, the origin of life and the origin of unearthed. As a result, instead of might one day crack the“To tough- me, the penguinorigin of or life a phallus.and the originEven after of an exhaustive Darwinian evolution are essentially the same chasing after art’s first appear- beads, and est question of all: WhatDarwinian was its evolutionmicroscopic are essentially study the concluded same that the thing,” says Szostak. And if Darwin was alive ance, many researchers seek to ochre dating evolutionary advantagething,” says to Szostak.Berekhat And Ram if Darwin object was had alive indeed been etched Protocell. Researchers at today, he might well be willing to write a lot understand its symbolic roots. back 100,000Protocell. humans?Researchers Did at symbols,today, as manyhe mightwith well abe tool willing to emphasize to write whata lot some consider Harvard are trying to make more about how life began. After all, art is an aesthetic This essay continuesHarvard our areresearchers trying to make suspect,more serve about as a howits life “head” began. and “arms,” many researchers have simple life forms, shown –CARL ZIMMER expression of something more monthlyor series. more See moresimple sociallife forms, glue shown that helped tribes of rejected it as a–CARL work of ZIMMER art. For some, proof of here in a computer image. on humans’ evolutionaryhere in a computer image. Carl Zimmer is the author of Microcosm: E. coli and the Carl Zimmer is the author of Microcosm: E. coli and the fundamental: the cognitive abil- journeyyears online ago. at blogs. early humans to survive and symbolic behavior requires evidence that the CREDIT: JANET IWASA JANET CREDIT: CREDITS (LEFT TO RIGHT): KATHERINE SUTLIFF/ SCIENCE; THE BOXGROVE PROJECT CREDITS (LEFT TO RIGHT): KATHERINE CREDIT: JANET IWASA JANET CREDIT: New Science of Life. (SUPERSTOCK) LONDON LIBRARY, ART RICHMOND/BRIDGEMAN GEORGE PROJECT; BOXGROVE THE SUTLIFF/ SCIENCE ; KATHARINE BOTTOM): TO (TOP CREDITS ity to construct symbols that sciencemag.org/origins. reproduce? New Science of Lifesymbols. had a commonly understood mean-

4 www.sciencemag.org SCIENCE VOL 323 9 JANUARY 2009 199 www.sciencemag.orgwww.sciencemag.orgSCIENCE VOLSCIENCE 323 9 JANUARYVOL 323 2009 6 FEBRUARY 2009 199 709 5 0206NewsFocus.qxp 2/2/09 7:13 PM Page 710 0206NewsFocus.qxp 2/2/09 7:13 PM Page 711

ORIGINS ORIGINS

tools, which Wynn and many others argue are Late Acheulean tools. Both methods turned ated with human burials. Some researchers A roaring start. Researchers agree that Chauvet Cave’s clear examples of an imposed form based on a on visual and motor areas of the brain. But argue that this represents an early case of magnificent paintings, including these lions, are full-blown art. mental template. Some have even argued that only Late Acheulean knapping turned on cir- “color symbolism,” citing the universal these skillfully crafted hand axes had symbolic cuits also linked to language, the team importance of red in historical cultures meanings, for example to display prestige or reported last year. worldwide and the apparently great lengths even attract members of the opposite sex. to which early humans went to gather red The half-million-year mark also heralded Color me red ochre. ”There is very strong circumstantial the arrival of H. heidelbergensis, which had At Twin Rivers, it’s not just the tools that hint evidence for the very great antiquity of the a much larger brain than H. erectus. Not at incipient symbolic behavior. Early humans color red as a symbolic category,” says long afterward, our African ancestors began there also left behind at least 300 lumps of anthropologist Sally McBrearty of the to create a wide variety of finely crafted ochre and other pigments in a rainbow of col- University of Connecticut, Storrs. blades and projectile points, which allowed ors: yellow, red, pink, brown, purple, and These finds of colorful ochre, fancy them to exploit their environment in more blue-black, some of which were gathered far tools, and beads have convinced many sophisticated ways, and so presumably from the site. Excavator Lawrence Barham researchers that the building blocks of enhance their survival and reproduction. of the University of Liverpool in the United symbolism had emerged by at least Archaeologists refer to these tools as Middle Kingdom thinks they used the ochre to 100,000 years ago and possibly Stone Age technology and agree that they paint their bodies, though there’s little much earlier. But why? What did require mental templates. “The tools tell hard evidence for this. Most archaeolo- selective advantages did using us that the hominid world was changing,” gists agree that body painting, as well symbols confer on our ancestors? says Wynn. as the wearing of personal orna- To some scientists, the ques- ing and were shared within groups of people. ity to hold an abstract concept in one’s head— As one moves forward in time, humans ments such as bead necklaces, tion is a no-brainer, especially For example, the hundreds of bone and stone and, in the case of the tool, to “impose” a pre- appear able to imagine and create even more was a key way that early humans when it is focused on the most “Venus figurines” found at sites across Eura- determined form on raw material based on an elaborate tools, sharpening their evolution- symbolically communicated sophisticated form of symbolic sia beginning about 30,000 years ago were abstract mental template. That kind of ability ary edge in the battle for survival. By social identity such as member- communication: language. The skillfully carved and follow a common motif. was probably not needed to make the earliest 260,000 years ago, for example, ancient ship in a particular group, much ability to communicate detailed, They are widely regarded not only as sym- known tools, say Wynn and other researchers. humans at Twin Rivers in what is now Zambia as people today declare social concrete information as well as bolic expression, but full-fledged art. These implements, which date back 2.6 mil- could envision a complex finished tool and allegiances and individual person- abstract concepts allowed early Thus many researchers are reluctant to lion years, consist mostly of rocks that have put it together in steps from different compo- alities by their clothing and jewelry. humans to cooperate and plan for the accept rare, one-off discoveries like the Tan- been split in two and then sharpened to make nents. They left behind finely made blades Yet while the Twin Rivers evidence is future in ways unique to our species, Tan or Berekhat Ram objects as signs of simple chopping and scraping implements. and other tools that had been modified— suggestive, it’s hard to be sure how the ochre thus enhancing their survival during symbolic behavior. “You can imagine [an Then, about 1.7 million years ago, large, usually by blunting or “backing” one edge— was actually used. There’s little sign that it was rough times and boosting their repro- ancient human] recognizing a resemblance teardrop-shaped tools called Acheulean hand to be hafted onto handles, presumably made ground into powder, as needed for decoration, ductive success in good times. “What but [the object] still hav[ing] no symbolic axes appeared in Africa. Likely created by of wood. These so-called backed tools have says Ian Watts, an independent ochre expert in aspects of human social organization and meaning at all,” says Philip Chase, an anthro- H. erectus and probably used to cut plants and been widely regarded as evidence of sym- Athens. And even ground ochre could have adaptation wouldn’t benefit from the evolu- pologist at the University of Pennsylvania. butcher animals, these hand-held tools vary bolic behavior when found at much younger had utilitarian uses, says archaeologist Lyn tion of language?” asked Terrence Deacon, a Thomas Wynn, an anthropologist at the greatly in shape, and archaeologists have sites. “This flexibility in stone tool manufac- Wadley of the University of Witwatersrand biological anthropologist at the University University of Colorado, Colorado Springs, debated whether creating the earliest ones ture [indicates] symbolic capabilities,” says in Johannesburg, South Africa. Modern-day of California, Berkeley, in his influential agrees: “If it’s a one-off, I don’t think it required an abstract mental template. But by archaeologist Sarah Wurz of the Iziko Muse- experiments have shown that ground ochre book The Symbolic Species: The Coevolu- counts. It’s not sending a message to anyone.” about 500,000 years ago, ancient humans were ums of Cape Town in South Africa. can be used to tan animal hides, help stone tion of Language and the Brain. Deacon creating more symmetrical Late Acheulean Similar cognitive abilities were possibly tools adhere to bone or wooden handles, and Eye of the beholder. Archaeologists debate whether went on to list just some of the advantages: Tools of the imagination required to make the famous even protect skin against mosquito bites. this modified stone was meant to represent a woman. organizing hunts, sharing food, teaching Given how difficult it is to detect 400,000-year-old wooden spears “We simply don’t know how ancient peo- toolmaking, sharing past experiences, and the earliest symbolic messages in “If it’s a one-off, I don’t think from Schöningen, Germany. One ple used ochre 300,000 years ago,” Wadley consider diagnostic elements of symbolic raising children. Indeed, many researchers the archaeological record, some recent study concludes that these says. And since at that date the ochre users behavior came together. And in work now have argued that symbolic communication is researchers look instead for proxy it counts. It’s not sending a spears’ creators—probably mem- were not modern humans but our archaic in press, the Blombos team reports finding what held groups of early humans together behaviors that might have bers of H. heidelbergensis—car- ancestors, some experts are leery of assign- engraved ochre in levels dating back to as they explored new environments and required similar cognitive abili- message to anyone.” ried out at least eight preplanned ing them symbolic savvy. 100,000 years ago (Science, endured climatic shifts. ties, such as toolmaking. Charles —THOMAS WYNN, UNIVERSITY OF COLORADO, steps spanning several days, Yet many archaeologists are 30 January, p. 569). As for art and other nonlinguistic forms Darwin himself saw an evolu- COLORADO SPRINGS including chopping tree branches willing to grant that our species, Online There are other hints that the of symbolic behavior, they may also have tionary parallel between tool- with hand axes and shaping the H. sapiens, was creating and sciencemag.org modern humans who ventured been key to cementing these bonds, by making and language, probably spears with stone flakes. using certain kinds of symbols Hear author out of Africa around this time expressing meanings that are difficult or Michael Balter the most sophisticated form of The idea that sophisticated by 75,000 years ago and per- discuss the roots of art at might also have engaged in sym- impossible to put into words. In that way, symbolic behavior. “To chip a toolmaking and symbolic haps much earlier. At sites such www.sciencemag.org/ bolic behavior. At the Skhul rock artistic expression, including music, may flint into the rudest tool,” Darwin thought require similar cognitive as Blombos Cave on South darwin. shelter in Israel, humans left have helped ensure the survival of the wrote in The Descent of Man, skills also gets some support Africa’s southern Cape, people 100,000-year-old shell beads fittest. This may also explain why great art demands a “perfect hand” as well from a surprising quarter: brain- left sophisticated tools, including elabo- considered by some to be personal orna- has such emotional force, because the most adapted to that task as the “vocal imaging studies. Stout’s team ran rately crafted bone points, as well as perfo- ments (Science, 23 June 2006, p. 1731). At effective symbols are those that convey organs” are to speaking. positron emission tomography rated beads made from snail shells and the 92,000-year-old Qafzeh Cave site their messages the most powerfully— To many researchers, making scans on three archaeologists— pieces of red ochre engraved with what nearby, modern humans apparently strongly something the artists at Chauvet Cave seem sophisticated tools and using Symbolic start. Some scientists argue that this 77,000-year-old engraved ochre all skillful stone knappers—as appear to be abstract designs. At this single preferred the color red: Excavators have to have understood very well.

symbols both require the capac- shows symbolic capacity. they made pre-Acheulean and D’ERRICO FRANCESCO AND HENSHILWOOD CHRIS ARCHAEOLOGY; OF RHONE-ALPES/DEPARTMENT COMMUNICATION/DRAC AND CULTURE OF MINISTRY FRENCH BOTTOM): TO (TOP CREDITS NOWELL APRIL AND D’ERRICO FRANCESCO CREDIT: site, a number of what many archaeologists studied 71 pieces of bright red ochre associ- –MICHAEL BALTER 6 7 710 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 711 0306NewsFocus.qxp 3/2/09 4:45 PM Page 1286 0306NewsFocus.qxp 3/2/09 4:46 PM Page 1287

ORIGINS

live in habitats such as scalding hot springs, Allen (no relation to John Allen) Catching rays. Long before plants got OnOn thethe OriginOrigin ofof don’t generate oxygen. Their photosynthetic and JoAnn Williams of Arizona in on the act, photosynthetic cyano- proteins huddle in relatively simple “reaction State University, Tempe, and col- bacteria living in pools like this one in centers” that may have been the predecessors leagues are working out how a Yellowstone National Park were chang- of the two photosystems. bacterial reaction center could ing the composition of the atmosphere. Photosynthesis Envisioning the steps that led to this have evolved photosystem II’s complex biochemistry is mind-boggling. appetite for electrons. says astrobiologist Roger Buick Similarities between proteins in photo- Taking a hands-on approach, of the University of Washington, piece together how and when organisms first synthetic and nonphotosynthetic bacteria they have been tinkering with the Seattle. These hints could push began to harness light’s energy. Although suggest that early microbes co-opted some reaction center of the purple bac- the origin back 600 million most modern photosynthesizers make oxygen photosynthesis genes from other metabolic terium Rhodobacter sphaeroides years or more. from water, the earliest solar-powered bacteria pathways. But protophotosynthesizers to determine if they can make One line of evidence is oil relied on different ingredients, perhaps hydro- might also have helped each other piece it more like photosystem II. biomarkers that researchers gen sulfide. Over time, the photosynthetic these pathways together by swapping First they targeted bacterio- think are the remains of cyano- machinery became more sophisticated, even- genes. Biochemist Robert Blankenship of chlorophyll, the bacterial version bacteria. They’ve turned up in tually leading to the green, well-oxygenated Washington University in St. Louis, Mis- of chlorophyll that’s at the core rocks that are up to 2.7 billion world that surrounds us today. In the lab, some souri, and colleagues say they’ve uncovered of the reaction center, and years old. And in western Aus- biochemists are recapitulating the chemical traces of these lateral gene transfers by altered the number of hydrogen tralia, thick shale deposits that steps that led to this increased complexity. comparing complete bacterial genomes. bonds. Adding hydrogen bonds are 3.2 billion years old hold Other researchers are locked in debates over For example, their 2002 study of more than hiked the molecule’s greed for rich bacterial remains but no just when this transition happened, 2.4 billion 60 photosynthetic and nonphotosynthetic electrons, they found. traces of sulfur or other possible years ago or much earlier. bacteria (Science, 22 November 2002, The water-cleaving portion electron sources, suggesting that Looking so far into the past is difficult. p. 1616) suggested that bugs had passed of photosystem II sports four the microbes were using water The geological record for that time is around several photosynthesis genes, manganese atoms that become to make energy. skimpy and tricky to interpret. Eons of evo- including some involved in synthesizing the oxidized, or lose electrons. So Geologist Euan Nisbet of lution have blurred the molecular vestiges bacterial version of chlorophyll. the team equipped the bacterial Royal Holloway, University of of the early events that remain in living Gene-sharing might also explain the puz- reaction center with one atom London, and colleagues found organisms. But “it’s a terribly important zling distribution of the photosystems, of the metal. In this modified additional support for an early problem,” says biochemist Carl Bauer of Blankenship says. A cell needs both photo- version, the added manganese origin when they went search- Indiana University, Bloomington, one well systems to carry out oxygenic photosynthesis. also underwent oxidation, the ing for traces of RuBisCO, a Try to picture the world without photo- worth the travails. Yet modern nonoxygenic bacteria have the researchers reported in 2005. key photosynthetic enzyme. synthesis. Obviously, you’d have to strip presumptive predecessor either of photo- James Allen says that their cre- RuBisCO feeds carbon dioxide away the greenery—not just the redwoods To catch a photon system I or of photosystem II, never both. To ations aren’t powerful enough to into the reactions that yield sug- and sunflowers, but also the humble algae Over more than 200 years, researchers have explain how the two protein complexes split water. But eventually, they ars. The enzyme version found and the light-capturing bacteria that nourish ironed out most of the molecular details of wound up together, Blankenship favors “a hope to engineer a reaction center in oxygenic photosynthesizers many of the world’s ecosystems. Gone, too, how organisms turn carbon dioxide and large-scale lateral [gene] transfer” or even a that can oxidize less possessive Oxygenic photosynthesis “was the plays favorites: It prefers carbon would be everything that depends on photo- water into food. Chlorophyll pigment and fusion of organisms carrying each photo- molecules, such as hydrogen dioxide that contains the carbon- synthetic organisms, directly or indirectly, about 100 other proteins team up to put light system. However, other researchers remain peroxide, that would have been last of the great inventions of micro- 12 isotope over the bulkier car- for sustenance—from leaf-munching bee- to work. Plants, some protists, and cyano- skeptical, arguing that one photosystem present on the early Earth. Even bial metabolism, and it changed the bon-13. In 2007, Nisbet and tles to meat-eating lions. Even corals, which bacteria embed their chlorophyll in two evolved from the other, possibly through the if the researchers never replicate his colleagues found dispropor- play host to algal partners, would lose their large protein clusters, photosystem I and duplication of genes, creating an ancient cell photosystem II, “if we define planetary environment forever.” tionately low carbon-13 values main food source. photosystem II. And they need both systems with both. No one knows for sure. the intermediate stages, we’ve indicative of RuBisCO activity Photosynthesis makes Earth congenial for to use water as an electron source. Light accomplished a lot,” he says. —Paul Falkowski, Rutgers University when they analyzed organic life in other ways, too. Early photosynthesizers jump-starts an electrical circuit in which The electron thief matter in rocks from three sites pumped up atmospheric oxygen concentra- electrons flow from the photosystems Either way, it took some fancy fiddling to Something in the air about 2.4 billion years ago, geologists see in Zimbabwe and Canada that are between tions, making way for complex multicellular through protein chains that convert the primitive reaction How the photosystems got their start is cru- the first unmistakable signs of significant, 2.7 billion and 2.9 billion years old. Nisbet life, including us. And water-dwellers were make the energy-rich molecules THE YEAR OF centers to oxygen-generating cial for understanding the origin of photo- sustained levels of atmospheric oxygen. concludes that oxygen-making photo- able to colonize the land only because the ATP and NADPH. These mole- Given its photosystems. Oxygenic photo- synthesis. But the question that’s drawn the These signs include red beds, or layers synthesis began at least 2.9 billion years ago. oxygen helped create the ozone layer that cules then power the synthesis DARWINimportance in synthesis was a huge upgrade, most attention—and provoked the most tinged by oxidized iron, i.e., rust. Further The early-origin case isn’t ironclad. For shields against the sun’s ultraviolet radiation. of the sugars that organisms making and leading to a land of plenty, says wrangling—is when photosynthesis began. support that the GOE marks an atmospheric example, a 2008 paper that has some Oxygen-producing, or oxygenic, photo- depend on to grow and multiply. biochemist John Allen of Queen Most researchers accept that nonoxygenic revolution comes from a technique that researchers fuming claims that the oil bio- synthesis “was the last of the great inventions Photosystem II—the strongest keeping earth Mary, University of London. photosynthesis arose first, probably shortly detects skewed abundances of sulfur iso- markers are contaminants that seeped in of microbial metabolism, and it changed the naturally occurring oxidant— lush, photo- “Water is everywhere, so the after life originated more than 3.8 billion topes that occur if the air lacks oxygen. from younger rocks. Advocates also have to planetary environment forever,” says geo- regains its lost electrons by organisms never ran out of elec- years ago. “Life needs an energy source, and These imbalances persisted until the GOE, explain why it took hundreds of millions of biologist Paul Falkowski of Rutgers Univer- swiping them from water, gener- synthesis ranks trons. They were unstoppable.” the sun is the only ubiquitous and reliable when they vanished. years for oxygen to build up in the air. sity in New Brunswick, New Jersey. ating oxygen as a waste product. high on the But water clings to its elec- energy source,” says Blankenship. Hard-liners construe these data to mean Although the last word on the origins of Given its importance in making and keep- However, some bacteria trons. With its oxidizing power, top- 10 list of The sharpest disputes revolve around that oxygenic photosynthesis could not have oxygen-making photosynthesis isn’t in, ing Earth lush, photosynthesis ranks high on don’t rely on water as an elec- This essay is the third photosystem II can wrench them when organisms shifted to oxygenic photo- emerged until shortly before the GOE. But researchers say they are making progress. One the top-10 list of evolutionary milestones. By tron source, using hydrogen sul- in aevolutionary monthly series. More away, but the reaction centers in synthesis. At issue is how to interpret a other scientists disagree. “We are finding thing is for certain, however: Without this delving into ancient rocks and poring over on evolution online at fide or other alternatives. These blogs.sciencemag.org/milestones. nonoxygenic photosynthesizers watershed in the fossil record known as the more and more hints that oxygenic photo- innovation, Earth would look a lot like Mars. CREDIT: K. SUTLIFF/ SCIENCE K. CREDIT: DNA sequences, researchers are now trying to nonconformists, which today origins. cannot. Biochemists James IMAGES CLARK/GETTY NEALE CREDIT: great oxidation event (GOE). In rocks from synthesis goes deeper into the fossil record,” –MITCH LESLIE

1286 8 6 MARCH 2009 VOL 323 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 323 6 MARCH 2009 1287 9 0403NewsFocus.qxp 3/30/09 5:55 PM Page 28 0403NewsFocus.qxp 3/30/09 5:23 PM Page 29

NEWSFOCUS ORIGINS

embryo that serves as its food supply. Stockholm. “When we started, Darwin was perplexed by the diversity of the search profile was bigger, OnOn thethe OriginOrigin ofof flowering plants; they were too numerous a magnolia [flower],” she and too varied, and there were too few fos- recalls. But 30 years ago, she sils to sort out which were more primitive. and others discovered tiny Flowering Plants Throughout much of the 20th century, mag- ancient flowers by sieving nolia relatives with relatively large flowers through sand and clay sedi- were leading candidates for the most primi- ments. With this technique, how flowers got started—and from which tive living flowers, although a few they have now collected hun- ancestor. Today, researchers have analytical researchers looked to small herbs instead. dreds of millimeter-size tools, fossils, genomic data, and insights that In the late 1990s, molecular systematics flowers, some preserved in Darwin could never have imagined, all of came to the rescue, with several reports pre- three dimensions, from Por- which make these mysteries less abom- senting a fairly consistent picture of the tugal and other locations with inable. Over the past 40 years, techniques lower branches of the angiosperm tree. An Cretaceous deposits 70 mil- for assessing the relationships between obscure shrub found only in New Caledonia lion to 120 million years old. organisms have greatly improved, and gene emerged as a crucial window to the past. This fossil diversity sequences, as well as morphology, now help Amborella trichopoda, with its 6-millimeter shows that angiosperms were researchers sort out which angiosperms greenish-yellow flowers, lives deep in the thriving, with several groups arose early and which arose late. New fossil cloud forests there. In multiple gene-based well-established, by 100 mil- Out of the past. finds and new ways to study them—with assessments, including an analysis in 2007 lion years ago. In some, the Tiny Amborella sits synchrotron radiation, for example—pro- of 81 genes from chloroplast genomes flower parts are whorled like at the bottom of the vide a clearer view of the detailed anatomy belonging to 64 species, Amborella sits those of modern flowers; in angiosperm family tree. of ancient plants. And researchers from var- at the base of the angiosperm family tree, others they are spiraled, con- ious fields are figuring out genomic changes the sister group of all the rest of the sidered by some researchers that might explain the amazing success of angiosperms. as the more primitive arrangement. Some from one of the nonflowering seed plants this fast-evolving group. Given that placement, Amborella’s tiny flower fossils have prescribed numbers of “We are realizing that this or gymnosperms, whose heyday was 200 These approaches have given researchers flowers may hint at what early blossoms petals, another modern feature, whereas in million years ago. Modern gymnosperms a much better sense of what early flowers were like. It’s one of “the most similar living others the petal count varies. huge diversity is probably include conifers, ginkgoes, and the cycads, IN 1879, CHARLES DARWIN PENNED A LETTER were like and the relationships among them. flower[s]” to the world’s first flower, says In 1998, Chinese geologist Ge Sun of with their stout trunks and large fronds. to British botanist Joseph Dalton Hooker, But one of Darwin’s mysteries remains: the James Doyle of the University of Califor- Jilin University in Changchun, China, came the result of one innova- Before angiosperms came along, these lamenting an “abominable mystery” that nature and identity of the angiosperm ances- nia, Davis. The petals and sepals of its sin- across what seemed to be a much older plants were much more diverse and threw a wrench into his theory of evolution: tor itself. When flowering plants show up in gle-sex flowers are indistinguishable and flower. The fossil, called Archaefructus, was tion piled on top of included cycadlike species, such as the How did flowering plants diversify and the fossil record, they appear with a bang, vary in number; so too do the numbers of an aquatic plant that looked to be 144 mil- another innovation.” extinct Bennettitales, and many woody spread so rapidly across the globe? From with no obvious series of intermediates, as seed-producing carpels on female flowers lion years old. By 2002, Sun and David plants called Gnetales, of which rice paddies to orange groves, alpine mead- Darwin noted. Researchers still don’t know and pollen-generating stamens on male Dilcher of the Florida Museum of Natural —Peter Crane, a few representatives, including the ows to formal gardens, prairies to oak- which seed- and pollen-bearing flowers. The organs are spirally History (FLMNH) in Gainesville had University of Chicago joint firs, survive today (see family tree, hickory forests, the 300,000 species of organs eventually evolved into THE YEAR OF arranged, and carpels, rather described an entire plant, from roots to flow- p. 31). Also common in the Jurassic were angiosperms alive today shape most terres- the comparable flower parts. than being closed by fused tis- ers, entombed on a slab of rock unearthed in These fossils often spark debate because seed ferns, a group now long gone; their trial landscapes and much of human life “We’re a bit mystified,” says DARWINFor more on sue as in roses and almost all Liaoning in northeastern China. specimens tend to be imperfectly preserved most famous member is Caytonia, which and culture. Their blooms color and scent botanist Michael Donoghue of flower origins, familiar flowers, are sealed by a In one sense, Archaefructus wasn’t much and leave room for interpretation. To help seems to have precarpel-like structures. our world; their fruits, roots, and seeds feed Yale University. “It doesn’t secretion. to look at. “It’s a flowering plant before remedy that, Friis and her colleagues have These groups’ perceived relevance to us; and their biomass provides clothing, appear that we can locate a close listen to a Most genetic analyses showed there were flowers,” Dilcher notes. It lacked begun to examine flowers using synchro- flower evolution and their relationships to building materials, and fuel. And yet this relative of the flowering plants.” podcast by that water lilies were the next petals and sepals, but it did have an tron radiation to generate a 3D image of angiosperms have ping-ponged between takeover, which took place about 100 mil- branch up the angiosperm tree, enclosed carpel. When Kevin Nixon and their inner structures, allowing the fossil to camps, depending on how the evolutionary lion years ago, apparently happened in a Seeking the first flower author Elizabeth followed by a group represented colleagues at Cornell University compared remain intact while Friis peers inside it trees were constructed. blink of geological time, just a few tens of One of two major living groups by star anise, which also has a Pennisi at its traits with those same traits in 173 living from many angles (Science, 7 December In the mid-1980s, Peter Crane, now at millions of years. of seed plants, angiosperms have primitive look about it, says plants, Archaefructus came out as a sister to 2007, p. 1546). “We can get fantastic reso- the University of Chicago in Illinois, pro- The father of evolution couldn’t quite “covered” seeds that develop This essaywww. is the fourth Doyle, “though each of these living angiosperms and closer to the com- lution,” says Friis. “It’s really exciting.” But posed a solution, the anthophyte hypothesis. fathom it. Darwin had an “abhorrence that encased in a protective tissue in a monthly series. has deviations from the ances- mon ancestor than even Amborella. so far, the flowers Friis finds are Using several lines of evidence and noting For moresciencemag. on evolutionary evolution could be both rapid and poten- called a carpel (picture a bean topics online, see the tral type.” Archaefructus’s distinction was short- too diverse to trace back to a that both Bennettitales and tially even saltational,” writes William pod). That’s in contrast to the Origins blogorg/ at lived, however. Within months, better dat- particular ancestor. “From Gnetales organize their male Friedman in the January American Journal nonflowering gymnosperms, blogs.sciencemag.org/ Fossil records ing of the sediments in which it was found these fossils, we cannot and female organs together origins.multimedia/ For more on of Botany, which is devoted to this such as conifers, which bear flower origins, listen Although some fossil pollen yielded younger dates, putting this first say what is the basic in what could be con- “abominable mystery.” Throughout his life, naked seeds on scales. An to a podcastpodcast. by author dates back 135 million years, no flower squarely with other early fossil form,” she says. strued as a preflower, Elizabeth Pennisi at Darwin pestered botanists for their angiosperm’s carpel sits at the www.sciencemag.org/ credible earlier fossil evidence flower parts, about 125 million years old. he considered them, thoughts on the matter, but they couldn’t center of the flower, typically multimedia/podcast. exists. In Darwin’s day, and for Also, a 2009 phylogenetic analysis of Before flowers along with angiosperms, give him much help. surrounded by pollen-laden sta- many decades afterward, pale- 67 taxa by Doyle and Peter Endress of the Although they have yet as comprising a single Now, 130 years later, evolutionary biol- mens. In most flowers, the carpel and stamens obotanists primarily found leaves or pollen University of Zurich, Switzerland, placed to find the oldest fossil Larger than life. Although merely angiosperm entity called ogists are still pestering botanists for clues are surrounded by petals and an outer row of but almost no fossil flowers. They had the the fossil in with water lilies rather than at flowers, researchers 2.2 millimeters in diameter, this 3D anthophytes. For the next about what has made this plant group so leaflike sepals. Seeds have a double coating wrong search image, says Else Marie Friis of the base of the angiosperms, although this assume that the ances- fossil flower shows that grasses date decade, most family trees CREDITS (TOP TO BOTTOM): REPRODUCED WITH THE KIND PERMISSION OF THE DIRECTOR AND THE BOARD OF TRUSTEES, ROYAL BOTANIC GARDENS, KEW; GEORGE RICHMOND/BRIDGEMAN ART LIBRARY, LONDON (SUPERSTOCK) LONDON LIBRARY, ART RICHMOND/BRIDGEMAN GEORGE KEW; GARDENS, BOTANIC ROYAL TRUSTEES, OF BOARD THE AND DIRECTOR THE OF PERMISSION KIND THE WITH REPRODUCED BOTTOM): TO (TOP CREDITS CREDITS (TOP TO BOTTOM): COURTESY OF STEPHEN MCCABE, UC SANTA CRUZ; PHOTO BY JENNIFER SVITKO, COURTESY OF WILLIAM L. CREPET, CORNELL UNIVERSITY CORNELL CREPET, L. WILLIAM OF COURTESY SVITKO, JENNIFER BY PHOTO CRUZ; SANTA UC MCCABE, STEPHEN OF COURTESY BOTTOM): TO (TOP CREDITS successful, as well as when, where, and as well as endosperm, tissue surrounding the the Swedish Museum of Natural History in GARDENS, KEW BOTANIC REPRODUCED WITH THE KIND PERMISSION OF DIRECTOR AND BOARD TRUSTEES, ROYAL CREDIT: conclusion is contested. tral angiosperm evolved back to 94 million years ago. based on morphology sup- 10 11 28 3 APRIL 2009 VOL 324 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 324 3 APRIL 2009 29 0403NewsFocus.qxp 3/30/09 5:23 PM Page 30 0403NewsFocus.qxp 3/30/09 6:07 PM Page 31

ORIGINS ORIGINS

such a causal relationship is not set- tled. Later, animals that ate fruit and SEED PLANT Paleozoic seed ferns dispersed seeds likely helped evolv- ing species expand quickly into new PHYLOGENY Asterids territory. Some think the answer lies » Eudicots in genes: duplications that gave the angiosperm genome opportunities to Rosids try out new floral shapes, new chem- ical attractants, and so forth. This flexibility enabled angiosperms to Monocots exploit new niches and set them up Angiosperms for long-term evolutionary success. Magnolias “My own view is that in the past, we » have looked for one feature,” says Crane. Now, “we are realizing that Water lilies this huge diversity is probably the result of one innovation piled on top ? of another innovation.” Archaefructus » The latest insights into diversifica- tion come from gene studies. From 2001 to 2006, Pamela Soltis of the Amborella FLMNH and Claude dePamphilis ? of Pennsylvania State University, Caytonia University Park, participated in the Floral Genome Project, which ? Flowers, food, fuel.fuel. DarwinDarwin marveled at thethe diversitydiversity of of angiosperms. angiosperms. Given Given thatmainstays they represent of both our nine welfare in 10 and land sense plants, of beauty.it’s no surpriseClockwise that from they left: serve aspens, as Bennettitales mainstaysthat they represent of both ournine welfare in 10 land and plants,sense of it’s beauty. no surprise Clockwise that they from serve top left:as aspens,orchids, orchids, grasses, grasses, sunflowers, sunflowers, tulips, apples, tulips, walnuts. apples, walnuts. searched for genes in 15 angiosperms. Now as a follow-up, the Ancestral ported this idea. Crane and others carefully work points you in another direction,” out. Today, almost nine in 10 land plants Angiosperm Genome Project looks dissected and described fossils of these Crane says. are angiosperms. at gene activity in five early angio- Cycads sperms and a cycad, a gymnosperm. groups, looking for the precursors to And if the molecular work is correct, The exact timing of the angiosperms’ gymnosperms carpels, the seed’s double coat, and other then the field doesn’t know in which direc- explosion and expansion is under debate, as is DePamphilis and his colleagues

Ginkgoes Living distinctive angiosperm traits. tion to turn, because in most analyses the the cause. At least one estimate based on the matched all the genes in each » But they have run into problems. “We do genetic data don’t place any living plant rate at which gene sequences change—that species against one another to deter- ? not really know how to compare them close to angiosperms. The angiosperms is, the ticking of the molecular clock— mine the number of duplicates. They Gnetales because the structures are very different- group together, the living gymnosperms pushes angiosperm evolution back to 215 then looked at the number of differ- looking; figuring out what’s homologous is group together, and there’s nothing in million years ago. “There appears to be a ences in the sequences of each gene quite a difficult thing,” says Crane. He and between. “The nonangiosperm ancestor just gap in the fossil record,” says Donoghue, pair to get a sense of how long ago Conifers » his colleagues argue, for example, that the isn’t there,” says paleobotanist William who also notes that molecular dating meth- the duplication occurred. In most Extinct taxa seeds in the Bennettitales have two cover- Crepet of Cornell. “I’m starting to worry ods “are still in their infancy” and, thus, early angiosperms, including water ings, which may be a link to angiosperms. that we will never know, that it transformed could be misleading. He and others think lilies and magnolias, they saw many Shifting branches. As this simplified family tree shows, gene studies have helped clarify the relationships of many But in the January American Journal of without intermediates.” that flowering plants lingered in obscurity simultaneous duplications—but not living angiosperms, but fitting in extinct species is still a challenge, and some nodes are hotly debated. Botany, Gar Rothwell of Ohio University, for tens of millions of years before radiating in Amborella, they reported in the Athens, and two colleagues disagree, say- Seeds of success toward their current diversity. January 2009 American Journal of Botany, The Floral Genome Project also looked are not as well-defined as they are in Ara- ing that what Crane calls the outer layer is The angiosperm’s ancestor may be missing, Whatever the timing, there was some- confirming earlier reports. The data suggest to see whether the genetic programs guiding bidopsis. This sloppiness may have made the only layer, and find but what is very clear—and was quite thing special about the angiosperm radia- that a key genome duplication happened flower development were consistent development flexible enough to undergo fault with the hypothesis annoying to Darwin—is that the angio- tion. During the 1980s and again in 1997, after the lineage leading to Amborella split throughout the angiosperms. “We found that many small changes in expression patterns in general. sperm prototype so readily proved a Cornell’s Karl Niklas compiled a database off but before water lilies evolved. “We’re there are fundamental aspects that are con- and functions that helped yield the great To make matters winner. Seed ferns and showing the first and last occurrences of beginning to get the idea that polyploidiza- served in the earliest lineages,” says Soltis. diversity in floral forms. worse for anthophyte other gymnosperms fossil plants. When he and Crepet used that tion may have been a driving force in creat- “But there are differences in how the genes In his letter to Hooker, Darwin wrote proponents, gene- arose about 370 and more recent information to look at ing many new genes that drive floral devel- are deployed.” that he would like “to see this whole prob- based evolutionary million years ago species’ appearances and disappearances, opment,” dePamphilis says. Take the avocado, a species on the lower lem solved.” A decade ago, Crepet thought trees break up this and dominated the they found that new angiosperms appeared Others have noted that a duplication branches of the angiosperm tree. In most Darwin would have gotten his wish by now. grouping, pulling the planet for 250 mil- in bursts through time, whereas other plants, occurred in the evolution of grasses, and the angiosperms, the flower parts are arranged in That hasn’t happened, but Crepet is opti- Gnetales off any lion years. Then in a such as gymnosperms, radiated rapidly only Floral Genome Project confirms that yet concentric circles, or whorls, around the mistic that he and his colleagues are on the angiosperm branch few tens of millions at first. Moreover, angiosperms proved less PHOTOS.COM IMAGES) OTHER (ALL DILCHER; DAVID IMAGE) another duplication paved the way for eudi- carpels, with stamens innermost, then petals, right track, as analyses of various kinds of and placing them of years, angio- likely to disappear, somehow resisting cots, the group that includes apples, roses, and finally sepals. Each tissue has its own data become more sophisticated. “We are among or next to the sperms edged them extinction, says Crepet. beans, tomatoes, and sunflowers. “There are distinct pattern of gene expression, but not less likely to go around in circles in the next other gymnosperms. Once the angiosperms arrived, how did some real ‘hot spots’ in angiosperm evolu- in the avocado. Genes that in Arabidopsis 10 years,” he says. “I believe a solution to tionary history,” says dePamphilis, who is are active only in, say, the developing petals the problem is within reach. … The mystery “The molecular work Inside and out. Synchrotron radiation helped pro- they diversify and spread so quickly? Darwin points in one direc- duce a 3D rendering (gold) of this fossil male flower suspected that coevolution with insect polli- working to fully sequence the genome of spill over in avocado to the sepals. Thus in is solvable.” CREDITS: ( ARCHAEFRUCTUS CREDITS:

CREDITS (TOP TO BOTTOM): PHOTOS.COM; ELSE MARIE FRIIS MARIE ELSE PHOTOS.COM; BOTTOM): TO (TOP CREDITS –ELIZABETH PENNISI tion; the paleobotanical (right) and insights into its internal structure. nators helped drive diversification, though FRIIS MARIE ELSE PHOTOS.COM; BOTTOM): TO (TOP CREDITS Amborella with his colleagues. the more primitive plants, petals and sepals 12 13 30 3 APRIL 2009 VOL 324 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 324 3 APRIL 2009 31 0501PerspectivesR1 4/28/09 12:07 PM Page 596

0501PerspectivesR10501PerspectivesR1 4/28/094/28/09 12:0712:07 PMPM PagePage 596596

PERSPECTIVESPERSPECTIVES

HISTORY OF SCIENCE HISTORY OF SCIENCE HISTORY OF SCIENCE atmospheric, oceanic, geological, ecological, typically producing detailed descriptive reports Global environmental change may be the and cultural phenomena across the globe. Hum- with little integration within or among the com- greatest challenge faced by human societies In the early 19th century,InIn thethe Alexander earlyearly 19th19th century,voncentury, AlexanderAlexander vonvon boldt’s obsession with geographically refer- ponent entities. However, for a few intellectu- since the advent of agriculture. Humboldt ad- Alexander von Humboldt and Humboldt laid the foundations for today’s Alexander vonAlexander Humboldt von and Humboldt andHumboldt laid the foundations for today’s enced measurements and collections was central ally nimble participants—including Charles vocated for science that spoke to human needs EarthEarth systemsystem sciences.sciences. Earth system sciences. to his vision. He recognized that spatial arrays Darwin, T. H. Huxley, Matthew Maury, Asa and concerns (5). It is fitting that on the 150th the General Physics of the Earth of observations could be aggregated to reveal Gray, C. Hart Merriam, and Peter Kropotkin— anniversary of his death, we recognize his role the General Physics of the Earth patterns that would in turn reveal underlying these explorations provided data and experience in fostering the sciences that speak to the most Stephen T. Jackson Stephen T. Jackson processes—such as the distribution of incident that spurred the development of biogeography, profound human concerns—sustainability of Stephen T. Jackson s scientists are celebrating the 200th radiation, the transport of heat and materials in ecology, oceanography, and other environmen- human societies and the ecosystems on which ss scientistsscientists areare celebratingcelebrating thethe 200th200th anniversary of Charles Darwin’s birth winds and ocean currents, the influence of tem- tal sciences (11). they depend. anniversaryanniversary ofof CharlesCharles Darwin’sDarwin’s birthbirth and the effect of latitude Unfortunately, the conceptual unification s scientists are celebrating the 200thand the 150th anniversary of the pub- perature on plant form, A andand thethe 150th150th anniversaryanniversary ofof thethe pub-pub- and continentality on mountain snowline. among the sciences of the Earth that Humboldt References and Notes anniversary of Charles Darwin’slication birth of his On the Origin of Species, licationlication of of his his On the Origin of Species,, He expanded this vision in the succeeding sought never developed in the century follow- 1. P. J. Bowler, Science 323, 223 (2009).[Abstract/ Darwin’s ideas continue to shape and enrich and the 150th anniversary of theDarwin’s pub- ideas continue to shape and enrich years, establishing international cooperative ing his death. Disciplinary specialization played Free Full Text] A the sciences (1). 6 May 2009 marks the 150th 2. A. de Humboldt, Essai sur la géographie des lication of his On the Origin of Speciesthethe sciencessciences, ((11).). 66 MayMay 20092009 marksmarks thethe 150th150th networks of meteorological and geomagnetic a large role in eclipsing Humboldt’s integration, anniversary of the death of another 19th-cen- plantes (Levrault, Schell & Co., Paris, 1807). Darwin’s ideas continue to shape andanniversaryanniversary enrich ofof thethe deathdeath ofof anotheranother 19th-cen-19th-cen- measurement stations, inventing isotherms and as did 20th-century trends toward reductionism, 3. An English translation of (2) is currently in turytury figure—Alexander figure—Alexander von von Humboldt— Humboldt— the sciences (1). 6 May 2009 marks thetury 150th figure—Alexander von Humboldt— other graphical devices to portray spatial pat- experimentalism, and fine-scale processes in press at the University of Chicago Press, Chi- whose scientific legacy also flourishes in the terns, and noting that plant form is often better many disciplines. cago. anniversary of the death of another 19th-cen-21st21st century. century. Humboldt Humboldt helped helped create create the the predicted by local environment than by taxo- A new incarnation of Humboldt’s general 4. A. de Humboldt, Personal Narrative of Trav- els to the Equinoctial Regions of the New intellectualintellectual worldworld DarwinDarwin inhabited,inhabited, andand hishis tury figure—Alexander von Humboldt—intellectual world Darwin inhabited, and his nomic affinity (a paradox resolved by Darwin). physics of the Earth began to emerge with the Continent, During the Years 1799–1804, by writings inspired Darwin to embark on whose scientific legacy also flourisheswritings in the inspired Darwin to embark on Humboldt’s genius lay in his geographical vi- plate tectonics revolution in the 1960s. Draw- Alexander de Humboldt and Aime Bonpland; H.M.S. Beagle. More pertinent to our time, 21st century. Humboldt helped createH.M.S. the Beagle. More pertinent to our time, sion, and in his intuition that Earth’s land sur- ing on Humboldtian spatial arrays of observa- with Maps, Plans, &c. (Longman, Hurst, Rees, Humboldt established the foundation for the face, oceans, atmosphere, and inhabitants form tions, this theory provided a unified explanatory Orme, & Brown, London, 1814 to 1829), vols. intellectual world Darwin inhabited, and his 1 to 7. Earth system sciences: the integrated system an integrated whole, with linkages among the framework for disparate geophysical, geologi- writings inspired Darwin to embarkof knowledge on on which human society may 5. A. von Humboldt, Cosmos: A Sketch of the ofof knowledgeknowledge onon whichwhich humanhuman societysociety maymay various components (4, 5). Humboldt’s general cal, paleontological, and biogeographic phe- Physical Description of the Universe (Johns

H.M.S. Beagle. More pertinent to ourdependdepend time, inin thethe faceface ofof globalglobal climateclimate change.change. physics of the Earth envisioned climate as a nomena. Hopkins Univ. Press, Baltimore, 1997), vol. 1. Humboldt established the foundation forLike the Darwin, Humboldt undertook a major control of Earth-surface phenomena, with Today, a second, even broader manifestation 6. M. Nicolson, Hist. Sci. 25, 167 (1987). [ISI] major voyage that would shape his ideas vegetation serving as both an index of climate of Humboldt’s vision aspires to understand the 7. M. Nicolson, in Romanticism and the Sciences, Earth system sciences: the integrated system A. Cunningham, N. Jardine, Eds. (Cambridge andand thinking.thinking. HumboldtHumboldt spentspent 55 yearsyears (1799(1799 and a proximal control of microclimate, animal interactions and feedbacks among the compo- of knowledge on which human societyto 1804) may with botanist Aimé Bonpland explor- Univ. Press, Cambridge, 1990), p. 169. toto 1804)1804) withwith botanistbotanist AiméAimé BonplandBonpland explor-explor- habitat, and cultural practices (6–8). nents of the Earth system, encompassing the 8. M. Dettelbach, in Cultures of Natural History, depend in the face of global climate change.inging Venezuela,Venezuela, thethe northernnorthern Andes,Andes, andand cen-cen- Humboldt’s dream of systematic observa- lithosphere, atmosphere, hydrosphere, cryo- N. Jardine, et al., Eds. (Cambridge Univ. Press,

Like Darwin, Humboldt undertooktraltral Mexico,Mexico, a withwith visitsvisits toto Tenerife,Tenerife, Cuba,Cuba, andand tional arrays across the globe took hold in the sphere, and biosphere as well ashuman societies Cambridge, 1996), p. 287. 9. S. F. Cannon, Science in Culture: The Early major voyage that would shape histhethe ideas UnitedUnited States.States. TheyThey collectedcollected botanical,botanical, 19th century. Throughout the century, countless and economies. This effort is often referred to as zoological, geological, and ethnological spec- Humboldt-inspired explorations were launched, Earth system science, but it could just as well be Victorian Period (Dawson, New York, 1978). zoological,zoological, geological,geological, andand ethnologicalethnological spec-spec- 10. W. H. Goetzmann, New Lands, New Men: and thinking. Humboldt spent 5 years (1799 each involving systematic measurement and designated “general physics of the Earth,” using imens,imens, mademade extensiveextensive atmosphericatmospheric andand geo-geo- Intellectual riches. The central portion of Humboldt’s Physical Tableau of the Andes and Neighboring America and the Second Great Age of Discov- to 1804) with botanist Aimé Bonpland physicalexplor- measurements, and recorded the IntellectualIntellectual riches.riches. TheThe centralcentral portionportion ofof Humboldt’sHumboldt’s PhysicalPhysical TableauTableau ofof thethe AndesAndes andand NeighboringNeighboring mapping of physical, biological, and often cul- the early-19th century definition of physics as physicalphysical measurements, measurements, and and recorded recorded the the CountriesCountries,, publishedpublished asas partpart ofof ((22,, 33),), showsshows ChimborazoChimborazo inin profile,profile, withwith vegetationvegetation zones,zones, plantplant species,species, andand ery (Viking, New York, 1986). geographic location of their thousands of Countries, published as part of (2, 3), shows Chimborazo in profile, with vegetation zones, plant species, and tural features of landscapes and oceans (8–10). the study of the material world and its phenom- 11. P. J. Bowler, The Norton History of the Envi- ing Venezuela, the northern Andes, andgeographicgeographic cen- location location of of their their thousands thousands of of snowlinesnowline depicteddepicted atat appropriateappropriate elevations.elevations. InIn thethe original,original, thethe profileprofile isis flankedflanked onon bothboth sidessides byby tablestables specimens and tens of thousands of measure- snowline depicted at appropriate elevations. In the original, the profile is flanked on both sides by tables These surveys were relentlessly inductive, ena (which we now call science). ronmental Sciences (Norton, New York, 1993). tral Mexico, with visits to Tenerife, Cuba,specimensspecimens and andand tenstens ofof thousandsthousands ofof measure-measure- describingdescribing elevationalelevational patternspatterns inin temperature,temperature, humidity,humidity, lightlight refractionrefraction andand intensity,intensity, agriculture,agriculture, fauna,fauna, andand ments. Humboldt spent the next 22 years and otherother physical,physical, chemical,chemical, andand biologicalbiological features.features. the United States. They collected botanical,most of his inherited fortune in Paris, pre- zoological, geological, and ethnologicalmost spec- of his inherited fortune in Paris, pre- paringparing andand publishingpublishing 4545 volumesvolumes ofof aa never-never- distributiondistribution ofof incidentincident radiation,radiation, thethe transporttransport servingserving asas bothboth anan indexindex ofof climateclimate andand aa prox-prox- imens, made extensive atmospheric andfinished geo- report on his travels. of heat and materials in winds and ocean cur- imal control of microclimate, animal habitat, finishedfinished reportreportIntellectualIntellectual onon hishis travels.travels. riches. riches. The central The portioncentral of ofofportion Humboldt’s heatheat andand of materials materials Humboldt’sPhysical Tableau inin windswinds Physical of the andand Andes ocean oceanTableau and cur-cur- Neighboring of theimalimal Andes Countries controlcontrol and of,of published microclimate,Neighboringmicroclimate, as part animalanimal habitat,habitat, Of these volumes, the first was a slim work rents, the influence of temperature on plant and cultural practices (6–8). physical measurements, and recordedOf the these volumes,ofCountries (2, 3), shows the, published first Chimborazo was a asslim inpart workprofile, of (2 ,withrents,rents, 3), vegetation shows thethe influenceinfluence Chimborazo zones, plant ofof temperature temperatureinspecies, profile, and withsnowline onon plantplantvegetation depictedandand zones, at culturalcultural appropriate plant practicespractices species,elevations. ((66–– 88and ).). geographic location of their thousandsentitledentitled of EssayIn theon original,the Geography the profile of is Plantsflanked onform, form,both sides andand thebythe tables effecteffect describing ofof latitudelatitude elevational andand continen-continen- patterns in temperature,Humboldt’s humidity, dream light of systematic observa- (2, 3). The modestrefractionsnowline title and depictedbelies intensity, the at intellectualagriculture, appropriate fauna,tality elevations. and onother mountain physical, In the snowline. chemical, original, and the biological profile features. is flankedtional on arrays both across sides theby tablesglobe took hold in the specimens and tens of thousands of measure-((22,, 33).). TheThe modestmodestdescribing titletitle belies belieselevational thethe intellectualintellectual patterns in temperature,talitytality onon mountainmountain humidity, snowline.snowline. light refraction and intensity,tionaltional agriculture,arraysarrays acrossacross fauna, thethe globeglobe and tooktook holdhold inin thethe richnessrichness within.within. InIn thethe texttext andand accompanyingaccompanying He expanded this vision in the succeeding 19thcentury.19thcentury. Throughout Throughout thethe century,century, countlesscountless ments. Humboldt spent the next 22 yearsrichness and within.other In the physical, text and chemical, accompanying and biologicalHe features.expanded this vision in the succeeding 19thcentury. Throughout the century, countless colorcolor plateplate (see(see thethe figure),figure), HumboldtHumboldt layslays outout years,years, establishingestablishing internationalinternational cooperativecooperative Humboldt-inspired explorations were launched, most of his inherited fortune in Paris,a vision pre- of a comprehensive “general physics networks of meteorological and geomagnetic each involving systematic measurement and In the early 19th century, Alexander von aa visionvision ofof aa comprehensivecomprehensive “general“general physicsphysics networksnetworks ofof meteorologicalmeteorological andand geomagneticgeomagnetic eacheach involvinginvolving systematicsystematic measurementmeasurement andand of the Earth”s scientists aimed at are nothing celebrating less than the a 200th syn- inhabited,measurement and hisstations, writings inventing inspired isothermsDarwin to atmosphericmapping of physical,and geophysical biological, measurements,and often cul- paring and publishing 45 volumes of aofof never- thethe Earth”Earth” distributionaimedaimed atat nothingnothing of less lessincident thanthan aa radiation,syn-syn- measurement the transport stations,serving inventing as isotherms both an indexmapping of climate of physical, and abiological, prox- and often cul- thesis ofanniversary atmospheric, of Charles oceanic,Darwin’s geological, birth embarkand otheron graphicalH.M.S. Beagle devices. More to pertinentportray spatial to our andtural recordedfeatures of the landscapes geographic and location oceans of(8 – their10). finished report on his travels. thesisthesis ofof atmospheric,atmospheric,of heat and oceanic,oceanic, materials geological,geological, in windsandand and otherother ocean graphicalgraphical cur- devicesdevicesimal toto control portrayportray spatialofspatial microclimate,turaltural featuresfeatures animal ofof landscapeslandscapes habitat, andand oceansoceans ((88––1010).). ecological,and and the cultural150th anniversary phenomena of acrossthe publi the- time,patterns, Humboldt and noting established that plantthe foundationform is often for thousandsThese surveys of specimens were relentlessly and tens inductive, of thousands typ- Humboldt laid the foundations for today’s Aecological,ecological, andand culturalcultural phenomenaphenomena acrossacross thethe patterns,patterns, andand notingnoting thatthat plantplant formform isis oftenoften These surveys were relentlessly inductive, typ- Of these volumes, the first was a slimcationglobe. workof Humboldt’s his Onrents, the Originobsession the influenceof Species with geographi-, Darwin’s of temperature thebetter Earth predicted onsystem plant bysciences: localand environment the culturalintegrated thanpracticessystem by ofically (6 measurements.–8 producing). detailedHumboldt descriptive spent the nextreports 22 globe.globe. Humboldt’sHumboldt’s obsessionobsession withwith geographi-geographi- betterbetter predictedpredicted byby locallocal environmentenvironment thanthan byby icallyically producingproducing detaileddetailed descriptivedescriptive reportsreports entitled Essay on the Geography ofideascally Plants continuereferencedform, to shape measurements and and the enrich effect andthe sciencesofcollec- latitude oftaxonomic knowledgeand continen- affinity on which (a humanparadoxHumboldt’s society resolved may dreamdeby- yearswith of little systematicand integrationmost of his observa- withininherited or amongfortune thein Paris,com- callycally referencedreferenced measurementsmeasurements andand collec-collec- taxonomictaxonomic affinityaffinity (a(a paradoxparadox resolvedresolved byby with little integration within or among the com- (2, 3). The modest title belies the intellectual(tions1). 6 wasMay central 2009tality marks to hison the vision.mountain 150th Heanniversary recognized snowline. of pendDarwin). in the Humboldt’sface of global geniustional climate layarrays change. in his across geo- thepreparingponent globe entities. andtook publishing However,hold in 45 thefor volumes a few intellectu- of a nev- Earth system sciences. thetionstions death waswas of centralcentral another toto 19th-century hishis vision.vision. HeHe figure—Alex recognizedrecognized- Darwin).Like Darwin, Humboldt’s Humboldt genius undertook lay in his a major geo- er-finishedponentponent entities.entities. report However,However, on his travels. forfor aa fewfew intellectu-intellectu- that spatial arrays of observations could be graphical vision, and in his intuition that ally nimble participants—including Charles richness within. In the text and accompanyinganderthatthat spatialspatial von Humboldt—whose arraysarraysHe ofof expanded observationsobservations scientific this couldcould vision legacy bebe invoyagegraphicalgraphical the succeeding that vision, vision,would shape and and 19thcentury. his in in hisideas his intuition intuition and thinking. Throughout that that allyallyOf thenimblenimble these century, volumes, participants—includingparticipants—including countless the first was a slim CharlesCharles work aggregated to reveal patterns that would in Earth’s land surface, oceans, atmosphere, and Darwin, T. H. Huxley, Matthew Maury, Asa color plate (see the figure), Humboldt alsolaysaggregatedaggregated flourishesout totoyears, reveal reveal in the establishing 21stpatternspatterns century. thatthat Humboldt wouldwouldinternational inin HumboldtEarth’s cooperative land spent surface, 5 years oceans,Humboldt-inspired (1799 atmosphere, to 1804) with and explorationsentitledDarwin, Essay T. H. were onHuxley, the launched, Geography Matthew ofMaury, Plants Asa (2, turn reveal underlying processes—such as the inhabitants form an integrated whole, with Gray, C. Hart Merriam, and Peter Kropotkin— helpedturnturn revealreveal create underlyingunderlying the intellectual processes—suchprocesses—such world Darwin asas thethe botanistinhabitantsinhabitants Aimé formform Bonpland anan integratedintegrated exploring whole,whole, Venezuela, withwith 3Gray,). The C. modest Hart Merriam, title belies and the Peter intellectual Kropotkin— rich- a vision of a comprehensive “general physics networks of meteorological andlinkages geomagnetic among the variouseach componentsinvolving systematic(4, these explorations measurement provided and data and experi- thelinkageslinkages northern amongamong Andes, thethe and variousvarious central componentscomponents Mexico, with ((44,, nessthesethese within. explorationsexplorations In the text providedprovided and accompanying datadata andand experi-experi- color of the Earth” aimed at nothing less than a syn- measurement stations, inventing5). Humboldt’s isotherms generalmapping physics of of the physical, Earth biological,ence that spurred and the often development cul- of biogeog- Botany Department and Program in Ecology, University of 55).). Humboldt’sHumboldt’s generalgeneral physicsphysics ofof thethe EarthEarth enceence thatthat spurredspurred thethe developmentdevelopment ofof biogeog-biogeog- BotanyBotany DepartmentDepartment andand ProgramProgram inin Ecology,Ecology, UniversityUniversity ofof visits to Tenerife, Cuba, and the United States. plate (see the figure), Humboldt lays out a vi- thesis of atmospheric, oceanic, geological,Wyoming, Laramie,and WY other 82071, graphical USA. E-mail: jackson@devices toenvisioned portray spatial climate astural a major features control of landscapes of raphy, ecology,and oceans oceanography, (8 –10). and other envi- Wyoming, Laramie, WY 82071, USA. E-mail: jackson@ Theyenvisionedenvisioned collected climate climate botanical, as as zoological, a a major major controlgeological, control of of sionraphy,raphy, of ecology,ecology,a comprehensive oceanography,oceanography,“general andand physics otherother of envi-envi- the uwyo.edu uwyo.edu Earth-surface phenomena, with vegetation ronmental sciences (11). PARIS MUSEUM, HISTORY NATURAL THE IN COPY ORIGINAL AN FROM REPRODUCED CREDIT: uwyo.edu Earth-surface phenomena, with vegetation ronmental sciences (11). PARIS MUSEUM, HISTORY NATURAL THE IN COPY ORIGINAL AN FROM REPRODUCED CREDIT: ecological, and cultural phenomena across the patterns, and noting that plantandEarth-surface form ethnological is often phenomena, specimens,These with made surveys vegetation extensive were relentlesslyEarth”ronmental aimed sciences inductive, at nothing (11). less typ- than a synthesis of PARIS MUSEUM, HISTORY NATURAL THE IN COPY ORIGINAL AN FROM REPRODUCED CREDIT: globe. Humboldt’s obsession with geographi- better predicted by local environment than by ically producing detailed descriptive reports 14 15 cally referenced measurements596 and collec- taxonomic affinity (a11 MAY MAYparadox 20092009 resolvedVOL 324324 bySCIENCEwith littlewww.sciencemag.org integration within or among the com- tions was central to his vision. He recognized Darwin). Humboldt’s genius lay in his geo- ponent entities. However, for a few intellectu- that spatial arrays of observations could be graphical vision, and in his intuition that ally nimble participants—including Charles aggregated to reveal patterns that would in Earth’s land surface, oceans, atmosphere, and Darwin, T. H. Huxley, Matthew Maury, Asa turn reveal underlying processes—such as the inhabitants form an integrated whole, with Gray, C. Hart Merriam, and Peter Kropotkin— linkages among the various components (4, these explorations provided data and experi- 5). Humboldt’s general physics of the Earth ence that spurred the development of biogeog- Botany Department and Program in Ecology, University of Wyoming, Laramie, WY 82071, USA. E-mail: jackson@ envisioned climate as a major control of raphy, ecology, oceanography, and other envi- uwyo.edu

Earth-surface phenomena, with vegetation ronmental sciences (11). PARIS MUSEUM, HISTORY NATURAL THE IN COPY ORIGINAL AN FROM REPRODUCED CREDIT:

596 1 MAY 2009 VOL 324 SCIENCE www.sciencemag.org 0227Books 2/20/09 1:38 PM Page 1170 0227Books 2/20/09 1:38 PM Page 1171

BOOKS ET AL. BOOKS ETAL.

HISTORY OF SCIENCE mer students who challenged ing the translation of Origin of Species into translations involved an attempt to recast exist- And so, finally, we come to Haeckel. The Tragic Sense of Life his ideas as they gained in- German, and (more intriguingly) diachroni- ing German terms in a newer, more up-to-date Gliboff’s key insight here is that Haeckel Making German Evolution: Ernst Haeckel and tellectual independence, and cally, as scientists reworked older words such mode that encompassed selection yet tamed originally read Bronn’s translation of Darwin, the Struggle over debated the pro-evolution as “perfection” and “type” to lend them new Darwin’s emphasis on unpredictability to meet not Darwin in the original. Gliboff shows Translation and Tragedy Evolutionary Thought (but anti-Haeckel) Jesuit priest meanings. Gliboff’s own clear, crisp prose is the more rigorous requirements of a German Haeckel as both echoing and responding to by Robert J. Richards and entomologist Erich Was- key to the success of this analysis, as he deftly academic scientist’s understanding of a “law” Bronn’s concerns, rather than either reflect- Lynn K. Nyhart mann—the list could go on leads his reader through dense philosophical of organic nature. Simultaneously, Bronn ing directly on Darwin’s original writing or University of Chicago and on. These were not iso- and terminological thickets with nary a thorn sought to translate Darwin’s ideas about selec- reaching directly back to the Romantic Press, Chicago, 2008. n this year of Darwin anniversaries (the duced important books. Robert 579 pp. $39, £27. lated episodes but rather scratch. This is some of the best close reading tion into a language without an exact equiva- embryologists. (Although Gliboff acknowl- 200th year of his birth and the 150th J. Richards, the director of the ISBN 9780226712147. moments in a lifelong cam- I have seen. It also represents a profound chal- lent for the term, and for an academic audience edges the centrality of monism to Haeckel’s Ianniversary of On the Origin of Species), University of Chicago’s Fish- paign to advance his philoso- lenge to our standard picture of 19th-century lacking the gentlemanly traditions of breeding thought, he focuses on the working evolution- The Tragic Sense of Life and H. G. Bronn, bein Center for the History of phy, which was accompanied German biology. pigeons and dogs so central to Darwin’s expo- ary theorist, not the popular ideologue.) Like Ernst Haeckel, and the Origins of German Science and Medicine and a H. G. Bronn, by a bitter hostility to orga- The old story, crudely put, is that Haeckel’s sition. The selection metaphor was further Bronn himself, Haeckel made further amend- Darwinism remind us that the history of evo- much-published author on Darwin Ernst Haeckel, and nized religion. version of evolution was a Darwinism in name fraught with an anthropomorphism foreign to ments both terminological and intellectual, lutionary thought in the 19th century extended and German Romantic biology, the Origins of Richards does not neg- only, best understood as an update on early- Germans, who were not brought up on British and Gliboff rereads Haeckel’s research pro- well beyond Darwin himself. Darwin did not has written a biography of Haec- German Darwinism lect Haeckel’s science pro- 19th-century idealistic morphologists such natural-theological assumptions about a per- gram as one not dominated by a typological launch his theory onto an unprepared public kel. Sander Gliboff, a professor A Study in Translation per, treating us to fascinat- as Carl F. Kielmeyer and J. F. Meckel that sonified God who had created a perfectly and linear-recapitulationist mindset but and scientific community, nor was the evolu- in Indiana University’s Depart- and Transformation ing and original discussions retained their teleology, their typological adapted nature. Bronn’s translation, though it rather as continuing to wrestle with the need tionism that developed after 1859 a mere ment of History and Philosophy by Sander Gliboff of his pathbreaking system- emphasis on form, and their linear recapitula- altered key ideas to make Darwin comprehen- to account for variability and unpredictable extension of his views—it was not even one of Science, places Haeckel at the atic and phylogenetic work tionism. This story, emphasizing the long per- sible to a German academic audience, was not change in terms of mechanistic laws of thing. How, then, should we think about the end of a study that examines the MIT Press, Cambridge, MA, on radiolaria and other marine sistence of a German transcendental approach a conservative throwback. It represented the nature—among which Haeckel included, at 2008. 271 pp. $35, £22.95. history of evolution in the 19th century? What larger process through which organisms, the importance to nature, has been deeply entrenched in the dynamic engagement of a leading paleontolo- the top of his list, natural selection. Haeckel’s ISBN 9780262072939. sorts of accounts best help us understand the Darwin’s words were translated, of linguistic analysis to his history of biology. gist who had also long been working on many Darwinism thus shows continuity with early- reception of Darwin’s theory, its relations to and his ideas modified, in the phylogenetic trees of the Gliboff challenges this history right from of the questions Darwin claimed as his own— 19th-century concerns, mediated through earlier ideas about nature, the directions that context of German biology. Both illuminate races of humans, and his remarkable experi- the beginning. The ascription of simple linear a critical yet generous equal, who saw himself Bronn. But those concerns were always more evolutionary investigation subsequently took, the twists and turns that evolutionary thought mental work with siphonophores. These con- recapitulationism to the views of Romantic as moving science forward through the modi- flexible than has been acknowledged, and and the relations of all of these to the broader took in Germany, but they do so in dramati- stitute important contributions to our under- embryologists, he notes, owes much to a carica- fications he made to Darwin’s flawed theory. their articulation changed over time. Of social, cultural, and religious concerns scien- cally different ways. standing of the technical development of ture developed by Karl Ernst von Baer in a Bronn’s death in 1862 afforded him little course Haeckel’s Darwinism was not tists shared with their contemporaries? Richards’s book, though over twice as long evolutionary biology. polemical context, then adopted uncritically by chance to steer the conversation further. Darwin’s own, but it was not an aberration or These questions become especially pointed as Gliboff’s, is the more entertaining read of The big picture here, however, is an argu- influential historians such as E. S. a distortion of some true theory, any more when one considers German Darwinism, and the two. In his characteristically rich and ment about the power of personality—at least Russell and Stephen Jay Gould. than any other post-Darwinian additions or especially Germany’s best-known follower of rolling prose, Richards weaves a compelling one personality—to shape the course of sci- Gliboff’s fresh reading of the orig- adjustments were. It was science moving on. Darwin, Ernst Haeckel. Most often remem- story of a life marked by tragedy and of an ence. In Richards’s presentation, German evo- inal sources interprets Kielmeyer Gliboff’s overall picture of scientific bered by biologists as the author of the bio- intense, larger-than-life figure whose passions lutionism was profoundly shaped by both and Meckel as far less rigidly advance, in contrast to Richards’s emphasis on genetic law (“ontogeny recapitulates phy- drove his scientific research and philosophy. In Haeckel’s charisma and his combativeness. typological in their orientation and charisma and passion, is one of scientists logeny”), Haeckel has also been accused of Richards’s rendering, the scientific Haeckel Perhaps the late-19th-century opposition of much more attentive to nature’s building and innovating incrementally, work- promoting European fascism via his monistic cannot be understood separately from the evolutionary science to Christianity would not variability than has been seen ing with what their predecessors have handed philosophy and of presenting a eugenic, bio- man’s personality and private circumstances. have been so fiery, he suggests, had Haeckel before. Both for these early-19th- them and sculpting it into something new yet logically determinist vision of humanity that His love of nature was surpassed only by his not continually fanned its flames. And century naturalists and for their understandable to those around them. His sen- led to Hitler’s “final solution.” Can one scien- love for his first wife, Anna Sethe, who died in although Richards absolves Haeckel of per- intellectual heirs, Gliboff argues, sitive reading allows us to see post-1859 tist be responsible for so much? Most histori- abdominal agony on his 30th birthday. Over sonal responsibility for fascism and Nazism, the critical issue was to understand German evolutionists as rational actors rather ans would say no, arguing that it takes a com- the next year, he wrote his way through the in part by situating him firmly in his time and nature’s manifold variety while than irrationally stuck in some early-19th- munity, rather than an individual, to make a despair that enveloped him, producing his place, he does show how the scientist’s ardent seeking out underlying strict natu- century moment with unmodern commit- movement; that single-cause explanations are foundational work, Generelle Morphologie temperament led him to the occasional intem- ral laws to account for it. ments. By challenging the very foundations of insufficient to account for something as broad (1). Although he remarried, the union was not perate statement that could be taken up by This provides a new starting the standard narrative of German morphol- as fascism; and that an individual cannot be happy, and passionate love would elude him extreme thinkers. One cannot leave this book point for analyzing Darwin’s first ogy, this careful, compelling account does at held responsible for the ways in which others until his sixties, when he had a secret affair that without a deep appreciation for Haeckel as a translator, the prominent paleon- least as much as Richards’s to undermine the (such as Hitler) took up his ideas and molded ended tragically with the death of his lover. tragic figure and for the force of personality in tologist H. G. Bronn—a figure lit- association of 19th-century German Darwin- them to new agendas after his death. But that Science remained his salvation and refuge. shaping the direction science may take. tle attended to in the standard ism with a dangerously exceptional view of still leaves open the questions of how to write His professional life was also filled with Gliboff’s account is of a completely differ- story but the lynchpin of Gli- nature. But the two books offer very different responsibly about what Haeckel actually drama, much of which centered on his philos- ent order. His is not a story of personalities or 1905) GERA-UNTERMHAUS, KOEHLER, (W. boff’s. Intriguingly and plausibly, reads. Is scientific progress a matter of per- believed and how we should situate him in the ophy of evolutionary monism—a science- private lives (although he mentions salient Gliboff argues that Bronn’s use sonal anguish and triumph, or of intellectual history of evolutionary thought. centered faith that became one of the most details), but of German academics seeking to of terms like “vervollkommnet” chugging along? Our concept of it should be

The historians under consideration here successful alternatives to the Judeo-Christian live up to the highest (if changing) ideals of WANDERBILDER (perfect) as translations for Dar- capacious enough to include both. have chosen two radically different strategies religion among those searching for a secular Wissenschaft and of the ways in which win’s “improved” or “favored” to understanding Haeckel’s place within spirituality. Haeckel could not turn down a Darwin’s theory was translated into this envi- were not about dragging Darwin References and Notes German evolutionism, and both have pro- fight: He battled the physician-statesman ronment. He thus situates Haeckel at the end backward into a German teleo- 1. E. Haeckel, Generelle Morphologie der Organismen (Georg Reimer, Berlin, 1866). Rudolf Virchow over the role of evolution in of a revised intellectual history of 19th- logical view of nature (as has 2. The reviewer previously served as a press reader for both the schools (Haeckel argued that it should century German evolutionism. Central to his been claimed by those who have books at the manuscript stage. The reviewer is at the Department of the History of Science, University of Wisconsin, Madison, WI 53706–1393, USA. replace religious education), sparred with reli- account is the idea of translation, which he paid attention to Bronn at all). A painter, too. Haeckel’s oil landscape of highlands in Java, from E-mail: [email protected] 10.1126/science.1169621 giously conservative scientists and with for- uses both synchronically, especially in treat- HAECKEL/FROM ERNST CREDIT: Instead, Gliboff asserts, Bronn’s Wanderbilder (1905). 16 17 1170 27 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 323 27 FEBRUARY 2009 1171 REVIEW

of the individual variants in a population was the relations among species. Several proposals his theory to predict an orderly pattern of essentially undirected ruled out any possibility available in the 1830s deflected attention away relations. that evolution could be shaped by a predeter- from the model of the branching tree (11). It has been argued that Darwin’s move to a REVIEWmined developmental trend. There was no ob- William Sharp Macleay’s quinary or circular more historical viewpoint was inspired by German vious goal toward which it was aimed, and it system of classification supposed that every romanticism [e.g. (12)], but a more practical didof not the produce individual an variants orderly in pattern a population of relations was thegenus relations contained among five species. species Several that proposals could behis ar- theoryincentive to predict was provided an orderly by pattern the biogeographical of betweenessentially species. undirected The accusation ruled outthat any possibilitythe theory availableranged in in a the circle; 1830s each deflected family attention five genera, away andrelations. so insights gained on the Beagle voyage (1831–36). ’ dependedthat evolution on “random could” variation be shaped indicated by a predeter- the con- fromon through the model the taxonomic of the branching hierarchy. tree Chambers (11). ’sIt hasThe been Galapagos argued that species Darwin provideds move the to most a obvious mined developmental trend. There was no ob- William Sharp Macleay’s quinary or circular more historical viewpoint was inspired by German cerns of his opponents on this score. As Darwin Vestiges of the Natural History of Creation example of how the relations within a group can vious goal toward which it was aimed, and it system of classification supposed that every romanticism [e.g. (12)], but a more practical himselfdid notmade produce clear, variationan orderly was pattern certainly of relations caused genusdepicted contained evolution five species in terms that of could parallel be linesar- ad-incentivebe was explained provided by by supposing the biogeographical that an original pop- by somethingbetween species. (later The identified accusation as that genetic the theory muta- rangedvancing in a circle;through each a family predetermined five genera, sequence and so insights of ulation gained on became the Beagle dividedvoyage up, (1831 in this–36). case by in- tions),depended but it on was“random not aimed” variation in any indicated one direction the con- onstages through within the taxonomic each family, hierarchy. driven Chambers by force’ deriveds The Galapagosdependent species acts provided of migration the most to obvious oceanic islands. REVIEW and,cerns thus, of left his adaptive opponents evolution on this score. essentially As Darwin open- Vestigesfrom individual of the Natural development. History of Creation exampleHere, of how Darwin the relations followed within Lyell a group in seeing can that bio- REVIEW ended.himself He made allowed clear, a variation limited was role certainly for variation caused depicted evolution in terms of parallel lines ad- be explained by supposinggeography that an must original become pop- a historical shapedby something by the organisms (later identified’ own activities as genetic (the muta- so- vancing through a predetermined sequence of ulation became dividedscience, up, in explaining this case by present in- distribu- opments that would push other naturalists toward calledtions), Lamarckian but it was effect), not aimed but in this any too one permitted direction stages within each family, driven by force derived dependent acts of migrationtions in to terms oceanic of islands. past migrations, REVIEW an evolutionary vision during the years he worked multipleand, thus, vectors left adaptive of change. evolution Evolution essentially had open- to be from individual development. Here, Darwin followedextinctions Lyell in seeing and (for that Darwinbio- but not opments that would pushREVIEW other naturalists toward ended. He allowed a limited role for variation geographyREVIEW must become a historical in isolation. By the late 1850s, the idea of pro- depicted as a branching tree in which each act of for Lyell) evolutionary adaptations. an evolutionary vision during the years he worked shaped by the organisms’ own activities (the so- science, explaining present distribu- ’ gressive evolution was widely recognized, and the branching was the result of a more or less un- Populations divided by geographical Darwinof thes individual Originality variants in a population was the relations among species. Severalinof isolation.progressive proposals By evolution thehis late theory 1850s,was towidely the predict idearecognized, ofan pro- orderly pattern of of thecalled individual Lamarckian variants effect), in but a this population too permitted was tionsIt has in termsbeen ofargued past migrations,that Darwin’s move opmentspositive role that ofwould individual push other competition naturalists was toward being predictable migration of organisms to a new loca- barriers will develop independently Darwinessentially’s Originality undirected ruled out any possibility available in the 1830s deflected attentiongressiveand the away evolutionpositiverelations. role was widelyof individual recognized, competition and the essentiallymultiple undirectedvectors of change. ruled Evolution out any had possibil to be- opments that would pushtoextinctions a more other naturalistshistorical and (for Darwin viewpoint toward but not was inspired Peter J. Bowlerthat evolution could be shaped by a predeter- from the model of the branchinganarticulatedevolutionary tree (11 by). thinkers visionIt during has such been the as arguedHerbert years he that Spencer worked Darwin’s move to a tion. At the same time, Darwin’s theory undermined as each adapts to its new environ- positivewas being role articulated of individual by thinkers competition such wasas Herbert being ity thatdepicted evolution as a branching could be tree shaped in which by each a prede act of- an evolutionary visionbyfor during German Lyell) the evolutionary years romanticism he worked adaptations. [e.g. (12)], but a mined developmental trend. There was no ob- William Sharp Macleay’s quinaryin(Fig. or isolation. circular1). But By keymore the aspects late historical 1850s, of theviewpoint Darwinian the idea ofwas vision pro- inspired by German thebranching old idea thatwas the species result were of a more idealized or less types, un- Populationsment divided in its by own geographical way, and the pos- Peter J. Bowler articulatedSpencer (Fig. by 1 thinkers). But key such aspects as Herbert of the Darwin Spencer- termined developmental trend. There was no in isolation. By the latemore 1850s, practical the ideaincentive of pro- was provided by Charles Darwin’s theory’ of natural selection has been hailed as one of the most innovative gressivewere truly evolution original was and widely would recognized, not have occurred and the fixedpredictable elements migration in a clearly of organisms defined to natural a new order. loca- barriers willsibility develop that independently barriers can be crossed Darwinviouss goal Originality toward which it was aimed, and it system of classification supposed(Fig. that 1). every But keyromanticism aspects of the [e.g. Darwinian (12)], vision but a moreDarwin practical ’s Originality gressive evolution was widely recognized, and the contributions to modern science. When first proposed in 1859, however, it was widely rejected positiveianto any vision other role were of naturalist individualtruly original at the competition time.and would Here, was not Wallace being have obvioustion. At goal the same toward time, which Darwin it’s theory was aimed, undermined and theas eachbiogeographical adapts to its newinsights environ- gained on the Charles Darwin’sdid theory not ofproduce natural an selection orderly patternhas been of hailed relations as onegenus of the contained most innovative five species thatwere could truly be ar-originalincentive and would was not provided have occurred by the biogeographical Species had to be treated as populations of vary- positive role of individual competitionoccasionally was allows being for the branch- by his contemporaries, even by those who accepted the general idea of evolution. This article articulatedoccurredprovides ato good byany thinkers comparison:other naturalist such He as tooat Herbert the moved time. Spencer towardHere, – ingit didthe individuals, not oldproduce idea thatwith an species orderly no fixed were pattern limit idealized on of therelations types, range Beaglement in voyage itsing own process way,(1831–36). of and evolution the The pos- thatGalapagos Darwin contributionsPeter J. Bowler tobetween modern science.species. WhenThe accusation first proposed that the in 1859,theory however,ranged it in was a circle; widely each rejected family five genera,to any andother so naturalistinsights at thegained time. on Here, the Beagle WallacevoyagePeter (1831 J.36). Bowler articulated by thinkers such as Herbert Spencer identifies those aspects of Darwin“ ’s work” that led him to develop this revolutionary theory, Wallace(Fig.the idea 1). But providesof’ branching key aspects a good evolution of comparison:the Darwinian driven by He vision local too ofbetween possiblefixed elementsspecies. variation. inThe a clearlyaccusation defined that natural the order.theo- speciessibility thatprovidedconceived barriers the can in most the be crossed lateobvious 1830s. example It was by his contemporaries,depended even on byrandom those whovariation accepted indicated the the general con- ideaon ofthrough evolution. the taxonomic This article hierarchy.provides Chambers a goods comparison:The Galapagos He too species moved provided toward the most obvious (Fig. 1). But key aspects of the Darwinian vision including his studies’ of biogeography and animal breeding, and his recognition of the role played movedadaptation, toward butthe even idea he of did branching not share evolution Darwin’s ry dependedSpecies had on to be “random” treated as variation populations indicated of vary- ofoccasionally how the relations allows for within the branch-a group can be identifiesCharles Darwin thoses aspectscerns theory of of ofhis Darwin natural opponents’s selection work on that this has ledscore. been him As hailedto Darwin develop as one thisVestiges of revolutionary the most of the innovative Natural theory, Historythewere of idea Creation truly of original branchingexample and evolution would of how not driven the have relations occurredby local within aCharles group canDarwin’s theory of natural selection has been hailed as one of the most innovative were truly original and wouldby not approaching have occurred the problem of the by the struggle for existence. driveninsight thatby local the work adaptation, of the animal but even breeders he did throws not Thethe ing concerns Tree individuals, of Life of his with opponents no fixed limit on on this the score.range explaineding process ofby evolution supposing that Darwin that an original includingcontributions his studies tohimself modern of biogeographymade science. clear, When variation and first animal was proposed certainly breeding, in caused 1859, and however, hisdepicted recognition it evolutionwas of widely the in role rejected terms played of paralleladaptation,to any lines other ad- but naturalist evenbe explained he at did the time.not by share supposing Here, Darwin Wallace that’s an originalcontributions pop- to modern science. When first proposed in 1859, however, it was widely rejected to any other naturalist at the time.origin Here, of new Wallace species through a study share Darwin’s insight that the work of the As of Darwin possible himself variation. made clear, variation was populationconceived in became the late divided 1830s. It up, was in this case by his contemporaries,by something even by (later those identified who accepted as genetic the general muta- ideavancing of evolution. through This a predetermined article provideslight sequence on athe good of process comparison:ulation of natural became He selection. too divided moved up,toward in this caseby his by contemporaries, in- One innovation even by atthose the heartwho accepted of Darwin the’s theorygeneral idea of evolution. This article provides a good comparison:by approaching Heof too biogeography moved the problem toward that of the Darwin was by the struggle for existence. ’ insight that the work of the animal breeders throws identifieshe those publicationtions), aspectsof but of Charles it Darwin was not’ Darwins work aimed thats inOn anyled him onepre-existing directionto develop ones thisstages in arevolutionary progressive within each sequence theory,family, driven lead- byanimalthe forceThe idea derived breeders theory of branching was dependentthrows both evolution light original acts on of driven andthe migration process disturbing. by local toof oceanic islands. certainlyseemsThe so Tree causedobvious of Life by’ today something that it is (later hardidentified for us to byorigin independent of newled to species construct acts through of his migration a study model ofto open- oce- light on the process of natural selection. identifies those aspects of Darwin s work that led him to develop this revolutionary theory, the idea of branching evolution driven by local includingthe his Origin studiesand, of thus, ofSpecies biogeography leftin adaptive 1859 evolutionandis widely’ animal essentially breeding,ing upopen- to and humans hisfrom recognition (5). individual But if the of development. the general role ideaplayed of naturaladaptation,It was notselection. just but thatevenHere, the he Darwin idea did not of followed natural share Darwin selection Lyell in’s seeingincluding that bio- his studiesasappreciate geneticOne of biogeographyinnovation mutations), just how at the new andbut heart animalit and ofwas Darwin how not breeding, radicalaimed’s theory andin it his recognition of the role played adaptation, but evenanicof he biogeography didislands. notended, share Here, divergent that Darwin Darwin Darwin evolution.’s followed was Wallace Lyell he publication of Charles Darwin s On pre-existing ones in a progressive sequence’ lead- The theory was both original and disturbing. by thesupposedhe struggle publication forended. to existence. have of He Charles initiated allowed Darwin’s a a revolution limited On role of forevolution evolution variation was was not not entirely entirely new, new, Darwin Darwin’ss visionvi- insightchallengedThe thattheory the the workwas belief both of that the original the animal worldgeography and breeders was disturbing. designed must throws become by a historical the struggleanywas forseems existence.one at the direction so time. obvious Lamarckand, today thus, that had left it is proposed adaptive hard for us evo that to- insight that the work ofinled the seeing to animal constructdeveloped that breedersbiogeography his model a similar throws of open- model must and become tested Tthe Origin of Species in 1859 is widely ing up to humans (5). But if the general idea of It was not just that the idea of natural selection both inthe science Origin and of in Species Westernin culture. 1859 is Yet widely there sionof how of how the processthe process worked worked certainly certainly was. was. Al- Itby was a wise not and just benevolentthat the idea God. of Therenatural was selection a wider lutionappreciate essentially just howopen-ended. new and He how allowed radical it a aended, historical divergent science, evolution. explaining Wallace present supposedshaped to have by initiated the organisms a revolution’ own activitiesevolution (the wasso- not entirely new, Darwin’s vision challengedlight on the the process belief of that natural the worldscience, selection. was explaining designed present distribu- there might be natural processes adapting spe- light on the process of naturalit selection. during his explorations in South Thave beensupposed frequent to haveclaims initiated that Darwinism a revolution’ was Althoughthough the the theory theory was was eventually eventually paralleled paralleled by challengedelement of the teleology belief that or goal-directedness the world was designed almost limitedwas role at the for time. variation Lamarck shaped had proposedby the organ that- distributionsdeveloped a similar in terms model ofand past tested migrations, both inhe science publication andcalled in Western ofLamarckian Charles culture. Darwin effect), Yet theresbutOn this tooofpre-existing how permitted the onesprocess in a worked progressive certainly sequence was. lead- Al- by aThe wise theory and benevolent was both God.originaltions There and in was disturbing. terms a wider of past migrations,he publicationcies to ofchanges Charles in their Darwin environment.’s On pre-existing But Darwin ones in a progressive sequence lead- The theory was both originalAmerica and disturbing. and the Malay Archipelago T “ ” — there might be natural processes adapting spe- it during his explorations in South bothsomehowinthe science Originin the andmultiple air of in Species atWestern the vectors time,in culture. 1859 of merely change. isYet waitingwidely there Evolution ingbyWallace, Wallace, had up to to Darwin humans be Darwin had (5 had). conceived But conceived if the its generalits basic basic idea outline out of- Itbyuniversally was a wise not just acceptedand that benevolent the at the idea time. ofGod.extinctions natural Most There thinkers selection and was (fora Darwin but not isms’was perhaps own activities the first (the to realize so-called that ifLamarckian adaptation extinctions(modern and Indonesia). (for Darwin but not for have been frequent claims that Darwinism was though the theory was eventually paralleled by element of teleology or goal-directedness almost the Origin ofcies Species to changesin 1859 in their is environment. widely ing But up Darwin to humans (5). But if the general idea of It was not just that theAmerica idea of and natural the Malay selection Archipelago havefor someone beensupposed frequent to put to aclaims have few readily initiated that Darwinism available a revolution points was evolutionlinein the in late the was 1830s,latenot 1830s, after entirely hisafter return new, his Darwinfromreturn the from’s voyage vision the widerchallengedincluding element Jean-Baptiste the beliefof teleology that Lamarck the or world goal-directedness and was Chambers designed— effect),to the local but environment this too permitted was the multiple only mechanism vectors ’ Lyell) evolutionaryAdrian Desmond adaptations. and Popula James- somehow “in thedepicted air” at the as time, a branching merely treewaiting in whichWallace, each act Darwin of had conceived its basic outline universally accepted at the time.for Most Lyell) thinkers evolutionary— adaptations.supposed to have initiated a revolution evolution was not entirely new, Darwin s vision challenged the belief that(modern the world Indonesia). was designed T T was perhaps the first to realize that if adaptation together in the right way [for instance (1)]. The of H.M.S. Beagle. He worked on it in relative took it for granted that the development of life on forbothsomehow someone in science “in to the put andbranching air” a in few Westernat readilythe was time, culture. the availablemerely result Yet points ofwait there a- moreinofvoyageor the how less late of theun- 1830s, H.M.S. process after Beagle worked his return. He certainly fromworked thewas. on voyage it Al- in includingalmostby a wise universally Jean-Baptiste and benevolentaccepted Lamarck God. atPopulations There and the Chambers wastime. a divided wider Most— by geographicalboth in science andof evolution, change.to in the Western local Evolution there environment culture. wouldhad Yet was bethere to themajor be only depictedof implications mechanism how as the a process worked certainly was. Al- by a wise and benevolenttionsAdrian God.dividedMoore There Desmond by havewas geographical a recently widerand James proposedbarriers will that fact that Alfred Russel Wallace (Fig. 1) indepen- isolation over the next 20 years, until the arrival of earth represents the unfolding of a coherent plan ’ togetherhaveing for been someone in frequentthe rightpredictable to claimsput way a [forfew migrationthat instance readily Darwinism of (available1 organisms)]. The was toofthoughrelative a H.M.S. new theisolation loca-Beagle theoryover. was He the worked eventually next 20 on years, paralleled it in until relative the by tookelementthinkers—including it for of granted teleology that Jean-Baptiste or the goal-directedness developmentbarriersLamarck will of lifealmost develop and on independentlyhave been frequentforbranching theof claims evolution, whole tree that system in there Darwinism which would by which each be was major act species ofthough implicationsbranching are theclas- theory was eventually paralleled by element of teleologydevelop orMoore goal-directedness have Darwinindependently recentlys hatred proposed almostas of each slavery that adapts prompted to its dently formulated“ a” theory of natural selection in Wallace’s paper in 1858 precipitated the flurry of aimed at a predetermined goal. (This assumption— ’ factsomehowpoints thattogether Alfredin the Russel intion. air the Atat Wallace right the same time, way (Fig. time, merely [for 1) Darwin indepen- instance waiting’s theory isolationWallace,arrival undermined of over Darwin Wallace’s the next had paper 20conceived years,in 1858 until its basic theprecipitated arrival outline of earthuniversallyChambers—took represents accepted the it unfoldingfor at thegranted time.as of that Most eacha coherent the thinkers adapts develop plan to- its newsomehow environ-“in thewassified airfor ”theat into the theresult whole groups. time, of system merelya Darwin more by waiting whichors mentorless species unpredictableWallace, in geology,are clas- Darwin had conceived its basic outline universally acceptednew atDarwin the environment time.’shis hatred Most move ofthinkers slavery towardin its prompted—own evolutionism way, and ( 13the). 1858 is taken as evidence for this position. But activity leading to the publication of the Origin. is still preserved in the very term “evolution”; the— ’ dentlyfor(1)]. someone The formulated fact to that putthe aAlfred a theory old few idea readily ofRussel natural that available Wallace species selection points were(Fig. in idealizedWallaceinthe the flurry late types,’s paper 1830s, of activity in after 1858 hisleading precipitated return to from the the thepublication flurry voyage of aimedincludingment of at life a Jean-Baptiste predetermined on earth represents Lamarck goal.ment (This andthe inunfolding Chambers assumption its own of way, andfor thesomeone pos- to putmigrationCharles asified few Lyell, readilyintoof groups.organisms had available shown Darwin to how points as mentor newhis uniformitarian in location. in the geology, late At 1830s, after his return from the voyage including Jean-Baptistepossibilityhis Lamarck moveBecause toward andthat Chambersbarriers evolutionism many slaveholderscan— (be13). crossed argued oc- Darwin had created the outlines of the theory Historians have quarried Darwin’s notebooks Latin evolutio refers to the unrolling of a scroll.) 1858together1) independently is taken in the as rightfixed evidenceformulated way elements [for for a instancethis theory in a position. clearlyof (1 natural)]. defined ButThe activity naturalof the H.M.S. Origin order. leadingBeagle. to. the He publication worked on of it the inOrigin relative. isatook stillcoherent it preserved for granted plan inaimed that the very theat a development termpredeterminedsibility“evolution that of barriers life ”goal.; the on cantogether be crossed in the rightthetheory sameCharles way would time, [for Lyell, instance allowDarwin’s had shown the (1 )].theory biogeographer how The hisundermined uniformitarianof H.M.S. to the re-Beagle . He worked on it in relative took it for granted thatcasionallyBecause the developmentthat many allows the slaveholders blackfor of lifethe race onbranching argued was separately process 20 years earlier, and there were significant dif- and letters to establish the complex process by The explanatory framework centered on the theory would allow the biogeographer to re- that the black race was separately Darwinfactselection that had Alfred in 1858 created RusselSpecies is thetaken Wallace had outlines as to evidence be(Fig. oftreated 1) the indepen- for theory as populationsthis isolationHistorians of vary- over the have next quarried 20 years, DarwinDarwin’s until’ thes notebooks arrival of Latin(Thisearth representsevolutio assumptionrefers the is unfolding to still the preserved unrollingoccasionally of a coherent ofin a the scroll.) allows veryplan forfact the branch- that Alfred Russeloldconstruct idea Wallace that the species migrations (Fig. 1)were indepen- of idealized speciesisolation ontypes, an fixed ever- over the next 20 years, until the arrival of earth represents the unfoldingof evolutioncreated of a that coherent fromDarwin plan the conceived white, Darwinin the dently formulateding a theory individuals, of natural with selection no fixed in limitWallace on the range’s paper in 1858 precipitated the flurry of aimed at a predetermined goal.ing (This process assumption of evolutiondently that Darwin formulatedelementschanging aconstruct theory in earth. of a the naturalclearly migrations Populations selectiondefined of species in couldnaturalWallace on sometimesorder. an ever-’ sSpe paper- in 1858 precipitated the flurry of aimed at a predeterminedlatecreated 1830s. goal.wanted from (This It was the assumptionto by white, show approaching that Darwin all races the prob share- 20 years earlier, and there were significant dif- and letters to establish the complex process by The explanatory framework centered on the changing earth. Populations could sometimes wanted to show that all races share 1858 is taken asof evidence possible for variation. this position. But activity leading to the publication of the Origin. is still preserved in the very termconceived“evolution in” the; the late 1830s. It was ciesbecome had dividedto be treated by geographical as populations barriers, of varying so that lem of thea common origin“ of ancestry, new” species and he throughrealized 1858 is taken as evidencebecome for divided this by position. geographical But barriers,activity so leadingthat to the publication of the Origin. is still preserved in thea very common term ancestry,evolution and he; the realized Darwin had created the outlines of the theory Historians have quarried Darwin’s notebooks Latin evolutio refers to the unrolling of a scroll.) what was once a single species could split into ’ that this claim could be defended by approaching the problemDarwin of thehad createdindividuals,what the was outlines with once no aof single fixed the theory specieslimit on could the Historians splitrange into of have quarried Darwin s notebooks Latin evolutio refersa tothat study the this unrolling of claim biogeography could of a scroll.) be defended that Darwin was The Tree of Life 20 years earlier, and there were significant dif- and letters to establish the complex process by The explanatory frameworkorigin centered of new on species the through20 years a study earlier,possiblemultiple andmultiple there variation. branches were branches significant adapting adapting to dif- to separate separateand environ- letters to establish the complex process by The explanatory frameworkledby to extending constructby centered extending the his idea onmodel throughout the the ideaof open-ended, throughout One innovation at the heart of Darwin’s theory of biogeography that Darwin was ments ments (10 ().10 Evolution). Evolution would would become become a divergent divergentthe animalthe evolution.kingdom. animal kingdom. As Wallace a basisfor As developed a basis for a seems so obvious today that it is hard for us to led to construct his model of open- Theprocess,process, Tree with of with Life some some branches branches splitting splitting over and and similarhis thinking, modelhis thinking, this and thesis tested this is sureit thesis during to is surehis ex to- appreciate just how new and how radical it ended, divergent evolution. Wallace Oneoverover innovation again, again, whereas whereas at the others othersheart cameof came Darwin’s to to a dead theory end end Fig.Fig. 2. 2.TreeTree of Life,of Life, from from Darwin Darwin’s notebooks’s notebooks (22). (22). plorationsgenerategenerate much in South controversy, much America controversy, but and if the Malay but if was at the time. Lamarck had proposed that developed a similar model and tested seemsthroughthrough so extinction. obvious extinction. today that it is hard for us to Archipelagoacceptedaccepted it would (modern it emphasize would Indonesia). emphasize the the The image of the tree of life had appeared in These rigidly structured models of taxo- crucial role played by his move toward a model there might be natural processes adapting spe- it during his explorations in South appreciateThe image just ofhow the new tree and of life how had radical appeared it was in TheThese idea rigidly of common structured descent models now seems of taxo- so crucialAdrian role Desmond played by hisand move James toward Moore a model have Darwin’s notebooks of the late 1830s (Fig. 2) nomic relations and evolution made good sense of branching evolution based on geographical cies to changes in their environment. But Darwin America and the Malay Archipelago Darwin’s notebooks of the late 1830s (Fig. 2) nomic relations and evolution made good sense of branching evolution based on geographical at theand time. was proposed Lamarck independently had proposed by Wallace that there in toobvious anyone embeddedthat we might in a visionwonder of what nature alternative as a diversity. recently proposed that Darwin’s hatred of slav- was perhaps the first to realize that if adaptation (modern Indonesia). and was proposed independently by Wallace in to anyone embedded in a vision of nature as a diversity. mighta paper be natural published processes in 1855. adapting Both realized species that itto predictable,models could orderly have system been governed proposed by to a divineaccount for Thisery model prompted was so his radical move that toward many late evolutionism a paper published in 1855. Both realized that it predictable, orderly system governed by a divine This model was so radical that many late to the local environment was the only mechanism Adrian Desmond and James changesexplained in their why environment. naturalists were But able Darwin to arrange was plan.the Suchrelations a world among view species. made it difficultSeveral to proposals ac- 19th-century (13). evolutionistsBecause many were unableslaveholders to accept argued that explained why naturalists were able to arrange plan. Such a world view made it difficult to ac- 19th-century evolutionists were unable to accept of evolution, there would be major implications Moore have recently proposed that perhapsspecies the into first groups to realize within groups,that if adaptation using descent to ceptavailable that the in history the 1830s of life deflected on earth mightattention be awayit in full.the Ernst black Mayr race argued was separately that the theory created of from the ’ species into groups within groups, using descent cept that the history of life on earth might be it in full. Ernst Mayr argued that the theory of for the whole system by which species are clas- Darwin s hatred of slavery prompted the localfrom aenvironment common ancestor was the to explain only mechanism the under- essentiallyfrom the irregular model andof the unpredictable, branching dependanttree (11). Wilcommon- white, descent Darwin was one wanted of Darwin to ’ shows greatest that all races ’ from a common ancestor to explain the under- essentially irregular and unpredictable, dependant common descent was one of Darwin’s greatest sified into groups. Darwin s mentor in geology, his move toward evolutionism (13). of evolution,lying similarities. there would Closely be related major species implications have di- onliam the Sharp hazards Macleay’s of migration, quinary isolation, or circular and local systemachievements, share a incommon addition ancestry, to natural and selection he realized that Because many slaveholders argued lyingverged similarities. recently Closely from a common related ancestor, species whereashave di- adaptation.on the hazards Darwin of was migration, led toward isolation, his alterna- and localitself (14achievements,). So it was, but in addition I think Mayr to natural over- selection Charles Lyell, had shown how his uniformitarian for the whole system by which species are clas- of classification supposed that every genus con- this claim could be defended by extending the theory would allow the biogeographer to re- that the black race was separately vergedthe recentlyancestry offrom more acommon distantly related ancestor, forms whereas must tiveadaptation. model in Darwinpart because was he led was toward more his inter- alterna-estimateditself the ( rapidity14). So with it was, which but other I natural-think Mayr over- Fig. 1. Charles Darwin, Alfred Russel Wallace, and Herbert Spencer. sifiedbe tracedinto groups. further backDarwin’s downmentor the family in geology, tree to estedtained in adaptationfive species than cosmicthat could teleology, be arranged thanks inists a wereidea converted throughout to the the theory. animal Many kingdom. of the As a ba- construct the migrations of species on an ever- created from the white, Darwin the ancestry of more distantly related forms must tive model in part because he was more inter- estimated the rapidity with which other natural- Charlesfind theLyell, common had shown point of how origin. his uniformitar- tocircle; the influence each family of Williamfive genera, Paley’s naturaland so the-on throughnon-Darwinian sis for his theories thinking, of evolution this thesis proposed is sure to gener- Fig. 1. Charles Darwin,changing Alfred earth. Russel Populations Wallace, could and Herbert sometimes Spencer. wanted to show that all races share be traced further back down the family tree to ested in adaptation than cosmic teleology, thanks ists were converted to the theory. Many of the position. But Darwin had created the outlines of and letters to establish the complex process by term “evolution”; the Latin evolutio refers to the ian theoryThe idea would of common allow the descent biogeographer now seems so to ology.the taxonomic Natural selection hierarchy. replaced Chambers’s divine benev- Vestigesduring of theate“ mucheclipse controversy, of Darwinism ”butin the if lateaccepted 19th it would ferences betweenbecome the ways divided in by which geographical he and barriers,which so he that developed his theory (6–9). Darwin theory of natural selection challengeda common this ancestry, vision and he realized find the common point of origin. to the influence of William Paley’s natural the- non-Darwinian theories of evolution proposed the theory 20 years earlier, and there were sig- which he developed his theory (6–9). Darwin unrolling of a scroll.) The explanatory framework reconstructobvious thatthe migrations we might wonder of species what on alternative an ever- olencethe Natural as an explanation History of ofCreation adaptation. depicted Unlike evolucentury- emphasize were introduced the crucial with the role aim ofplayed subvert- by his move Wallace formulatedwhat their was ideas. once a In single this essay, species I couldwas split a highly into creative thinker who synthesized– a of nature as an orderly patternthat of relations. this claim could be defended The idea of common descent now seems so ology. Natural selection replaced divine benev- during the “eclipse of Darwinism” in the late 19th ferencesnificant betweendifferences the between ways inthe which ways in he which and whichwas a highly he developed creative his thinker theory who (6 synthesized9). Darwin theorycentered of on natural the theory selection of natural challenged selection this vision chal- changingmodels earth. could have Populations been proposed could to account sometimes for Macleaytion in terms and Chambers, of parallel Darwin lines didadvancing not expect throughing thetoward implications a model of the of principle branching of commonevolution basedCREDIT: SYNDICS OF CAMBRIDGE UNIVERSITY LIBRARY on Fig.argue 1. thatCharles Darwinmultiple Darwin, was truly Alfred branches original Russel adapting in Wallace,his think- to separate andnumber Herbert environ- of Spencer. key insights, some derived from his Darwin’s world view wasby profoundly extending differ- the ideaFig. throughout 1. Charlesobvious Darwin, that Alfred we Russel might Wallace, wonder whatand Herbert alternative Spencer.olence as an explanation of adaptation. Unlike century were introduced with the aim of subvert- Wallacehe and Wallace formulated formulated their ideas. their ideas. In this In essay, this es I- wasa number a highly of key creative insights, thinker some who derived synthesized from his a oflenged nature this as vision an orderly of nature pattern as ofan relations.orderly pattern become divided by geographical barriers, so a predetermined sequence of stages within each geographical diversity.

ing, and I support this claim by addressing the scientific work and others from currents circulat- ent because he argued that the adaptation of pop- CREDIT: SYNDICS OF CAMBRIDGE UNIVERSITY LIBRARY argue that Darwinments was truly (10). original Evolution in wouldhis think- becomenumber a divergent of key insights, some derived from his Darwin’s world view wasthe profoundly animal kingdom. differ- As a basis for models could have been proposed to account for Macleay and Chambers, Darwin did not expect ing the implications of the principle of common say,related I argue issue that of definingDarwin was just truly why theoriginal theory in washis scientificing in his culturalwork and environment. others from Few currents– would circu now- ofulations relations. to their local environment was the sole 224that what was once a single species could split family, driven by forceSCIENCE derived from individual This model was so radical that many late ing,ferences and I between supportprocess, thisthe waysclaim with byin some whichaddressing branches he and the splittingscientificwhich over he workand developed and others his fromtheory currents (6 9). circulat- Darwin enttheory because of natural he argued selection that the challengedhis adaptation thinking, this of vision pop-this thesisferences is sure tobetween the ways in which he and which9 JANUARY he developed 2009 his VOL theory 323 (6–9). Darwinwww.sciencemag.orgtheory of natural selection challenged this vision thinking,so disturbing and toI support his contemporaries. this claim by addressing latingaccept in the his claim cultural that environment. evolution by Few natural would se- causeDarwin’s of transmutation. world view Many was profoundly people found dif it- into multiple branches adapting to separate en- development. 19th-century evolutionists were unable to accept relatedWallace issue formulated of definingover their again, just ideas. whywhereas In the this theory others essay, was came I toingwas a indead a his highly end cultural creativeFig. environment. 2. thinkerTree ofwho Few Life, synthesized would fromDarwin now a ’sulationsof notebooks nature to as their (an22 orderly). local environment patterngenerate of relations. was much the sole controversy,Wallace but formulated if their ideas. In this essay, I was a highly creative thinker who synthesized a of nature as an orderly pattern of relations. the Darwinrelated issue was of certainly defining not just the why first the totheory sug- nowlection accept was the in theclaim air. that Darwin evolution approached by natural the ferenthard to because see’ natural he argued selection that as the agentadaptation of either of 224 vironments (10). Evolution would become9 JANUARY a 2009These rigidly VOL 323 structuredSCIENCE models ofwww.sciencemag.org taxonomic it in full. Ernst Mayr argued that the theory of soargue disturbing that Darwin to histhrough was contemporaries. truly extinction. original in his think- acceptnumber the of claimkey insights, that evolution some derived by natural from se-his causeDarwin of transmutation.s world view Many wasaccepted profoundly people foundit differ- would it emphasizeargue that the Darwin was truly original in his think- number of key insights, some derived from his Darwin’s world view was profoundly differ- wasgest so the disturbing idea of evolutionto his contemporaries. as an alternative to selectionsubject in was a way in the that air. was Darwin significantly approached different the populationsdivine benevolence to their orlocal of aenvironment rationally structured was the divergent process, with some branches splitting relations and evolution made good sense to any- common descent was one of Darwin’s greatest ing,Darwin and I support was certainly thisThe claim image not by theof addressing the first tree to of sug- life the hadlectionscientific appeared was work in in and the othersair.These Darwin from rigidly currents approached structured circulat- the modelshardent because of to see taxo- natural he arguedcrucial selection that role the as played adaptationthe agent by his of ofeither move pop- towarding, a and model I support this claim by addressing the scientific work and others from currents circulat- ent because he argued that the adaptation of pop- theDarwin creation was of species certainly by God.not the J. B. first Lamarck to ’sug-s subjectfrom any in ofa way the otherthat was efforts significantly being made different to ex- solecosmic cause teleology. of transmutation. Selection adapted Many people species found to an over and over again, whereas others came to a one embedded in a vision of nature as a predict- achievements, in addition to natural selection it- gestrelated the issue idea of of definingDarwin evolution’ justs notebooks as why an the alternative theory of the was late to 1830ssubjecting in (Fig. his in cultural a 2) waynomic that environment. was relations significantly Few and wouldevolution different now madedivineulations good benevolence sense to theirof local branching or environment of a rationally evolution was structured the based sole on geographicalrelated issue of defining just why the theory was ing in his cultural environment. Few would now ulations to their local environment was the sole gesttheory, the published idea of evolution in 1809, hadas an been alternative widely dis- to fromplain any the historyof the other of life efforts on earth. being He hadmade a uniqueto ex- itever-changing hard to see environment,natural selection and it as did the so byagent killing of dead end through extinction. able, orderly system governed by a divine plan. self (14). So it was, but I think Mayr overestimat- theso disturbing creation of to species hisand contemporaries. was by proposed God. J. B. independently Lamarck’s byfromaccept Wallace any the of in claim theto other that anyone efforts evolution embedded being by made natural in a to vision ex- se- ofcosmiccause nature of teleology. as transmutation. a diversity. Selection Many adapted people species found to an it so disturbing to his contemporaries. accept the claim that evolution by natural se- cause of transmutation. Many people found it cussed, although generally rejected (2–4). Robert combination of scientific interests that alerted off useless variations in a ruthless “struggle for theory,the Darwincreation published wasof speciesa certainly in paper 1809, by published God. had not beenthe J. B. in first widely Lamarck’s 1855. to sug- dis-Both realizedplainlectionplain thethe that was historyhistory it in the ofpredictable,of lifelife air. onon Darwin earth.earth. orderly HeHe approached hadhad system aa unique governed the ever-changinghardeither by toa divine divinesee natural environment, benevolence selectionThis model and asor the it of did was agent a so so rationally by of radical killing either that manyDarwin late was certainlyThe image not of the the first tree to of sug- life hadlection appeared was in theSuch air.a world Darwin view approached made it the difficulthard to to accept see natural ed selectionthe rapidity as thewith agent which of eitherother naturalists were Chambers’s Vestiges of the Natural History of him to topics ignored by other naturalists. He existence.” This did not seem the kind of process cussed,gesttheory, the althoughpublished idea ofexplained generally evolutionin 1809, rejected whyhad as anbeen naturalists alternative(2 –widely4). Robert were dis to- ablecombinationsubjectcombination to arrange in a way of of plan. scientific that Suchwas significantlyinterests interestsa world thatview that different alerted made alerted it difficultoffdivinestructured useless benevolencetoac- cosmic variations19th-century teleology. or in of a a ruthless rationally evolutionistsSelection“struggle structured adapted were for unablegest to theaccept idea ofin evolution Darwin’s asnotebooks an alternativeof the late to 1830ssubject (Fig. in 2 a) waythat that the was history significantly of life on different earth mightdivine be essen benevolence- converted or of ato rationally the theory. structured Many of the non-Dar- Creation of 1844 sparked a debate over the pos-’ certainly drew on ideas widely discussed at the that would be instituted by a benevolent God, Chambersthecussed, creation although’s ofVestiges speciesspeciesgenerally of by into the God. rejected Naturalgroups J. B. within(2 History Lamarck–4). groups,Rob ofs- usinghimfromhim to to descentany topics of theignored cept other that efforts by the other being history naturalists. made of life to Heex- He on earthexistence.cosmicspecies might teleology.to” an beThis ever-changing didit Selection in not full. seem Ernst adapted environment, the Mayrkind species of argued process and to an thatit thethe theory creation of of speciesand was by proposed God. J. B.independently Lamarck’s byfromWallace any ofin thetially other irregular efforts being and unpredictable, made to ex- dependantcosmic teleology. on winian Selection theories adaptedof evolution species to proposed an during the sibility that new species were produced from time, but was forced by his scientific interests to especially because its essentially “selfish” nature Creationtheory,ert Chambers’s publishedof 1844 Vestigesfrom sparked in 1809, a commonof a had debatethe beenNatural ancestor over widely theHistory to pos- dis- explain certainlyplaincertainly the the under- history drewdrew ononessentially of ideasideas life on widelywidely earth. irregular discussedHe had and a unpredictable, unique atat the thatever-changingdid dependantso would by killing be environment, institutedcommonoff useless by descent and variations a itbenevolent did was so in by one a killing God,ruth of Darwin- theory,’s greatest publisheda paper in 1809, published had been in widely1855. Both dis- realizedplain the that history it the of lifehazards on earth. of migration, He had a uniqueisolation, andever-changing local ad- environment,“eclipse of andDarwinism” it did so by in killing the late 19th century – use those sources of inspiration in a highly orig- meant that a parasitic way of life was“ a perfectly sibilitycussed,of Creation althoughthat newof 1844lying generally species sparked similarities. were rejected a produceddebate Closely (2 4 ).over Robert related fromthe speciestime,combinationtime, have butbut waswas di- offorced forcedon scientific the by by hazards his his interests scientific scientific of migration, that interests interests alerted isolation, to especiallyoffless uselessand“struggle local because variations forachievements, existence.” its essentially in a ruthlessThis in“selfish additiondidstruggle not” nature seem to for natural cussed, selection althoughexplained generally why rejected naturalists (2–4). Robert were ablecombination to arrange ofaptation. scientific Darwin interests was that led alertedtoward hisoff alternative useless variations were introduced in a ruthless with“struggle the aim for of subverting the ’ inal way. natural adaptive” response in some circumstances. ChamberspossibilitySchool of Philosophy sthatVestiges newverged and Anthropologicalspecies of recently the were Natural from Studies, produced a History common The Queen's from of ancestor, usehimto use those towhereas those topics sources sources ignoredadaptation. of inspiration of by inspiration other Darwin in naturalists. a highly wasin a ledhighly orig- Hetoward meantexistence.the his kind alterna- that of a Thisprocess parasitic diditself notthat way (14 seem would). of So life the itbe waskind was, instituted a of perfectly but process Iby think MayrChambers over-’s Vestigesspecies of into thegroups Natural within History groups, of usinghim descent to topics ignoredmodel in by part otherbecause naturalists. he was Hemore existence.interested” inThis implications did not seem of the the kind principle of process of common descent To some extent, Darwin may have been More seriously for the idea of cosmic tele- Creationpre-existingUniversity ofof Belfast, 1844onesthe University sparked in ancestry a Road progressive a debate ofBelfast, more overBelfast, distantly sequence the Northern pos- related inalcertainlyoriginal forms way. mustway. drew ontive ideas model widely in part discussed because at hethe wasnaturalthata morebenevolent would adaptive inter- be God, instituted responseestimated especially in by the some a because rapidity benevolent circumstances. its with essen God, which- otherCreation natural-of 1844from sparked a common a debate ancestor over the to pos-explaincertainly the under drew- onadaptation ideas widely than discussedcosmic teleology, at the thanksthat would to the be instituted(15). The byAmerican a benevolent neo-Lamarckians God, Edward School of Philosophy and Anthropological Studies, The Queen's “ ” ’ CREDIT (LEFT TO RIGHT):sibilityleadingIreland, STAPLETON COLLECTION/CORBIS; HULTON-DEUTSCH COLLECTION/CORBIS; MICHAEL NICHOLSON/CORBIS BT7 up that 1NN, to newhumans UK. speciesE-mail: (5). [email protected] But were if the produced general fromidea time,merelyTo but someahead was extent,forced of his by Darwin time, his scientificanticipating may have interests devel- been to especiallytiallyology, “selfish” Darwin because natures supposition its essentiallymeant that that “a theselfish parasitic production” nature way lying similarities. Closely related species have influence of William Paley’s natural theology. Drinker Cope and“ Alpheus” Hyatt proposed that University of Belfast, Universitybe traced Road further Belfast, back Belfast, down Northern the familyTo tree some to extent,ested inDarwin adaptation may than have cosmic been teleology,More thanks seriouslyists for were the converted idea of cosmic to the tele- theory. Manysibility of that the new species were produced from time, but was forced by his scientific interests to especially because its essentially selfish nature find the common point of origin. usemerely those ““ahead sources ofto ofhis the inspiration time,” influence” anticipating in of a Williamhighly devel orig- Paley- ’smeantof natural life thatwas the- a parasiticperfectly’ non-Darwinian way natural of life adaptive theories was a response perfectly of evolution proposed diverged recently from a common ancestor, Natural selection replaced divine benevolence the evolution of each group should be seen as a CREDIT (LEFT TO RIGHT):Ireland, STAPLETON COLLECTION/CORBIS; HULTON-DEUTSCH COLLECTION/CORBIS; MICHAEL NICHOLSON/CORBIS BT7 1NN, UK. E-mail: [email protected] merely ahead of his time, anticipating devel- ology, Darwin s supposition that the production use those sources of inspiration in a highly orig- meant that a parasitic way of life was a perfectly The idea of common descent nowinalopments seems way. that so wouldology. push Natural other selection naturalists replaced to- divinenaturalin some benev- adaptive circumstances. responseduring the in“ someeclipse circumstances. of Darwinism” in the late 19th whereas the ancestry of more distantlyinal way. related as an explanation of adaptation. Unlikenatural Macleay adaptive series response of parallel in some lines circumstances. moved through the same School of Philosophy and Anthropological Studies, Thewww.sciencemag.org Queen's SCIENCE VOL 323 9 JANUARY 2009 223School of Philosophy and Anthropological Studies, The Queen's University of Belfast, Universityobvious Road that Belfast, we might Belfast, wonder Northern whatward alternativeTo an some evolutionary extent,olence vision Darwin as an during explanation may the have years of been he adaptation. MoreMore Unlike seriously seriouslycentury for the were idea introduced of cosmic with tele tele- the- aim of subvert- forms must be traced further back downTo some the extent,and Chambers, Darwin mayDarwin have did beennot expect hisMore theory seriously hierarchy for the of idea developmental of cosmic tele- stages, an updated “ ” ’ University of Belfast, University Road Belfast, Belfast, Northern CREDIT (LEFT TO RIGHT): STAPLETON COLLECTION/CORBIS; HULTON-DEUTSCH COLLECTION/CORBIS; MICHAEL NICHOLSON/CORBIS SCIENCE 223

Ireland, BT7 1NN, UK. E-mail: [email protected] merely ahead of his time, anticipating devel- ology, Darwin s supposition that the production CREDIT: SYNDICS OF CAMBRIDGE UNIVERSITY LIBRARY “ ” ’ models could have beenwww.sciencemag.org proposed toworked account in forisolation.Macleay By the andVOL late Chambers, 1850s, 323 the 9 DarwinJANUARY idea didology, not 2009 expectDarwin’s ingsupposition the implicationsthat the ofproduction the principle CREDIT (LEFT TO RIGHT): ofIreland, STAPLETON COLLECTION/CORBIS; HULTON-DEUTSCH COLLECTION/CORBIS; MICHAEL NICHOLSON/CORBIS common BT7 1NN, UK.family E-mail: tree [email protected] to find the common pointmerely of origin.ahead ofto predict his time, an orderlyanticipating pattern devel- of relations.ology, Darwin sversion supposition of the that idea the production suggested in Chambers’s 18 19 224 www.sciencemag.org9 JANUARYSCIENCE 2009 VOLVOL 323 323SCIENCE 9 JANUARYwww.sciencemag.org 2009 223 www.sciencemag.org SCIENCE VOL 323 9 JANUARY 2009 223 REVIEW

Vestiges. The similarities linking the species in directed variation among individual organisms. tion thinking, and although he may have ex- descent (15). TheThe American Struggle for neo-Lamarckians Existence scale and thatfor could that be transition investigated in the directly. late 19thinterbreeding century. It individuals.the Men Traditionally, Who Discovered species It (Doubleday, New a genus were due not to a recent common ances- Although convinced that the degree of variabil- aggerated the extent to which Darwin himself Edward DrinkerOne Cope of andthe most Alpheus disturbing Hyatt aspects pro- Thereof Darwin’s was a well-developeddid, however, network highlight of breeders the harsherwere aspects treated of as idealizedYork, types1958). with a fixed es- try, but to parallel trends independently reaching ity was artificially enhanced under domestica- made the conceptual transition, the subsequent posed that the evolutiontheory was of each its appeal group should to the be struggleby for this exis time,- andthe although consequences their of ideas struggle. about Thesence, potential any variation 2. P. from Corsi, the The norm Age being of Lamarck: trivial Evolutionary seen as a series of parallel lines moved through heredity and variation were distinctly pregenet- and impermanent. TheTheories breeders in France, knew that1790–1830 they (Univ. of the same stage of development. Like Chambers, tion, Darwin, nevertheless, accepted that there development of the selection theory brought this tence as the natural process that equates with the implications’ were drawn out even more clearly California Press, Berkeley, 1988). they endorsed the recapitulation theory (ontog- must be some equivalent variability in every implication out more clearly. the same hierarchybreeder’s of developmental activity as stages, a selecting an up- agent.ical This (like very Darwin whens own), Galton they argued had a that very it clearwould becould necessary produce huge changes in structure by ac- dated version of the idea suggested in Chambers’s appreciation of how they produced changes in cumulating normal3. variationsA. Desmond, over The a Politics number of of Evolution: Mor- eny recapitulates phylogeny, in the terminology wild population. The analogy with artificial se- In the debates that followed the publication harsh vision of nature certainly threatened the to apply artificial selection to the human race phology, Medicine and Reform in Radical Vestiges. The similarities linking the species in a their artificially small populations. The insight generations. When Darwin linked this informa- introduced by Ernst Haeckel) and saw evolution lection then allowed him to depict natural selec- of On the Origin of Species, the analogy with traditional belief in a benevolent Creator. The in order to prevent “unfit” individuals from London (Univ. of Chicago Press, Chicago, genus were due not to a recent common ancestry, that they worked by selection may have been tion with his conviction that species could change as the addition of preordained stages to ontog- tion as a parallel process in which a few variant artificial selection continued to play a keyREVIEW role term “struggle for existence” occurs in Thomas reproducing and undermining the biological 1989). but to parallel trends independently reaching the important (this is the point of contention among indefinitely over4. time, P. J. he Bowler, was driven Evolution: toward The aHistory of an eny. Adaptation was not crucial once the basic individuals, in this case with characters useful to by forcing even Darwin’s critics to think about Robert Malthus’s An Essay on the Principle health of ’the population. This was the eugenics Idea (Univ. of California Press, Berkeley, ed. descent (15). The American neo-Lamarckians scale and that could be investigated directly. interbreeding individuals. Traditionally, species same stage of development. Like Chambers, they experts studying Darwin s notebooks), but the new form of species concept in which the pop- character of the groupEdward was Drinker established, Cope and and Alpheus the Hyatt species pro- ratherThere than wasa the well-developed human breeder, network sur of- breedersthe problemswere treated of heredity as idealized and types variation with a in fixed a new es- of Population, although used in the context of program, and in its most extreme manifestation 3, 2003). endorsed the recapitulation theory (ontogeny breeders certainly taught him one thing. He ulation becomes paramount. The natural range of the linear, orthogeneticposed evolution that the evolutionof the of group each groupvive shouldand bereproduce.by this Those time, andwith although harmful their charac ideas- aboutway (1sence,8). Opponents any variation such from as the Fleeming norm being Jenkin, trivial recapitulates phylogeny,tribal groups in thecompeting terminology for intro- limitedrealized resources. that inat a domesticatedthe hands of populationthe Nazis, thereit led notvariability just to the becomes 5. partJ. A. of Secord, the species Victorian’ character, Sensation: The Ex- traordinary Publication, Reception, and Secret might eventually generateseen as abizarre series of nonadaptive parallel lines movedters throughare eliminatedheredityby and the variationstruggle were for existence, distinctly pregenet- who sawand impermanent.selection working The breederson large knew variations that they duced by ErnstDarwin Haeckel) saw and that saw population evolutionpressure as is would always lead a fundsterilization of apparently but purposelessalso to the actual and eliminationnot the result of of accidental deviations from a the same hierarchy of developmental stages, an up- ical (like Darwin’ s own), they had a very clear could produce huge changes in structure by ac- Authorship of Vestiges of the Natural History characters as a prelude to extinction—the the- just as the breeder will not permit any animal or “sports of nature,” were, nevertheless, still the addition of preordainedto competition stages between to ontogeny. individuals undirectedand was per variation- those among unfortunates individual deemed organisms. unfit by fixedthe state. norm.Did This is what Mayr called the transi- dated version of the idea suggested in Chambers’s appreciation of how they produced changes in cumulating normal variations over a number of of Creation (Univ. of Chicago Press, Chicago, ory of “racial senility.” Darwin could make no to reproduce if it does not have the character he working within the framework defined by this Adaptation washaps not crucialthe first once to therealize basic that char- it mightAlthough represent convinced Darwin’s that the emphasis degree onof variabil- the natural tion elimination from typological thinking to population think- Vestiges. The similarities linking the species in a their artificially small populations. The insight generations. When Darwin linked this informa- 2000). sense of the theory proposedgenus were by due Cope not toand a recentHyatt, common wants. ancestry, It wasthat the theybreeders worked who by selectiontaught Darwin may have beenanalogy.tionFor with supporters his conviction such that speciesas Francis could Galton, change acter of the groupa wasmeans established,by which andthe thepopulation linear, couldity was change artificially of enhanced maladaptive undervariants domestication, help toing, create and although a 6. heD. may Kohn, have Ed., exaggerated The Darwinian the Heritage: A

because he could notbut imagine to parallel an trends evolutionary independently that reaching variation the importantis not directed (this is toward the point some of contention preor- amongartificialindefinitely selection overhelped time, he to was clarify driven the toward nature a orthogenetic evolutionthrough of time the group(19, 2 might0). The even- processDarwin, worked nevertheless, by climate accepted of opinion that there in must which be suchextent atrocities to which DarwinCentennial himself Retrospect made the (Princeton con- Univ. Press,

process driven by predeterminedsame stage of development.trends. But Likethe Chambers,dained theygoal, allowingexperts studying him to Darwinbuild ’son notebooks), his exist- butof the bothnew heredity form of and species selection, concept inpaving which the pop-way tually generateeliminating bizarre nonadaptive the least fit characters variants withinsomethe equivalentpopu- became variability possible? in every wild popu- ceptual transition, thePrinceton, subsequent NJ, development 1985). of 7. J. Browne, Charles Darwin: Voyaging (Jona- fact that such theoriesendorsedflourished the recapitulationin the late 19th theory ing (ontogeny convictionbreeders that adaptive certainly evolution taught him must onebe thing. for He theulation revolutionary becomes paramount. impact Theof Mendelian natural range ge of- as a prelude to extinctionlation and— allowingthe theory the ofbetter“racial adaptedlation. to survive The analogy It with has to artificial be admitted selection that, then by makingthe selectiondeath it- theory brought this implication out recapitulates phylogeny, in the terminology intro- realized that in a domesticated population there variability becomes part of the species’ character, than Cape, London, 1985). century demonstrates just how radical the theory an open-ended, branching process. netics. The notion of “hard” heredity was in- senility.” Darwinand could breed. make This nowas sense what ofthe the philosopherallowed Her him- to depictself a creative natural force selection in nature, as a par- Darwinmore introduced clearly. duced by Ernst Haeckel) and saw evolution as is always a fund of apparently purposeless and not the result of accidental deviations from a 8. J. Browne, Charles Darwin: The Power of of open-ended, divergentthe addition evolution of preordained was to the stages to ontogeny.At the sameundirected time, variationthe breeders’ among individualattitude to organisms.- troducedfixed in norm. opposition This is what to Mayrthe “soft” called the form transi- of theory proposedbert by CopeSpencer and would Hyatt, later becauserefer heto as theallel “survival process in whicha new and a few profoundly variant individuals, disturbing insightIn into the the debates thatPlace followed (Jonathan the Cape, publication London, 2002).

naturalists of the time.Adaptation was not crucial once theward basic variation char- Although pushed convincedDarwin toward that the degreethe view of variabil- inheritancetion from implied typological by the thinking Lamarckian to population process. think- could not imagineof anthe evolutionary fittest.” Strictly process speaking, drivennaturalin this selection case with world, characters an insight useful that to the seems species to haveof resonatedOn the Origin9. ofM. Species J. S. Hodge,, the analogy G. Radick, with Eds., The Cam- acter of the group was established, andthat the the linear, speciesity was is artificially just a population enhanced under of domestication,inter- The undirecteding, and although nature he of may variation have exaggerated was clarified the by predeterminedrequires trends. only But the differential fact that reproduction such rather among than the humanwith the breeder, thinking survive of many and who re- didartificial not under selection- continuedbridge Companion to play ato key Darwin role (Cambridge Artificial Selection orthogenetic evolution of the groupbreeding might even- individuals.Darwin, nevertheless, Traditionally, accepted that species there mustboth be throughextent to the which study Darwin of large himself populations made the con- by theories flourishedvariants, in the but late Darwin 19th century thought dem- that theproduce. pressure Those of withstand harmful or accept characters the details are of elim- his theory.by forcingDarwin even- DarwinUniv.’s Press, critics Cambridge, to think about 2003). tually generate bizarre nonadaptive characters some equivalent variability in every wild popu- ceptual transition, the subsequent development of 10. M. J. S. Hodge, Stud. Hist. Biol. 6, 1 (1982). These non-Darwinian models were ultimately— were “treated as idealized types with a fixed es- Galton and through the breeding studies of the onstrates just howcompetition radical thewas theory necessary of open- to makeinated it effective. by the struggleism was for not existence, “responsible” just as for the socialtheDarwinism problems of heredity and variation in a new as a prelude to extinction the theory of racial lation. The analogy with artificial selection then the selection theory brought this implication out 11. P. F. Rehbock, The Philosophical Naturalists: marginalized by thesenility.synthesis” Darwin of the could selection make no sensesence, of any the variationallowed himfrom tothe depict norm natural being selection trivial as a par-geneticists.more clearly. Although it took some time for the ended, divergentIt evolution seems that was without to the the naturalistsinput frombreeder Malthus, will he not permitor eugenics any animal in any to simple reproduce way. if Afterway all, (18 some). Opponents Themes such asin Early Fleeming Nineteenth-Century Jenkin, British

theory and genetics theory in the proposed early 20th by Cope century. and Hyatt, and because impermanent. he allel process The breeders in which aknew few variant that they individuals, geneticistsIn theto accept debates the that situation, followed the their publication studies of the time. would not have come up with the theory.it does not have theearly charactergeneticists he wants. endorsed It was eugenics the who by analogy saw selection workingBiology (Univ. on large of Wisconsin variations Press, Madison, Genetic mutations seemedcould not to imaginebe essentially an evolutionary plu- processcould driven producein huge this case changes with characters in structure useful by to ac the- speciesof mutationof On theultimately Origin of endorsed Species, theDarwin’s analogy claim with The idea of struggle was pervasivebreeders in the wholit- taughtwith Darwin animal that breeding variation even is not whileor dismissing“sports of nature,1983).” were, nevertheless, still ralistic and undirected,by providing predetermined just trends. the source But the factcumulating that such normalrather thanvariations the human over breeder, a number survive of andthat re- theartificial only way selection the continuedenvironment to play could a key affect role Artificial Selectionerature of the period, but could be directedexploited toward in somenatural preordained selection as goal, the mechanism allowing ofworking evolution. within 12. the R. framework J. Richards, defined The Meaning by this of Evolution: The ’ of “random” variationtheories that Darwin’s flourished mechanism in the late 19th centurygenerations. dem- produce. When DarwinThose with linked harmful this characters infor are- elim-the populationby forcing was even by Darwin selection.s critics Modern to think evolu about- These non-Darwinianmany different models ways. were ultimatelyIn the 1850s, himSpencer to build had on hisAnd existing the Nazis conviction wanted that to adapt- purify a analogy. fixed racial For supportersMorphological such as Francis Construction Galton, and Ideological onstrates just how radical the theory of open- inated by the struggle for existence, just as the the problems of heredity and variation in a new Reconstruction of Darwin’s Theory (Univ. of marginalized by the synthesis of the selection ive evolution must be an open-ended, branching artificial selection helped to clarify the nature required as its raw ended, material. divergent This evolution later devel was- to themation naturalists with breeder his conviction will not permit that any species animal tocould reproducetionary if waydevelopmental (18). Opponents biology such as has Fleeming reopened Jenkin, the already seen how competition could be turned type, which they certainly did not want to admit Chicago Press, Chicago, 1992).

opment highlights theof the importance time. of another change indefinitelyit does not haveover the time, character he he wants.was Itdriven was question the who of saw whether selection variation workingand on largeevolution variations can theory and geneticsinto a in very the different, early 20th and century. in some respectsprocess. less had evolved gradually from an ape ancestry.of both heredityBut 13. and A. selection, Desmond, paving J. R. Moore, the way Darwin’s Sacred insight gained by Darwin in the late 1830s, his toward a newbreeders form of who species taught concept Darwin that in variationwhich isbe not quiteor “assports open-ended of nature, ”aswere, Darwin nevertheless, and his stillfol- Genetic mutationsdisturbing, seemed tomechanism be essentially of plu-progress (At21). the For same by time, proposing the breeders that evolution’ attitude workedfor primarily the revolutionary Cause: impact Race, ofSlavery Mendelian and the Quest for Hu- Artificial Selection decision to investigate the work of the animal the populationdirected becomes toward paramount. some preordained The natural goal, allowing lowersworking believed. within But the the framework non-Darwinian defined byvision this ralistic and undirected,Spencer, providing the interaction just the source betweentoward individuals variation through pushed the Darwinelimination toward of useless the variants,genetics. Dar The- notionman of Origins“hard” (Allenheredity Lane, was London, 2009). breeders (Fig. 3) andThese his recognition non-Darwinian that models their wererange ultimately of variabilityhim to buildbecomes on his part existing of the conviction species’ that adapt-of evolutionanalogy. unfolding For supporters to an such orderly, as Francis predictable Galton, of “random” variationstimulated that their Darwin efforts’s mecha-to adapt toview the chang that the- specieswin created is just an image a population that could of all toointroduced easily be in opposition14. E. Mayr, to One the “Longsoft” Argument:form of Charles Dar- artificial selection helped to clarify the nature marginalized by the synthesis of the selection ive evolution must be an open-ended, branching nism required as its raw material. win and the Genesis of Evolutionary Thought method of artificialtheory selection and genetics offered in thea earlyuseful 20thcharacter, century. notprocess. the result of accidental deviations plan hasof bothbeen heredity essentially and selection,marginalized pavingby the accep way- ing social and physical environment. He then exploited by those who wanted the human race (Harvard Univ. Press, Cambridge, MA, 1991). This later development high- way ofunderstandingGenetic how the mutations equivalent seemed natural to be essentiallyfrom a plu-fixed norm.At This the sameis what time, Mayr the called breeders the’ attitude tance offor the the key revolutionary insights on impactwhich Darwin of Mendelian based invoked Lamarck’s concept of the “inheritance to conform to their own pre-existing ideals. In 15. P. J. Bowler, The Eclipse of Darwinism: Anti- process operated. Theralistic exact and role undirected, played by providing Dar- justtransition the source fromtoward typological variation thinking pushed Darwinto popula toward- his the theorygenetics. of natural The notion selection. of “hard ” heredity was lights the importanceof acquired of another characteristics” to explain how these the same way, his popularization of the struggle Darwinian Evolution Theories in the Decades win’s study of breedingof “inrandom the formulation” variation thatof his Darwin ’s mecha- view that the species is just a population of introduced in opposition to the “soft” form of insight gained byself-improvements Darwin in the accumulated over many gen- metaphor focused attention onto the individual- Around 1900 (Johns Hopkins Univ. Press, theory is much debatednism by required historians as its raw (16 material.–17), late 1830s, his decisionerations, to leading inves- to biological evolution and so- istic aspects of Spencer’s philosophy. Baltimore, MD, 1983). This later development high- tigate the work of the animal 16. R. J. Richards, Stud. Hist. Philos. Sci. 28, 75 but there can be littlelights doubt the importance of how important of another cial progress. Spencer’s self-improvement mod- Modern science recognizes the importance (1997). the analogy between artificial and natural selec- breeders (Fig. 3)el andof progress his recog- became immensely popular in the of Darwin’s key insights when used as a way insight gained by Darwin in the nition that their method of artifi- 17. M. Ruse, J. Hist. Ideas 36, 339 (1975). tion became in his laterlate 1830s, thinking. his decision In this to case, inves- later 19th century, and because it too seemed to of explaining countless otherwise mysterious 18. J. Gayon, Darwinism’s Struggle for Survival: cial selection offered a useful way Darwin was truly unique,tigate because the work even of theWallace animal rely on struggle as the motor of change, it was aspects of the natural world. But some of those Heredity and the Hypothesis of Natural Se- of understanding how the equiv- did not take this stepbreeders and dissociated (Fig. 3) and his himself recog- often confused with the Darwinian mechanism. insights came from sources with profoundly dis- lection (Cambridge Univ. Press, New York, nition that their method of artifi- alent natural process operated. 1998). from the link with artificial selection expressed In fact, Spencer thought that all humans will turbing implications, and many historians now cial selection offered a useful way The exact role played by Darwin’s 19. P. J. Bowler, J. Hist. Ideas 37, 631 (1976). in Darwin’s later writings. eventually acquire the faculties needed to inter- recognize that the theory, in turn, played into of understanding how the equiv- study of breeding in the formula- 20. A. Desmond, J. R. Moore, Darwin (Michael alent natural process operated. Darwin turned to the breeders in search of a tion of his theoryact is muchharmoniously debated with one another. But his occa- the way those implications were developed by Joseph, London, 1991). The exact role played by Darwin’s 21. M. Francis, Herbert Spencer and the Invention clue as to how a population could be changed— by historians (16sional–17), use but of there highly individualistic language al- later generations. This is not a simple matter of study of breeding in the formula- of Modern Life (Acumen, Stocksfield, UK, here at least was a situation where modifications can be little doubtlowed of how him impor- to be perceived as the apostle of free science being “misused” by social commenta- tion of his theory is much debated 2007). were actually being producedby historians on ( 16a –human17), buttime there tant the analogyenterprise. between artificial Much of what later became known tors, because Darwin’s theorizing would almost 22. C. Darwin, Transmutation Notebook B, from scale and that couldcan be be investigated little doubt of how directly. impor- and natural selectionas “social became Darwinism” in his was, in fact, Spencerian certainly have been different had he not drawn Natural Selection portfolio p. 36 (Cambridge

There was a well-developedtant the analogy network between of breed artificial- later thinking. Insocial this case, Lamarckism Darwin expressed in theterminology inspiration from social, as well as scientific, in- Univ. Library, Cambridge, 1838); P. H. Bar- and natural selection became in his ers by this time, and although their ideas about was truly unique,of struggle because popularized even by Darwin. fluences. We may well feel uncomfortable with rett, P. J. Gautrey, S. Herbert, D. Kohn, S. later thinking. In this case, Darwin Smith, transcribers and Eds., Charles Dar- heredity and variation were distinctly pregeneti- Wallace did not takeThis this steppoint and is important in the context of those aspects of his theory today, especially in was truly unique, because even win’s Notebooks p. 180 (British Museum of cal (like Darwin’s own),Wallace they did had not take a very thisstep clear and dissociated himselfthe from charge the linkraised by modern opponents of light of their subsequent applications to human Natural History, Cornell Univ. Press, Ithaca, appreciation of howdissociatedthey produced himself changes from the link in with artificial selectionDarwinism expressed that the theory is responsible for the affairs. But if we accept science’s power to up- NY, 1987); available at the Darwin Digital their artificially smallwith artificial populations. selection The expressed insight in Darwin’s laterappearance writings. of a whole range of unpleasant social set the traditional foundations of how we think Library, http://darwinlibrary.amnh.org/. ’ that they worked byin selection Darwin s later may writings. have been Darwin turnedpolicies to the based breeders on struggle. Darwin exploited the about the world, we should also accept its po- 23. G. Neumeister, Das Ganze der Taubenzucht important (this is the pointDarwin of contention turned to the among breeders idea of the struggle for existence in a way that tential to interact with moral values. (B. F. Voigt, Weimar, 1876). in search of a clue as to how a in search of a clue as to how a — experts studying Darwin’spopulationnotebooks), could be changed but the— population couldwas be unique changed until paralleled by Wallace nearly breeders certainly taughthere athim least one was thing. a situation He real where- here at least was20 a situation years later. where Their theory certainly fed into ized that in a domesticatedmodifications population were actually there be-is modifications werethe movements actually be- that led toward various kinds of References and Notes 23 CREDIT: AMERICAN MUSEUM OFing NATURAL HISTORY produced on a human time Fig. 3. Pigeons ( ). 23 1. L. Eiseley, Darwin’s Century: Evolution and always a fund of apparently purposeless and un- CREDIT: AMERICAN MUSEUM OFing NATURAL HISTORY produced onsocial a human Darwinism, time butFig. it was 3. Pigeonsnot the only ( ).vehicle

20 www.sciencemag.org SCIENCE VOL 323 9 JANUARY 2009 225 21 www.sciencemag.org SCIENCE VOL 323 9 JANUARY 2009 225 Speciation

disentangle key aspects of clade histories. Clades REVIEW are monophyletic, including all descendants of an SPECIALSECTION ancestor, whereas taxa may be monophyletic or Speciation vestigations,paraphyletic, excluding provide an some outline descendantsof how of these the new taxa couldmarine become ecosystem established became saturated.only by onplants, diversity, insects,SPECIAL orand vertebrates, other correction becauseSECTION regimes ’ 5 The Red Queen and the Court Jester: andancestor. other Comparativestudies correspond macroecological to the predictions studies driving othersSepkoski to extinction.s coupled logistic Key evidence model ( ) maythese be groups so complex seem to haveas to radiated produce data in 600

disentangle key aspects of clade histories. Clades 3000 of the Red Queen, Court Jester, and multilevel marineis that both ecosystem identifiedorigination became three and equilibria, saturated.extinction in the Cam-rates whichexplosively,plants, geologic without insects, anddiversity or biologic vertebrates, plateaus, signals because are REVIEW addare monophyletic, rigor to analyses including showing all descendants that sister clades of an mixed models (Table 1), and outline some phy- Sepkoskiappear to’ shave coupledbrian, been most logistic density-dependent of the Paleozoic, model (5 and) per-(5– notparticularly obviouslythese in groups theseparated. past seem 100 to million have radiated Species Diversity and the Role of Biotic mayancestor, vary whereasin rate of taxa evolution, may be timing monophyletic of increases or haps a third, beginning in the Pliocene years (My) (10). 600 logenetic studies of the macroevolution of spe- identified7), limiting three rises equilibria, in diversity in the and Cam- promoting 3000 Lifeexplosively, on land today without may diversity be as much plateaus, as in species richness and morphospace occupation, 400 paraphyletic, excluding some descendants of the and continuing to the present (Fig. 2000 Resolution between the equilibri- brian, most of the Paleozoic, and per- particularly in the past 100 million The Red Queen and the Court Jester: ciesandancestor. distributionsdiversity. Comparative of evolutionary macroecological novelties studies across rapid recovery2A). after These extinction three equilibrium events. levels 25um times models, as anddiverse between as life these in andthe sea, so it and Abiotic Factors Through Time lineages and subclades. Here, I will explore the hapsAlternative a third,correspond beginning models to in three for the global setsPliocene of phyla, diversi the- mayexpansionist beyears wrong (My) models, to (10 generalize). might seem from marine

add rigor to analyses showing that sister clades (empirical; red) (corrected; blue) 400 200 Cambrian, Paleozoic, and Modern, that Number of genera straightforward, but the solution de- 2000 1000

and continuing to the present (Fig. Number of genera Resolution between the equilibri- Thelargest-scalemay Global vary in rate global,Pattern of evolution, taxic of Diversification investigations, timing of increases provide fication are expansionist, allowing global paleontological studies to all life. Perhaps MichaelSpecies J. Benton Diversity and the Role of Biotic Through Time 2A).species These diversityinteracted three to equilibrium rise, and successivelywith damping, levels replaced but landpends andum onadequate sea models, show assessment and similar between ofpatterns the these of andex- anin speciesoutline ofrichness how these and morphospace and other studies occupation, corre- each other through the Cambrian- quality of the fossil record. The long- Aspond key to question the predictions about of the the origin Red Queen, of modern Court correspondwithout a predictable to three sets limit of phyla, (8–1 the1). Den- ponentialexpansionist increase in models, species mightnumbers seem (8, 0 and distributions of evolutionary novelties across Ordovician and Permo-Triassic inter- 0 term saturation model for global diver- (empirical; red) Cm O S D C P Tr J K Pg Ng (corrected; blue) 200 Evolution may be dominated by biotic factors, as in the Red Queen model, or abiotic factors, Cambrian, Paleozoic, and Modern, that Number of genera straightforward, but the solution de- 1000 and Abiotic Factors Through Time biodiversityJester, and multilevel is how mixedtoday’s models 10 million (Table species 1), and sity dependencevals, reachingof origination higher equilibrium and extinc levels- Number of genera 9sification, 11), or (Fig. perhaps 2, blue they curve) differ arises in their key lineages and subclades. Here, I will explore the 500 400 300 200 100 0 as in the Court Jester model, or a mixture of both. The two models appear to operate interacted and successively replaced pends on adequate assessment of the aroseoutlinelargest-scale from some a single global, phylogenetic ultimate taxic investigations, studies species ofof the microbial provide macro- tion rates doesaftereach not long-term preclude replacement expansionist event. rulesfrom extensive (13, 14), attemptswith the to correctsea acting data as a giant predominantly over different geographic and temporal scales: Competition, predation, and other each otherThe through second themodel Cambrian- (6, 7) identifies a Time (Ma) sets forquality sampling of errorthe fossil (6, 7), record. where- The long- Speciation Michael J. Benton lifeevolutionan outline3500 ofmillion of species how years these diversity. ago and (Ma) other ( studiesFig. 1). corre- Two models because they may be dampened by Gaussian petri dish, where species diver- 0 biotic factors shape ecosystems locally and over short time spans, but extrinsic factors such as single equilibrium level from the early 0 as the multiple-equilibria and expan- models for global diversification are termed the Ordovicianlimiting factors and Permo-Triassic such as shortage inter- of food Fig. 2. Patterns of marine animal genus diversification through the past sity isterm equilibrial saturation and model density-dependent, for global diver- disentangle key aspects of clade histories. Clades spond to the predictions of the Red Queen, Court Paleozoic, perhaps 400 Ma, to the Fig. 2. Cm Patterns O S of marine D C animalP genusTr J diversificationK Pg Ng through sion models originally used raw data, REVIEW climateEvolution and may oceanographic be dominated and by tectonic biotic factors, events as shape in the larger-scale Red Queen patterns model, regionally or abiotic and factors, The Global Pattern of Diversification vals, reaching higher equilibrium levels 530 My, the Phanerozoic. The two lines compare current estimates from sification (Fig. 2, blue curve) arises are monophyletic, including all descendants of an saturation/equilibriumJester, and multilevel mixedmodel models (5–7) (Table and the 1), andex- or space, orpresent active (Fig. predation, 2B). In both as well models, as theby the past500 530 My, the400 Phanerozoic.300 The200 two lines 100compare current0 andwithout the correctionland witnessing (5, 8–11), continuing although (damp- globally, and through thousands and millions of years. Paleobiological studies suggest that species Through Time after each long-term replacement event. the empirical (uncorrected) Sepkoski database (red line) and sampling- from extensive attempts to correct data as in the Court Jester model, or a mixture of both. The twoancestor, models whereas appear taxa to mayoperate be monophyletic or pansion model (8–11). The equilibrium model climate andequilibria other physical correspond factors. to biodiversity Further, ened)more recentexponential analyses rise return in adiversity some- as ever outline some phylogenetic studies of the macro- 6 7 standardizedestimates from (corrected) the empirical analysisTime of the (uncorrected)(Ma) Paleobiology SepkoskiDatabase (blue database line). 6 7 diversitypredominantly is driven over largely different by abiotic geographic factors and such temporal as climate, scales:paraphyletic, landscape, Competition, excluding or food predation, some supply, descendants and and other of the A key question about the origin of modern bio- The second model ( , ) identifies a sets for sampling error ( , ), where- The Red Queen and the Court Jester: hasevolution prevailed, of species among diversity. marine paleobiologists at it seems thatsaturationthe coupled in which logistic new taxa model could The(red empirical line) and curve sampling-standardized (red line) suggests that global(corrected) marine diversityanalysis reached of the newwhat dampened sectors of but ecospace congruent signal are conquered comparativebiotic factors phylogenetic shape ecosystems approaches locally offer and newover insights short time into spans,ancestor. clade but dynamics. Comparative extrinsic macroecological factors suchas studies diversity is how today’s 10 million species arose single equilibriumbecome level established from the only early by driving whenas corrections the multiple-equilibria areimposed. Cor- and expan- add rigor to analyses showing that sister clades least, for a long time, and represents a classic may be partly an artifact of taxonomicFig. 2.aPaleobiology possiblePatterns plateau of Database marine through animal(blue the Paleozoic line). genus The (450 diversification empirical to 250 Ma) curve and through (red has risen,line) the past(9, 14). Any model for global diversifica- Paleozoic,others perhaps to extinction. 400 Ma, Key to evidence the is rectionsion for models sampling originally is clearly used es- raw data, climate and oceanographic and tectonic events shape larger-scalemay vary patterns in rate of evolution, regionally timing and of increases RedfromThe Global Queen a single Patternviewpoint ultimate of speciesbecause Diversification of it microbial implies pri life- scale (Fig. 2, red curve); it was worked out530 My,apparentlysuggests the Phanerozoic. that exponentially, global marine The ever two since.diversity lines The compare reached sampling-standardized currenta possible estimates plateau curve from tion must encompass the independent evi- Species Diversity and the Role of Biotic present (Fig.that 2B). both In origination both models, and the extinction sentialwithout (6, 7), and correction future investigation (5, 8–11), although globally,herehere and are are through two two ways ways thousands of of viewing viewing and evolution,millions evolu of- years.portantscales, Paleobiological and notin toit species is export likely richness studiesorganism-level that and evolution suggest morphospace processes thatoperates occupation, species in to marily3500Through million biotic Time years controls ago (density (Ma) (Fig. dependence) 1). Two mod- on at ordinal and familial levels but does notthe empirical(bluethrough line) the (uncorrected) suggests Paleozoic that (450 global Sepkoski to marine 250 database Ma) diversity and (redhas reached risen, line) near-modern apparently and sampling- dence for increasing complexity of organ- rates appear to have been density- levels some 400 Ma and there has been only modest increase since then. must determine appropriate indepen- diversitythrough is driven the largely spectacles by abiotic of either factors the Red such asregional climate, orand landscape, global distributions scales, or of food evolutionary and supply, it is novelties likely and across that els for global diversification are termed the equilibria correspond to biodiversity exponentially, ever since. The sampling-standardized curve (blue more recent analyses return a some- and Abiotic Factorstion, through Through the spectacles Timeof either a pluralistic way (3). globalA key diversity. question about the origin of modern bio- work convincinglydependent (at5– 7 generic), limiting or risesspecific instandardized di- Cm, Cambrian; (corrected) C, Carboniferous; analysis of D, the Devonian; Paleobiology J, Jurassic; Database K, Cretaceous; (blue line).isms,dent proxies increases for preservation in the occupation and hu- of novel comparativeQueen phylogenetic or the Court Jester. approaches The Red offer Queen new insightsevolution into operateslineages clade and dynamics. in subclades. a pluralistic Here, way I will (3 explore). the saturation/equilibrium’ model (5–7) and the ex- saturation in which new taxa could The empiricalline) suggests curve (redthat line)global suggests marine thatdiversity global reached marine near-modern diversity reached what dampened but congruent signal T the Red Queen or the Court Jester. The There arelargest-scale two broad global, methodologies taxic investigations, for providestud- diversityThere is are how two today versionss 10 million of the species equilibrium arose levels (10, versity11), thereand promoting are problems rapid recovery with Ng, Neogene; O, Ordovician; P, Permian; Pg, Paleogene; S, Silurian; Tr, ecospace,man error; someexplosive current proxiesevolution (such within par- – become established only by driving levels some 400 Ma and there has been only modest increase when corrections are imposed. Cor- Michael J. Benton modelT (1) stems from Darwin, who viewed evo- There are two broad methodologies for studies pansion model (8 11 ). The equilibrium model has after extinction events. a possibleTriassic. plateau Based on through (6). the Paleozoic (450 to 250 Ma) and has risen,as number of fossiliferous localities) Red Queen model (1) stems from Darwin, who ies of speciesan outline diversity of howthrough these and time, other studiestaxic corre-and model,from a singlediffering ultimate in the species time when of microbial the global life key numerical assumptions (11, 12), and ticular clades, and addition of novel clades others to extinction. Key evidence is since then. Cm, Cambrian; C, Carboniferous; D, Devonian; J, Ju- rection for sampling is clearly es- lutionviewed ashere evolution primarily are two as a ways balanceprimarily of of viewing a biotic balance pressures,evolution, of bi- phylogeneticofportant species notspond diversity to (export4 to). the The through predictions organism-level taxic time, ofapproach the taxic Red processes Queen, andinvolves phy- Court to marineprevailed,3500 million ecosystem among years marinebecame ago (Ma) paleobiologists saturated. (Fig. 1). Sepkoski’s Two at mod-least, the backgroundAlternative assumption models of for aglobal globalapparently di- exponentially, ever since. The sampling-standardized curvewithoutare themselves the loss dependent of precursors on diversity, (9, 11, 15), Evolution may be dominated by biotic factors, as in the Red Queen model, or abiotic factors, that both originationversification and areexpansionist, extinction allowingrassic; global K, Cretaceous;partly an artifact Ng, Neogene; of taxonomic O, Ordovician; scale (Fig. 2, P, red Permian;and otherPg, correctionsential regimes (6, may7), and be so future complex investigation mostotic pressures, notablythrough competition, most the spectacles notably and competition, of it waseither character- the and Red it treatinglogeneticregional species, or (Jester,4). global The and genera, multilevel taxic scales, approach mixed or and families models it involves is (Tablelikely as indetreat- 1), that and- coupledforels a for long global time,logistic and diversification representsmodel (5 a) classic identified are termed Red Queen three the carrying capacity is doubtful (8, 10, 1(blue1). line) suggests that global marine diversity reached near-modernall of which have happened many times in as in the Court Jester model, or a mixture of both. The’ two models appear to operate outline some phylogenetic studies of the macro- rates appearspecies to have diversity been to rise, density- with damping, butPaleogene; with- curve);S, Silurian; it was Tr, worked Triassic. out Based at ordinal on (6 and). familial as to produce datamust in which determine geologic andappropriate biologic indepen- ized byQueen the Red or the Queen CourtsJester. statement The to Red Alice Queen in ingevolution species, operates genera, in or a pluralistic families as way independent (3). viewpointsaturation/equilibrium because it implies model primarily (5–7) and biotic the con- ex- – levels some 400 Ma and there has been only modest increase since then. predominantly over differentwasT geographic characterized and temporalby the scales:Red Queen’s Competition,“ statement predation, andpendent other entitiesevolution and of species counting diversity. their occurrences equilibria, in the Cambrian, most of the dependentFurther, it has (out5– 7proved a), predictable limitinghard risesto limit export in (8 di-11 the). Density lo- depen- levels but does not work convincingly at generic signals arethe not obviouslypastdent500 proxies separated.My. for preservation and hu- Through the Looking-Glass that it takes all the entities and counting their occurrences against trols (density dependence) on global diversity. Cm, Cambrian; C, Carboniferous; D, Devonian; J, Jurassic; K, Cretaceous; biotic factors shape ecosystemstomodel Alice locally (1 in) stemsand Through over from short the Darwin, time Looking-Glass spans, who but viewed extrinsic that evo- factors “it suchagainst asThere time are and two other broad factors. methodologies The phylogenetic for studies Paleozoic,pansion model and ( 8 perhaps–11). The a equilibrium third, beginning model has in gistic modeldence to the of originationmuch more and speciose extinction terrestri rates does- notcurve)or specific arises levels from (10 extensive, 11), there are attempts problems to with cor- Life on land today may be as much as 25 climate and oceanographicrunning and tectonic you events can do, shape to larger-scale keep in the patterns same regionally place.” andtime andThe other Global factors. Pattern The of Diversification phylogenetic ap- There are two versions of the equilibrium versity and promoting rapid recovery Ng, Neogene; O, Ordovician; P, Permian; Pg, Paleogene; S, Silurian; Tr, man error; some current proxies (such takeslution all as primarily the running a balance you can of biotic do, to pressures, keep in approachof species uses diversity cladograms through or time, molecular taxic and trees phy- to theprevailed, Pliocene among and continuing marine paleobiologists to the present at ( least,Fig. al realm, whetherpreclude one expansionist considers models plants, because insects, they mayrectkey data numerical sets for assumptions sampling (11error, 12), (6 and, 7 the), whereas back- times Large-Scale as diverse as lifeControls in the sea, on soSpecies it may beDiversity globally, and through thousandsThe Court and millions Jester of years. model Paleobiological (2) is that studies evolution, suggest thatproach species usesThrough cladograms Time or molecular trees to model, differing in the time when the global after extinctionbe dampened events. by limiting factors suchTriassic. as shortage Based onground (6). assumption of a global carrying capacity is wrong to generalizeas number from marine of fossiliferous paleontolog- localities) diversity is driven largelythe bymost abiotic same notably factors place.” competition, such The as climate,Court and landscape,Jester it was model or character- food (2 supply,) is anddisentanglelogenetic (4 key). The aspects taxic of approach clade histories. involves Clades treat- 2forA). a These long time, three and equilibrium represents alevels classic correspond Red Queen or vertebrates, because these groups seem to the multiple-equilibria and expansion models Taxic paleobiological studies have provided a speciation, and extinction’ rarely happen except in A key question about the origin of modern bio- Alternativeof food models or space, for or global active predation, di- as well as by doubtful (8, 10, 11). Further, it has proved hard to ical studies to allare life. themselves Perhapsdependent land and sea on diversity, comparative phylogeneticthatized approaches evolution, by the offer Red newspeciation, Queen insightss into and statement clade extinction dynamics. to Alice rarely in areing monophyletic, species,diversity genera, is including how or today families’s 10all milliondescendants as independent species arose of toviewpoint three sets because of phyla, it implies the Cambrian, primarily Paleozoic, biotic con- have radiated explosively, without diversity pla- originally used raw data, without correction (5, great deal of evidence about controls, mainly response to unpredictable changes“ in the physical versificationclimate are expansionist, and other physical allowing factors. global Further, itpartlyexport an artifact the logistic of taxonomic model to the scale much (Fig. more 2, redshowand similar other patterns correction of exponential regimes increase may be in so complex happenThrough except the Looking-Glass in responsethat to it unpredictable takes all the anFig.entities ancestor, 1. andOperationfrom whereas counting a single oftaxa ultimate their Red may occurrencesQueen species be monophyletic of microbial against life andtrols Modern, (density dependence) that interacted on global and successively diversity. teaus, particularlyseems that in thethe coupledpast 100 logistic million model years may be8–11speciose), although terrestrial more realm, recent whether analyses one considers return aspecies abiotic, numbers on species (8, 9, 11diversity.), or perhaps Biotic they factors, such environment, recalling the capricious behavior” A species diversity to rise, with damping, but with- curve); it was worked out at ordinal and familial as to produce data in which geologic and biologic here are two ways ofchangesrunning viewing youin evolution, the can physical do,portant to keep environment, not into export the same organism-level recalling place. processesor(biotictime paraphyletic, to and causation)3500 other million factors. excluding and years Court ago The (Ma)some Jester phylogenetic (Fig. descendants 1). Two mod- ap- replacedThere each are two other versions through of the the equilibrium Cambrian- (My) (10). 8–11 somewhat dampened but congruent signal when as body size, diet, colonizing ability or ecologi- through the spectaclesofthe of either licensed the Red foolregional of Medieval or global times. scales, Neither and it is likely that els for global diversification are termed the out a predictableTable 1. limitMacroevolutionary ( ). Density phenomena depen- and theirlevels support but doesfor either not thework Red convincingly Queen (biotic, at intrinsic) generic or Courtsignals Jester are (physical, not obviously extrinsic) separated. models. Many The Court Jester model (2) is that evolution, (abioticproach causation)uses cladograms models or at differ-molecular treesGenus to model, differing in the time when the global “ ” TQueen or the Court Jester.modelthe capricious The was Red Queen proposedbehaviorevolutionas of exclusive,the operates licensed in aand pluralistic fool both of way (of3). the ancestor.saturation/equilibrium Comparative model macroecological (5–7) and thelife ex- spanOrdovician and Permo-Triassic intervals, denceResolution of originationcould between fit either and the extinction worldview, equilibrium andrates so doesmodels, are notednot asorcorrectionsmultilevel specific levels mixed. are (imposed.10, 11), thereCorrection are problems for sam with- cal Lifespecialization, on land today appear may to have be as little much effect as 25on model (1) stems from Darwin,Medievalspeciation, who viewed times. and evo- extinction NeitherThere model rarely are two was happen broad proposed methodologies except as in forstudiesent studies geographic addpansion rigor and model to analyses temporal (8–11). The showing scales equilibrium that model sister(3-6 has My)reaching higher equilibrium levels after each and between these and expansionist models, pling is clearly essential (6,11 7), 12and future inves- the diversity of modern organisms, although Darwin and Van Valen (1) allowed for extrinsic preclude expansionistRed Queen models because they may key numericalCourt Jester assumptions ( , ), and the back- times asMultilevel diverse mixed as life in the sea, so it may be lution as primarily a balanceresponse of biotic to pressures, unpredictableof species changes diversity in through the physical time, taxic ( andA). phy- The Redprevailed, Queen among may marine prevail paleobiologists at at least, Court influencesexclusive, onand evolution both Darwin in their and primarily Van Valen biotic, (1) cladesFig. 1. mayOperationvary in rate of Redof evolution, Queen timing of long-term replacement Jesterevent. The second model bemight dampened seem bystraightforward, limiting factors but such the as shortagesolution groundtigation assumption must determine of a global appropriate carrying capacity indepen is- abundancewrong to generalize and r-selected from marinelife-history paleontolog- charac- most notably competition,environment, and it was character- recallinglogenetic the ( capricious4). The taxic approachbehavior involvesorganismic treat- for and a long species time, and level represents on short a classic RedA QueenSpecies Interspecific competition Waxing and waning of clades in association with Vicariance and dispersal in major phylogenetic splits (17) ’ allowed for extrinsic influences on evolution in increases(biotic causation) in species and richness Court Jesterand morphospace (6, 7) identifies a single equilibrium level from offooddepends or on space, adequate or active assessment predation,as of the well quality as by doubtfuldent proxies (8, 10for, 11 preservation). Further, it hasand provedhuman hard error; to teristicsical studies (short to gestation all life. period, Perhaps large landlitter and size, sea ized by the Red Queen sRed statementof the Queen licensed to Alice views. in fooling of species, Medieval genera, times. or families Neither as independenttime scales,viewpoint whereas because the Court it implies Jester primarily biotic con-life span tectonic and oceanographic events (2, 17) their “primarily biotic, Red Queen views. occupation, and distributions of evolutionaryGenus the early Paleozoic, perhaps 400 Ma, to the of the fossil record. The long-term saturation some current proxies (such as number of fos- and short interbirth intervals) sometimes corre- Through the Looking-Glass thatSpeciesit takes diversity all the entities in a Red and counting Queen worldtheir occurrences de- (abiotic against causation)trols (density models dependence) at differ- on global diversity.(1-2life span My) climate andCharacter other displacementphysical factors. Further,Mass extinctions it export and smaller the logistic extinction model events to the muchLatitudinal more diversityshow similar gradient patterns (22–24) of exponential increase in model was proposed” as exclusive, and both holds his own on larger scales. The running you can do, to keeppends inSpecies the primarily same place. diversity on intrinsictime in and a factors,Red other factors.Queen such as The world body phylogenetic noveltiesent geographic ap- acrossThere and lineages are temporal two versions and scales subclades. of the equilibrium Here,(3-6 My)present (Fig. 2B). In both models, the equilibria seemsmodel that for globalthe coupled diversification logistic model(Fig. may2, triggered blue be byspeciosesiliferous extrinsic terrestrial causes localities) such realm, as are eruptions, themselves whether one dependent considers latespecieswith numbershigh species (8, richness9, 11), ( or16). perhaps they Darwin and Van Valen (1) allowed for extrinsic stippled green shape shows an area scale Temporal Red The Court Jester model (depends2) is that primarily evolution, onproach intrinsic uses cladogramsfactors, such or molecular as I( A treeswill). The toexplore Redmodel, Queenthe differing largest-scale may in prevail the time global, at when taxic the global inMilankovitch- correspond to biodiversityCourtApparent saturation in which climate change, anoxia, impact (10, 11) speciation, and extinction rarelysize,influences happen breadth except on of evolution physiologicalin in their tolerance, primarily or adapt-biotic, where Red Queen effects might be Queen? JesterRed Queen Table 1. Macroevolutionary phenomena and their support for either the Red Queen (biotic, intrinsic) or Court Jester (physical, extrinsic) models. Many organismic and species level on short ScaleSpecies (100 ky) Evolutionary arms races (1) Coordinated turnovers, originations, and extinctions Occupation of new ecospace (25) response to unpredictable changesabilitybody insize, to the hard physical breadth times. Inof a physiological Court Jester world,tolerance, spe- identified erroneously, but these are from scale effect could fit either worldview, and so are noted as “multilevel mixed.” Red Queen views. Fig. 1. Operation of Red Queen time scales, whereas the Court Jester life span in response to physical perturbations– termed environment, recalling theciesor capricious adaptability diversity behavior depends to hard on times. fluctuations In a Court in climate, Jester likelyA the result of spatial averaging of Fig. 1. Operation of Red Queen (biotic causation) Species diversity(biotic in a causation)Red Queen and world Court Jester de- holds his own on larger scales. The (1-2 My) Locality Average Biotic Entire “coordinated stasis” or “turnover pulse” hypothesis of the licensed fool of Medievallandscape,world, times. species and Neither diversity food supply. depends In reality, on fluctuations of course, Genus and Court Jester (abioticspecies causation)region/ modelscontinent at dif- Red Queen Court Jester Multilevel mixed pends primarily on intrinsic(abiotic causation) factors, modelssuch as at body differ- regionallife span responses to climate change (2, 29, 30) model was proposed as exclusive, and both stippled green shape shows an area scale Temporal ferent geographicRed andrange temporalclimate scales (A). The Red bothin climate, aspectslandscape, might prevailent and geographic in food different supply. and temporalways In and real scales at- and other(3-6 My) complex physical pertur- Milankovitch Apparent Constancy of ecological Nonconstant probability of extinction (3, 11) Subdivision of niches/specialization (10, 25) Darwin and Van Valen (1)size, allowed breadth for extrinsic of physiological tolerance, or adapt- Court Queen? zone Interspecific competition Waxing and waning of clades in association with Vicariance and dispersal in major phylogenetic splits (17) (A). The Red Queen may prevail at where Red Queen effects might be Scale (100Queen ky) may prevail at organismicRed Queen and species level influences on evolution indifferentity, their of primarily course, times, biotic, both what aspects could perhapsmight prevail be called in dif the- bations that may be in oppositeJester di- guilds through time (25) ability to hard times.organismic In a Court and Jester species world, level on spe- short Species from scale effect tectonic and oceanographic events (2, 17) identified erroneously, but these are on short time scales,Geographical whereas the scaleCourt Jester holds Incumbency advantage Lack of evidence for a global carrying capacity and Declining global extinction rates through time (1, 5) Red Queen views. multilevelferent ways mixed and at model. timedifferent scales, Traditionally, whereastimes, thewhat biologists Courtcould Jester rections,life span and so cancel each other, 22–24 cies diversity depends on fluctuations in climate, likely the result of spatial averaging of his own onLocality larger scales. Average The stippledBiotic greenEntire shape Character displacement3 24 Mass extinctions and smaller8 10 extinction events Latitudinal diversity gradient ( ) Species diversity in a Redhaveperhaps Queen tended be world called to de-think theholds in multilevel a Redhis own Queen, on mixed larger Darwinian, scales.model. The suggesting(1-2 My) no controlling effect of the ( , ) equilibrium levels ( , ) pends primarily on intrinsiclandscape, factors, such andas body food supply. In reality, of course, regional responses to climate change shows an area wherespecies Red Queenregion/ effectscontinent might be triggeredLack of by cohesiveness extrinsic causesof the great such“evolutionary as eruptions, Onshore-offshore patterns and disturbance (3) stippled green shape shows an area scale Temporal Red intrinsic,Traditionally, bioticbiologists factors way, have and tended geologists to think in in a physicalMilankovitch environment on evolution.Apparent range climate size, breadth of physiologicalboth tolerance, aspects or adapt-might prevailwhere Redin different Queen effects ways might and atbe and other complexQueen? physical pertur- B identified erroneously, but these are likely the re- climatefaunas change,” (12) anoxia, impact (10, 11) Physical-environmentalScale (100 ky) disruptionsRed Queen zone ability to hard times. In a CourtCourtadifferent Red Jester Jester, Queen, world, times, extrinsic, spe- Darwinian, whatidentified could physical perhaps intrinsic, erroneously, factors bebiotic way. but called these fac the are- bations that may be in oppositefrom scale di- effect Genus sult of spatial averaging of regional responses to Evolutionary arms races (1) CoordinatedSpecies turnovers, richness–energy originations, relationship and (18, extinctions19) ResourceOccupation use: stenotopes of new areecospace more speciose (25) than may elicit biotic responses along the Geographical scale cies diversity depends on fluctuationstorsmultilevelMuch way, in and of mixed climate,the geologists model.divergencelikely the Traditionally,in result a betweenCourt of spatial Jester,biologists averaging the Redex of- rections, and so cancelLocality each Average other,Biotic Entire life spanclimate change and other complex physical pertur- in response to physical perturbations– termed eurytopes (29, 30) landscape, and food supply. In reality, of course, regional responses to climate change red line separating Red Queenspecies andregion/ continent(3-6 My)bations that may be in opposite directions, and so Inverse relationship between global temperature and Queentrinsic,have tended andphysical Court to think factors Jester in worldaway. Red viewsQueen, may Darwinian, depend suggesting no controlling effectrange ofclimate the “coordinated stasis” or “turnover pulse” hypothesis both aspects might prevail in different ways and at and other complex physical pertur- Court Jester outcomes (B). The usagezone cancel each other, suggesting no controlling effect biodiversity (21) different times, what couldonintrinsic, perhapsMuch scale be biotic(called of2) the(Fig. the factors divergence 1):bations Bioticway, that andbetween mayinteractions geologists be in opposite the drive Red in di- a physical environment on evolution. Court (2, 29, 30) here is the microevolutionaryGeographical Red Queen, scale B Speciesof the physical environmentJester on evolution. Physical- Lack of clear correlation of species richness with multilevel mixed model. Traditionally,muchQueen of and the biologists local-scale Court rections, Jester success andworld so or cancel failure views each of may other, in- Physical-environmental disruptions Constancy of ecological Nonconstant probability of extinction (3, 11) Subdivision of niches/specialization (10, 25) Court Jester, extrinsic, physical factors way. as opposed to the macroevolutionary lifeGenus spanenvironmental disruptions may elicit biotic re- body size or other biotic factors (16) have tended to think in a Reddividuals,depend Queen, on Darwinian, populations, scale (2) suggesting(Fig. and 1 species): noBiotic controlling (Red interactions effect Queen), of the may elicit biotic responses along the (1-2 My) 25 Much of the divergencephysical environment between on the evolution. Red Red Queen that posits constant ex- life spansponses along the red line separating Red Queen guilds through time ( ) intrinsic, biotic factors way,but and perhaps geologists these in a processes are overwhelmed by redB line separating Red Queen and (3-6 My) 1 5 Court Jester, extrinsic, physicaldriveQueen factors muchand way. Court of Jester thePhysical-environmental local-scaleworld views maysuccess disruptions depend or tinction risk, a view that has been largely scale Temporal and Court Jester outcomes (B). The usage here is Incumbency advantage Lack of evidence for a global carrying capacity and Declining global extinction rates through time ( , ) CourtGenus Jester outcomes (B). The usage Milankovitch Red substantialfailure of individuals, tectonic andmaypopulations, climatic elicit biotic processes responses and speciesalong at time the life span31 2 Court (3, 24) equilibriumwww.sciencemag.org levels (8, 10) SCIENCE VOL 323 6 FEBRUARY 2009 729 Much of the divergenceon between scale ( the2) (Fig.Red5 1): Biotic interactions drive rejected ( ). Illustration based on ( ). Scale (100the ky) microevolutionaryQueen Red Queen, as opposed to Queen and Court Jester worldscales views above may depend 10 yearsred line (Court separating Jester). Red It Queen is im- and here is(3-6 the My) microevolutionary Red Queen, Species Jester Lack of cohesiveness of the great “evolutionary Onshore-offshore patterns and disturbance (3) (Redmuch ofQueen), the local-scale but Court perhaps success Jester outcomes these or failure (B processes). The of usage in- life spanthe macroevolutionary Red Queen that posits con- on scale (2) (Fig. 1): Biotic interactions drive as opposed to the macroevolutionaryCourt aredividuals, overwhelmed populations, byhere substantialand is thespecies microevolutionary (Redtectonic RedQueen), Queen, and Species stant extinctionLess risk, a view that has Morebeen largely faunas” (12) much of the local-scale success or failure of in- Red Queen that posits constantJester ex- (1-2 My) Departmentclimaticbut perhaps processes of theseEarth Sciences, processes asat opposed time University are to scales the overwhelmed ofmacroevolutionary Bristol, above Bristol 10 by5 life span rejected (31). Illustration based on (2). Species richness–energy relationship (18, 19) Resource use: stenotopes are more speciose than dividuals, populations, and species (Red Queen), (1-2 My) scale Temporal Environmental scale BS8 1RJ, UK. E-mail: [email protected] Queen that posits constant ex- tinction risk, a view that has been largely Red but perhaps these processesyearssubstantial are overwhelmed (Court tectonic Jester). by and It is climatic important processes not to export at time Milankovitch eurytopes (29, 30) tinction risk, a view that has been largely rejected scale Temporal (31). Illustration based on (2). Queen substantial tectonic and climatic processes at time5 Milankovitch Red Scale (100 ky) Inverse relationship between global temperature and organism-levelscales above 10 processesyearsrejected (Court (to31 ).regional Illustration Jester). or based It global is on im- (2). Scale (100 ky) Queen scales above 105 years (Court Jester). It is im- 21 728 SCIENCE Less More biodiversity ( ) 6 FEBRUARY 2009 VOL 323Less www.sciencemag.orgMore Department of Earth Sciences, University of Bristol, Bristol Lack of clear correlation of species richness with Department of Earth Sciences, University of Bristol, Bristol Environmental scale BS8 1RJ, UK. E-mail: [email protected] 1RJ, UK. E-mail: [email protected] Environmental scale body size or other biotic factors (16)

728 22 SCIENCE 23 728 6 FEBRUARY 2009 VOL 323 6 FEBRUARYwww.sciencemag.org 2009 VOL 323 SCIENCE www.sciencemag.org www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 729 SPECIALSECTION

first one-third of the history of the clade and that There is no geometric reason that diversification for the relative importance of the Red Queen and their continuing diversification in the Late Jurassic shifts should mainly occur low in a clade’s history: Court Jester worldviews. If the majority of diver- and Cretaceous was mainly indistinguishable from Clade shapes vary from bottom-heavy to top- sification shifts are coordinated, and associated a random walk. In particular, dinosaurs did not heavy, and diversification shifts may be concen- with particular climatic, tectonic, and geographic participate in the Cretaceous Terrestrial Revolu- trated low (dinosaurs and bats) or high (insects drivers, then the Court Jester model of macro- tion, some 130 to 100 Ma, when flowering plants, and ants) in a clade (26). evolution would prevail. This would link most leaf-eating insects, social insects, squamates, and In the future, the identification of diversification increases in species diversity to particular large- many other modern groups radiated substantially. shifts across numerous taxa may provide evidence scale radiation events, such as the Cretaceous Ter- restrial Revolution (26), or recoveries after mass

A extinctions. If, on the other hand, the majority of and upland habitats of the later Paleozoic when macroevolution, and dinosaurs provide a classic Comparing Sister Taxa macroevolution. Viewed close up, evolution is 245.9 237 232.5 228 203.6 199.6 diversification shifts are unique to particular clades,

Anisian Lad Crn Norian Rh. E J and not coordinated temporally with others, then land animals first burrowed, climbed, and flew, example. The standard view was that dinosaurs A powerful element of the comparative phyloge- all about biotic interactions in ecosystems (Red the Red Queen worldview might be considered.

PTEROSAURIA through the introduction of herbivory, giant size, originated in the Late Triassic, some 230 Ma, by netic approach to species diversity through time Queen model), but from further away, the large Comparing Sister Taxa

“Dinosauromorphs” A powerful element of the comparative phyloge- endothermy, and intelligence among vertebrates, a process of competition in which they prevailed is the opportunity to compare sister taxa. Sisters patterns of biodiversity are driven by the physi- ORNITHISCHIA Fig.netic approach3. Phylogenetic to species diversity relationships through time and mor- is the opportunity to compare sister taxa. Sisters and the great blossoming of flowering plants over their precursors, the crocodile-likecrurotar- arose from a single ancestor, and so their trajec- cal environment (Court Jester model). Dinosauria SAUROPODOMORPHA phospace occupation for Triassic archosaurs. arose from a single ancestor, and so their tra- (with associated vast expansions in diversity of sans and others, because of superior adaptations. tories occupy the same amount of time, and they THEROPODA (jectoriesA) Framework occupy the same amountphylogeny of time, and for Triassic cru- they started with the same genotype and phe- PHYTOSAURIA plant-eating and social insects and modern ver- A comparative phylogenetic study (28) shows, started with the same genotype and phenotype. Crurotarsi rotarsansnotype. Any similarities scaled in their to subsequent the Triassic evo- time scale. AETOSAURIA References and Notes Numberslution probably reflect at top this phylogenetic refer to signal millions of of years tebrates) during the Cretaceous Terrestrial Revo- however (Fig. 3), that the Dinosauria expanded Any similarities in their subsequent evolution CROCODYLOMORPHA a common origin, but differences reflect inde- 1. L. M. Van Valen, Evol. Theory 1, 1 (1973). beforependent aspects the of theirpresent; separate histories.gray bars represent the lution 100 Ma (26). in two steps, one after an extinction event 225 probably reflect this phylogenetic signal of a “Rauisuchids” Comparisons of sister taxa have allowed tests 2. A. D. Barnosky, J. Vert. Paleont. 21, 172 observed durations of major lineages; ver- POPOSAUROIDEA of the resource-use hypothesis (29), that general- The other mode of species increase globally Ma that removed dominant herbivores, and the common origin, but differences reflect indepen- (2001). ists are less speciose and have longer species ORNITHOSUCHIDAE tical dashed lines denote two extinction durations than specialists. Specialists divide the or regionally is by niche subdivision, or increas- secondfollowing the end-Triassic extinction 200 dent aspects of their separate histories. 3. D. Jablonski, Evolution 62, 715 (2008). events,physical environment at the into Carnian-Norian small patches, each and Triassic- occupied by a species, and each probably more ing specialization. This is hard to document be- Ma that removed most of the crurotarsans. Dino- Comparisons of sister taxa have allowed tests 4. A. Purvis, Annu. Rev. Ecol. Syst. 39, 301 Jurassic boundaries; arrowheads indicate B subject to environmental crises than their gener- cause of the number of other factors that vary sauria remained at moderate diversity and low of the resource-use hypothesis (29), that gener- (2008). lineagesalist relatives. Classicthat examplessurvived in support the of latter the event. Lad, 5. J. J. Sepkoski Jr., Paleobiology 10, 246 (1984). resource-use hypothesis come from studies of Ne- between ecosystems through time. However, disparity, and at lower disparity than the cruro- alists are less speciose and have longer species 0.12 Ladinian;ogene mammals Crn, (29). ForCarnian; example, two Rh, antelope Rhaetian; EJ, Early 6. J. Alroy et al., Science 321, 97 (2008). Jurassic.subgroups, the tribes(B) AlcelaphiniEmpirical and Aepycerotini,morphospace for Late mean species number in communities (alpha di- tarsans they supposedly out competed, during durations than specialists. Specialists divide the 7. J. Alroy, Proc. Natl. Acad. Sci. U.S.A. 105, 0.06 diverged 6 to 8 Ma. The former is now highly Triassicspeciose, with archosaurs, some 7 living and 25 based extinct spe- on the first two versity) has increased through time in both ma- the 25 My between the events, suggesting that physical environment into small patches, each 11536 (2008). 0 cies, and the latter is represented by two species, Dinosaur morphospace principalonly one, the impala coordinates.Aepyceros, surviving. Large The circles, dino- rine (15, 25) and terrestrial (10) systems, even there was no insistent competition driving other occupied by a species, and each probably more 8. T. D. Walker, J. W. Valentine, Am. Nat. 124, -0.06 slowly evolving Aepycerotini consists of few 887 (1984). saurs;species at anyovals, time, and pterosaurs; each of those is squares, long poposau- though niche subdivision may be less important groups to extinction but rather that the dinosaurs subject to environmental crises than their gen- -0.12 roids;lived, whereas hexagons, the speciose phytosaurs; Alcelaphini consists stars, aetosaurs; than occupation of new ecospace in increasing occupied new ecospace opportunistically, after it eralist relatives. Classic examples in support of 9. M. J. Benton, Science 268, 52 (1995). of many short-lived species. The ecological -0.18 Pterosaur morphospace 10. M. J. Benton, B. C. Emerson, Palaeontology crosses,habits of both clades crocodylomorphs; differ: The impala has a smaller black biodiversity. Further, morphological complexity had been vacated. the resource-use hypothesis come from studies Principal coordinate 2 broad, generalist diet, whereas the speciose al- 50, 23 (2007). -0.24 dots, “rauisuchids”; larger black dots, nondi- celaphines show more dietary specialization. In may be quantified, and a comparative study of A further study on Dinosauria explored the of Neogene mammals (29). For example, two 11. S. M. Stanley, Paleobiology 33, S1 (2007). nosaurianwider studies of manydinosauromorphs, clades of Neogene African Scleromochlus. -0.3 crustaceans shows, for example, that complexity subsequent evolution of the clade (26). Classic antelope subgroups, the tribes Alcelaphini and 12. J. Alroy, Evol. Ecol. Res. 6, 1 (2004). Crurotarsan morphospace Basedand South Americanon (28 mammals). (30), the resource- -0.36 use hypothesis was supported, and some subsidiary 13. G. J. Eble, Geobios 32, 223 (1999). predictions confirmed: Specialists are more com- has increased many times in parallel in separate views that the dinosaurs arose with a flourish, Aepycerotini, diverged 6 to 8 Ma. The former mon than generalists, carnivores include more lineages (27). and then finally gave way in the Cretaceous to is now highly speciose, with some 7 living and 14. M. J. Benton, Geol. J. 36, 211 (2001). -0.24 -0.16 -0.08 0 0.08 0.16 0.24 0.32 generalists than herbivores, and there are more 15. R. K. Bambach, A. H. Knoll, J. J. Sepkoski Jr., Principal coordinate 1 specialists in habitats that underwent recent en- the superior mammals, or that they dwindled to 25 extinct species, and the latter is represented vironmental change (tropical rain forests and Proc. Natl. Acad. Sci. U.S.A. 99, 6854 (2002).

Fig. 3. Phylogenetic relationships and morphospace occupation for Triassic archosaurs. (A) Framework deserts). The resource-use model then stresses Phylogenetic Studies of Clade Histories extinction because of “racial senility,” had long by two species, only one, the impala Aepyceros, 16. N. J. B. Isaac, K. E. Jones, J. L. Gittleman, A. phylogeny for Triassic crurotarsans scaled to the Triassic time scale. Numbers at top refer to millions of years the role of climate and tectonic movements in before the present; gray bars represent the observed durations of major lineages; vertical dashed lines denote determining species diversity rather than biological Species are not randomly distributed; they have been abandoned. The dinosaurs seemed to be surviving. The slowly evolving Aepycerotini Purvis, Am. Nat. 165, 600 (2005). two extinction events, at the Carnian-Norian and Triassic-Jurassic boundaries; arrowheads indicate lineages that controls such as competition and predation. 17. D. E. Wildman et al., Proc. Natl. Acad. Sci. survivedGeographic the latter event. Lad, Ladinian; and Crn, tectonic Carnian; Rh, Rhaetian; history EJ, Early has Jurassic. gener (B) Empirical- morphospacetween temperature and biodiversity on the glob- an evolutionary history, and so occur as twigs radiating actively in the Cretaceous, with many consists of few species at any time, and each atedfor Late Triassicpatterns archosaurs, of based species on the first diversity two principal coordinates.through Large time. circles, dinosaurs; al ovals,scaleOutlook for both marine and terrestrial organisms on a great phylogenetic tree. Studying species new clades appearing through their last 55 My, of those is long lived, whereas the speciose Al- U.S.A. 104, 14395 (2007). pterosaurs; squares, poposauroids; hexagons, phytosaurs; stars, aetosaurs; crosses, crocodylomorphs; smaller Paleontologists and evolutionary ecologists have 18. G. Hunt, T. M. Cronin, K. Roy, Ecol. Lett. 8, Theblack dots, slow“rauisuchids dance”; larger black of dots, the nondinosaurian continents dinosauromorphs, as PangaeaScleromochlus. Based( on21 (28),). wheredebated species generic diversity and largely familial independently. richness were as members of clades is a fruitful approach to and especially in their final 15 My. The new celaphini consists of many short-lived species. 739 (2005).

broke up during the www.sciencemag.orgpast 200 My hasSCIENCE affectedVOL 323relatively 6 FEBRUARY 2009low during warm “greenhouse”731 phas- understanding the drivers and controls on spe- study (26) shows that most diversification shifts The ecological habits of both clades differ: The 19. G. G. Mittelbach et al., Ecology 82, 2381 modern distribution patterns. Unique terrestrial es of Earth history, coinciding with relatively ciation. Key questions include (i) Do species (departures from ERM assumptions) fall in the impala has a broad, generalist diet, whereas the (2001). faunas and floras, notably those of Australia and high origination and extinction rates. diversify early in a clade’s history? (ii) How do first one-third of the history of the clade and that speciose alcelaphines show more dietary spe- 20. D. N. Schmidt, H. R. Thierstein, J. Bollmann, South America, arose because those continents A much-studied manifestation of energy and diversity and disparity (variance in characters their continuing diversification in the Late Juras- cialization. In wider studies of many clades of R. Schiebel, Science 303, 207 (2004). were islands for much of the past 100 My. Fur- temperature gradients is the latitudinal diversity or morphology) covary? (iii) Do major lineages sic and Cretaceous was mainly indistinguishable Neogene African and South American mammals 21. P. J. Mayhew, G. B. Jenkins, T. G. Benton, Proc. R. Soc. London B. Biol. Sci. 275, 47 ther, major geologic events such as the forma- gradient (LDG), namely the greater diversity of within a clade follow similar, or different, pat- from a random walk. In particular, dinosaurs (30), the resource-use hypothesis was support- (2008). tion of the Isthmus of Panama have permitted life in the tropics than in temperate or polar re- terns, and if different, why? (iv) Do evolutionary did not participate in the Cretaceous Terrestrial ed, and some subsidiary predictions confirmed: 22. G. G. Mittelbach, Ecol. Lett. 10, 315 (2007). the dispersal of terrestrial organisms and have gions, both on land and in the sea. There are two radiations follow the acquisition of new charac- Revolution, some 130 to 100 Ma, when flow- Specialists are more common than generalists, 23. D. Jablonski, K. Roy, J. W. Valentine, Science split the distributions of marine organisms. A explanations (22): (i) the time and area hypoth- ters or emptying of ecospace? (v) How do major ering plants, leaf-eating insects, social insects, carnivores include more generalists than herbi- 314, 102 (2006). classic example of vicariance is the fundamental esis, that the tropical belt is older and larger than clades of apparent competitors interact over long squamates, and many other modern groups radi- vores, and there are more specialists in habitats 24. J. W. Valentine, D. Jablonski, A. Z. Krug, K. Roy, Paleobiology 34, 169 (2008). division of placental mammals into three clades, temperate and polar zones, and so tropicalclades spans of geologic time? and (vi) How do sister ated substantially. There is no geometric reason that underwent recent environmental change 25. R. K. Bambach, A. M. Bush, D. H. Erwin, Edentata in South America, Afrotheria in Africa, have had longer to speciate, or (ii) the diversifi- clades vary in species diversity and why? that diversification shifts should mainly occur (tropical rain forests and deserts). The resource- Palaeontology 50, 1 (2007). and Boreoeutheria in the northern hemisphere, cation rate hypothesis, that there are higher rates For such analyses, the ideal is a complete low in a clade’s history: Clade shapes vary from use model then stresses the role of climate and 26. G. T. Lloyd et al., Proc. R. Soc. London B. presumably triggered by the split of those conti- of speciation and lower rates of extinction in the species tree, a phylogenetic tree that contains bottom-heavy to top-heavy, and diversification tectonic movements in determining species di- Biol. Sci. 275, 2483 (2008). nents 100 Ma (17). Other splits in species trees tropics than elsewhere. There is geological and all species living and extinct, plotted accurately shifts may be concentrated low (dinosaurs and versity rather than biological controls such as 27. S. J. Adamowicz, A. Purvis, M. A. Wills, Proc. Natl. Acad. Sci. U.S.A. 105, 4786 (2008). may relate to dispersal events, or there may be paleontological evidence for a mixture of both against geologic time (4). Simple to say; hard bats) or high (insects and ants) in a clade (26). competition and predation. 28. S. L. Brusatte, M. J. Benton, M. Ruta, G. T. no geographic component at all. hypotheses (23, 24). to achieve. More commonly, incomplete trees In the future, the identification of diversifi- Lloyd, Science 321, 1485 (2008). Species richness through time may correlate Species diversity may increase by the occu- have been used, with the risk of error in calcu- cation shifts across numerous taxa may provide Outlook 29. E. S. Vrba, Am. J. Sci. 293A, 418 (1993). with energy. The species richness–energy re- pation of new ecospace. The number of occu- lations of evolutionary rates or comparisons of evidence for the relative importance of the Red Paleontologists and evolutionary ecologists have 30. A. M. Bofarull et al., BMC Evol. Biol. 8, 97 lationship (18) posits correlations with evapo- pied guilds, that is, broad ecological groupings subclades. In paleontology, it has proven much Queen and Court Jester worldviews. If the ma- debated species diversity largely independently. (2008). 31. S. Finnegan, J. L. Payne, S. C. Wang, transpiration, temperature, or productivity, and of organisms with shared habits, has increased easier to work with higher taxa such as genera or jority of diversification shifts are coordinated, The realization that the Red Queen and Court Paleobiology 34, 318 (2008).

studies of terrestrial and marine ecosystems in several steps through time, from 20 in the families because species fossil records are less and associated with particular climatic, tectonic, Jester models may be scale-dependent, and that 32. Thanks to S. J. Braddy, P. C. J. Donoghue, D. have shown that these factors may explain as early Paleozoic to 62 in post-Paleozoic ma- complete than those of higher taxa, and yet it and geographic drivers, then the Court Jester evolution may be pluralistic (3), opens oppor- Jablonski, A. Purvis, M. Ruta, D. N. Schmidt, much as 90% of current diversity, although rela- rine faunas (25). Further, marine animals have is not clear how higher-level patterns relate to model of macroevolution would prevail. This tunities for dialog. Taxic studies in paleontol- and anonymous referees for comments and to tionships between species diversity and produc- shown several step increases in tiering, the abil- those at species level. Many key questions can would link most increases in species diversity ogy continue to have great value in highlighting S. Powell for drafting the figures. Supported by the Natural Environment Research Council and tivity change with spatial scale (19). Over long ity to occupy and exploit different levels in the be tackled by comparing a real tree to a hypo- to particular large-scale radiation events, such correlations between species richness and other the Royal Society. time spans, there are strong correlations between habitat: At times, burrowers have burrowed thetical tree that follows an equal-rate Markov as the Cretaceous Terrestrial Revolution (26), or factors, but comparative phylogenetic methods plankton morphology and diversity and water deeper, and reef-builders have built taller and (ERM) model, equivalent to tree growth after a recoveries after mass extinctions. If, on the other will illuminate questions about clade dynam- temperature: Cooling sea temperatures through more complex reefs. Analogous, if even more random walk, where equal chances of speciation hand, the majority of diversification shifts are ics, species richness, and the origin of novelties. the past 70 My, and consequent increasing ocean dramatic, expansions of ecospace have occurred and of extinction are shared by all species (4). unique to particular clades, and not coordinated Further, methods are shared by paleontologists stratification, drove a major radiation of Fora- on land, with numerous stepwise additions of Major biotic replacements, where one clade temporally with others, then the Red Queen and neontologists, and this allows direct com- minifera, associated with increasing body size new habitats, from the water-margin plants and replaces another, have been a focus of debate worldview might be considered. munication on the patterns and processes of (20). More widely, there is close tracking be- arthropods of the early Paleozoic to the forests about the roles of competition and progress in 24 25 SPECIALSECTION

41. D. Jablonski, Evolution 62, 715 (2008). 52. O. Seehausen et al., Nature 455, 620 (2008). 61. R. C. Albertson, J. T. Streelman, T. D. Kocheer, J. Hered. 42. R. J. Whittaker, K. A. Triantis, R. J. Ladle, J. Biogeogr. 35, 53. R. Lande, Genetics 128, 443 (1991). 94, 291 (2003). 977 (2008). 54. D. A. Joyce et al., Nature 435, 90 (2005). 62. K. Schwenk, N. Brede, B. Streit, Philos. Trans. R. Soc. 43. R. Gillespie, Science 303, 356 (2004). 55. G. Fryer, Environ. Biol. Fishes 45, 109 (1996). London Ser. B 363, 2805 (2008). 44. J. Todd Streelman, P. D. Danley, Trends Ecol. Evol. 18, 56. J. W. Valentine, D. Jablonski, in Evolution, Time and 63. D. Garant, S. E. Forde, A. P. Hendry, Funct. Ecol. 21, 434 126 (2003). Space, R. W. Sims, J. H. Price, P. E. S. Whalley, Eds. (2007). 45. D. Ackerly, D. Schwilk, C. Webb, Ecology 87, S50 (2006). (Academic Press, London, 1983), pp. 201–226. 64. O. Seehausen, Curr. Biol. 16, R334 (2006). 46. C. E. Parent, B. J. Crespi, Evolution 60, 2311 (2006). 57. C. O. Webb, D. D. Ackerly, M. A. McPeek, M. J. Donoghue, 65. C. Darwin, Voyage of the Boogle (Collier, New York, 47. J. A. Coyne, T. D. Price, Evolution 54, 2166 (2000). Annu. Rev. Ecol. Syst. 33, 475 (2002). 1909), p. 383. 48. J. B. Losos, D. Schluter, Nature 408, 847 (2000). 58. J. J. Wiens, C. H. Graham, Annu. Rev. Ecol. Syst. 36, 519 66. We thank A. Birand, R. Glor, A. Harrison, L. Harmon, 49. D. Lack, Darwin’s Finches (Cambridge Univ. Press, (2005). D Jablonski, L. Mahler, C. Marshall, and the anonymous Cambridge, 1947). 59. S. Skulason, T. B. Smith, Trends Ecol. Evol. 10, 366 reviewers for useful comments. Supported by the NSF 50. U. Dieckman, M. Doebeli, J. A. J. Metz, D. Tautz, Adaptive (1995). (grant DEB-0519777) and the NIH (grant GM56693). Speciation (Cambridge Univ. Press, Cambridge, 2004). 60. H. D. Bradshaw, D. W. Schemske, Nature 426, 176 51. U. K. Schliewen, D. Tautz, S. Pääbo, Nature 368, 629 (1994). (2003). 10.1126/science.1157966

REVIEW theories of the process (8, 11). I leave out dis- Speciation cussion of sympatricSPECIAL and allopatricSECTION speciation but insteadUnder identify this the Darwinian likelihood perspective, of ecological link and- once again receiving attention. The two most again theBoth parasitic models lampreys of speciation, have evolved ecologi into non-- A Gambusia way because, by chance, the pop- mutation-ordering speciation speciation with adaptation when was there relatively is gene general hypotheses involving selection are parasiticcal and forms...correlated mutation-order, with are theoretically life in small ulations do not acquire the same Evidence41. D. Jablonski, Evolution 62, for 715 (2008). Ecological52. O. Seehausen Speciationet al., Nature 455, 620 (2008). 61. R. C. Albertson, J. T. Streelman, T. D. Kocheer, J. Hered. streams, where a suitable food supply in the way mutations or fix them in the same straightforward, requiring only a test of whether ecological and mutation-order speciation. Eco- plausible, and only data can determine 42. R. J. Whittaker, K. A. Triantis, R. J. Ladle, J. Biogeogr. 35, 53. R. Lande, Genetics 128, 443 (1991). flow.94 I,ignore 291 (2003). reinforcement, a special type of of large fish is scarce or seasonal” (12). When cor- order. Divergence is therefore sto- 977 (2008). 54. D. A. Joyce et al., Nature 435, 90 (2005). 62. K. Schwenk, N. Brede, B. Streit, Philos. Trans. R. Soc. naturalphenotypic selection differences thought between to favor speciesstronger were pre- logical speciation is defined as the evolutionrelated their with relative environmentalimportance factors, in such nature. repetition The chastic but the process is distinct and43. R. Gillespie, ItsScience Alternative303, 356 (2004). 55. G. Fryer, Environ. Biol. Fishes 45, 109 (1996). matingcausedLondon by reproductive natural Ser. B 363 selection., 2805 isolation (2008). For once example, postzygotic at the of reproductive isolation between populations key is to figure out by which mechanism 44. J. Todd Streelman, P. D. Danley, Trends Ecol. Evol. 18, 56. J. W. Valentine, D. Jablonski, in Evolution, Time and 63. D. Garant, S. E. Forde, A. P. Hendry, Funct. Ecol. 21, 434 is unlikely to result from chance; environmental from genetic drift. It can occur in isolationAmerican has Association evolved. I alsofor the ignore Advancement speciation byof by divergent natural selection arising from dif-selectionreproductive pressures mustisolation therefore first be evolved the cause (3 of). both small and large (though not Dolph126 Schluter (2003). Space, R. W. Sims, J. H. Price, P. E. S. Whalley, Eds. (2007). 45. D. Ackerly, D. Schwilk, C. Webb, Ecology 87, S50 (2006). (Academic Press, London, 1983), pp. 201–226. polyploidy,Science64. O. Seehausen, 1939 even speciation thoughCurr. Biol. selection16 symposium, R334 might (2006). be [the crucial last ferences between ecological environments (2, speciation.Once “ theAs a earliest result of genetic our recent differences studies on infinite) populations. Selection 46. C. E. Parent, B. J. Crespi, Evolution 60, 2311 (2006). 57. C. O. Webb, D. D. Ackerly, M. A. McPeek, M. J. Donoghue, inmajor65. the C. earlysymposium Darwin, stages.Voyage on of thespeciation Boogle (Collier, before New the York, bio- 8, 9, 14). It predicts that reproductive isolation fishes...weighthave accumulated is constantly between being addedpopulations to the can be ecologically based under Natural selection commonly drives the origin of species, as Darwin initially claimed. Mechanisms of theory that speciation is...under the rigid control of mutation-order speciation, but 47. J. A. Coyne, T. D. Price, Evolution 54, 2166 (2000). Annu. Rev. Ecol. Syst. 33, 475 (2002). logical1909), species p. 383. concept (7)], an extensive com- should evolve between populations adapting to by either process, subsequent mutations speciation48. J. B. Losos, by D. selection Schluter, Nature fall into408, two 847 broad (2000). categories:58. ecological J. J. Wiens, C. and H. Graham, mutation-order.Annu. Rev. Ecol. Under Syst. 36, 519 66. We thank A. Birand, R. Glor, A. Harrison, L. Harmon, ” 12 Speciation and Adaptation from the environment ( ). However, this case is only ecology does not favor diver- ecological49. D. Lack, speciation,Darwin’s Finches divergence(Cambridge is Univ. driven Press, by divergent(2005). natural selection between environments, parativeD Jablonski, and biogeographic L. Mahler, C. Marshall, study and showcased the anonymousin- contrasting environments but not between pop-referringmight to the be origin favored of morphological in one population species. gence as such. It can lead to the Darwin to Dobzhansky whereasCambridge, under 1947). mutation-order speciation, divergence59. occurs S. Skulason, when T. B.different Smith, Trends mutations Ecol. Evol. arise10, 366 and stancesreviewers in which for useful derived comments. morphological Supported by theand NSF life ulations adapting to similar environments. The Theand turning not the point other for speciation because studiesof epistatic came evolution of any type of repro- 50. U. Dieckman, M. Doebeli, J. A. J. Metz, D. Tautz, Adaptive (1995). Appreciation(grant DEB-0519777)of the connection and the NIH between (grant GM56693). adapta- are fixed in separate populations adapting to similar selection pressures. Tests of parallel evolution history forms of fishes had arisen over and over basic idea has been around for a while (7), al-withinteractions the modern conceptwith genetic of speciation background“Species ductive isolation, with the excep- Speciation (Cambridge Univ. Press, Cambridge, 2004). 60. H. D. Bradshaw, D. W. Schemske, Nature 426, 176 tion and speciation began with Darwin when a separation is defined as a stage of the evolu- tion of ecologically based pre- and of51. reproductive U. K. Schliewen, isolation, D. Tautz, S. Pääbo, trait-basedNature 368 assortative, 629 (1994). mating,(2003). and reproductive isolation by active again10.1126/science.1157966 from the same ancestral type (12). The re- though it was tested only recently. The agents of (10). Hence, epistasis, including that morphologicalpeated, parallel concept origin of of nonparasitic species largely lamprey pre- divergent selection are extrinsic and can include tionaryproducing process Dobzhansky-Muller at which physiological incom isolat-- B Mimulus postzygotic isolation. selection have demonstrated that ecological speciation is a common means by which new species ” 6 vailed. In On the Origin of Species, Darwin wrote ing mechanisms become developed ( ) (here, Speciation resulting from in- arise. Evidence for mutation-order speciation by natural selection is more limited and has been in streams from the same migratory, parasitic abiotic and biotic factors such as food resources, “ patibilities” in hybrids between species that “I look at the term species, as one arbitrarily physiological is interpreted to mean evolved tragenomic conflict such as mei- best documented by instances of reproductive isolation resulting from intragenomic conflict. ancestor showed that “Again and again the para- climate, habitat, and interspecies interactions reproductive(3), can result isolation from between either populations, ecological asor otic drive or cytoplasmic male givensitictheories lampreys for of the thesake haveprocess of evolved convenience (8, 11 ). into I leave nonparasitic to a out set dis- of such as disease, competition, and behavioral in- mutation-order speciation. However,REVIEW we still have not identified all aspects of selection, and identifying the underlying genes ” distinct from geographical barriers to interbreed- sterility (Fig. 1B) is likely to be individualsforms...correlatedcussion of sympatric closely resembling with and lifeallopatric in each small speciation other... streams,and but terference. Ecological speciation can lead to the ing). Subsequently,Speciation can species be rapid were under defined both as mutation-order speciation be- for reproductive isolation remains challenging. “ whereinsteadThe amount a identifysuitable ofdifference thefood likelihood supply is onein of the very ecological way important of large and evolution of any type of reproductive isolation, “groupsspeciation of interbreeding models, natural because populations alleles that are cause, by chance, the initial muta- criterionfishmutation-order is scarce in settling or seasonal” speciation whether (1 twowhen2). When forms there correlated should is gene be including premating isolation, hybrid sterility, aredriven reproductively to fixation isolated by from natural other selection such tions causing drive and those Evidencet took evolutionary biologists for nearly Ecological 150 years, ries: ecological Speciation speciation and mutation-order ranked as species or varieties” (1). Under this groups” (7). From this point on, the study of countering it are unlikely to be t took evolutionary biologists nearly 150 would nevertheless be advantageous in both of withflow. environmental I ignore reinforcement, factors, such a special repetition type ofis and intrinsic hybrid inviability as well as extrin- in both cases. However, under the mu- but at last we can agree with Darwin that the speciation. Ecological speciation refers to the view, speciation is defined as the accumulation of speciation was the study of the evolution of the same in separate populations. years, but at last we can agree with Darwin their environments. The relative importance of unlikelynatural selection to result thought from chance; to favor environmental stronger pre- sic, ecologically based pre- and postzygotic iso- tation-order process, the same alleles, andorigin of Its species, Alternative“that mystery of mysteries” evolution of reproductive isolation between pop- sufficiently many differences between popula- reproductive isolation (3). Progress up to then in Speciation by sexual selection is I that the origin of species, “that mystery of these two categories of mechanism for the ori- selectionmating reproductive pressures must isolation therefore once be postzygotic the cause lation. Speciation by sexual selection is ecologi- if present, would be favored in every (1), really does occur by means of natural selection ulations or subsets of a single population by ad- tions to warrant their classification as separate understanding the link between morphological mutation-order speciation if di- I isolation has evolved. I also ignore speciation by mysteries” (1), really does occur by means of gin of species in nature is unknown. of speciation. “As a result of our recent studies cal speciation if ecologically based divergent se- population, at least in the early stages of (Dolph2–5). NotSchluter all species appear to evolve by aptation to different environments or ecological taxonomic species. Darwin understood the im- speciation and adaptation was largely forgotten, vergence of mate preferences or natural selection (2–5). Not all species appear to In this review, I summarize progress in un- onpolyploidy, fishes...weight even though is selectionconstantly being might beadded crucial to lection drives divergence of mating preferences, its contributionsdivergence. uncertain For this under reason, the new mutation- concept. gamete recognition occurs by the selection, but the evidence suggests that most of niches (2, 8, 9). Natural selection is divergent, portance of reproductive barriers between species Fig. 1. (A) Example of ecological speciation. Repeatedly and in the early stages. The biological species concept must surely Fig. 1. (A) Example of ecological speciation. Repeatedly and fixation of alternative advanta- themevolveNatural do. by selection The selection, effort commonly leadingbut the up drives evidence to this the conclusion originsuggests of species,actingderstanding as in Darwin contrasting the initially general directions claimed.features between Mechanisms of speciation environ- of (the1), buttheory the that study speciation of speciation is...under after the pub-rigid for example by sensory drive (15). order speciation is difficult when there independently, the mosquito fish, Gambusia hubbsi, inhabiting have made it more difficult to investigate any link independently,blue holes in the the Bahamas mosquito has evolved fish, Gambusia a larger caudal hubbsi region, inhabit and- geous mutations in different pop- involvedthatspeciation most many of by them selection experimental do. The fall effort into and two conceptual leading broadup categories: ad- to ments,by selection. ecological which Idrivesand do mutation-order. not the differentiate fixation of Underspeciation different licationcontrol of thisthe environment” work focused ( mainly12). However, on the evo- this In accordance with (10), mutation-order spe- is gene flow, because gene flow increas- Speciationand Adaptation from between speciation and natural selection. T. ingsmaller blue head holes in in the the presence Bahamas of has predators evolved (top) a larger than incaudal their ulations, as by sexual conflict vances,thisecological conclusion including speciation, involved a revision divergence many of the is experimental driven notion by of divergent alleles,by sexual natural each selection advantageous selection here between because in one environments, natural environment selec- lutioncase is ofonly speciesreferring differences, to the origin particularly of morpho of- ciation is defined as the evolution of reproduc-Dobzhanskyes the possibility (13) suggested that that favorable the genes under- muta- (16). Divergence in song and Darwin to Dobzhansky regionabsence and (bottom) smaller (29 ).head In laboratory in the presence trials, the of probability predators of (top) two speciationandwhereas conceptual under itself,mutation-order 80advances, years after including publication speciation, a revision of divergenceOn buttion notoccursdrives in the whenthe other.divergence different Following mutationsof mate G. S.preferences, arise Mani and and morphologicallogical species. traits but also of behavioral and tive isolation by the fixation of different advan-lyingtions differences occurring betweenin onepopulations population in ordinary will thanindividuals in their mating absence was (bottom) higher when (29). they In laboratory were from trials, different the other learned components of be- Appreciation of the connection between adapta- theofare the Origin fixed notion in of separate the of Speciesspeciation populations, to aitself, definition80 adapting years based after to on similar B.by selection C. either Clarke ecological pressures. (10), I defineor Tests mutation-order mutation-order of parallel evolution mecha specia-- otherThe phenotypic turning traits. point for speciation studies tageous mutations in separate populations ex-phenotypicspread traitsto other were populations, unlikely to be preventing the basis of probabilitypopulations havingof two theindividuals same predation mating environment was higher (andwhen similar they havior under purely social selec- tion and speciation began with Darwin when a publicationof reproductive of On isolation, the Origin trait-based of the Species assortative, to mating,nisms, andin most reproductive theories of isolation the process by active(8, 11). I came with the modern concept of speciation periencing similar selection pressures. Whereas reproductivedivergence isolation. (17, 1 He8). later Any changed process his result mind,- werebody shape)from different than when populations they were having from opposite the same predation preda- tion, not molded by selection for reproductive isolation instead of morphological tion as the evolution of reproductive isolation by morphologicalUnder this Darwinianconcept of perspective, species largely linking pre- 5 aselection definition have6 based7 demonstrated on reproductive that ecological isolation in speciation- leave isout a discussion common means of sympatric by which and new allopatric species “Species separation is defined as a stage of the different alleles are favored between populations buting at the in timelow thislevels viewpoint, of gene andflow, the including generally tionenvironments. environment [Photo (and credit: similar Brian body Langerhans shape) ( 29than)]. (Bwhen) Example they efficient signal transmission ( ), differences ( , ). the chance occurrence and fixation of different speciationvailed. In On with the adaptation Origin of was Species relatively, Darwin straight- wrote is the cultural equivalent of the steadarise. of Evidence morphological for mutation-order differences ( speciation6, 7). by naturalspeciation selection but instead is more identify limited the and likelihood has been of evolutionary process at which physiological under ecological speciation, the same alleles greaterselection, difficulty facilitates of studying subsequent reproductive diver isola-- wereof reproductive from opposite isolation predation evolving environments. under the mutation-order[Photo credit: The main question today is how does selec- alleles between populations adapting to similar forward,that “I look requiring at the only term a species, test of whether as one phenotyp- arbitrarily tion than morphology, must have discouraged Brianmechanism. Langerhans Male-fertile (29)]. (left)(B) Example and male-sterile of reproductive (right)flowers isolation of mutation-order process. Addition- bestThe documented main question by instances today is ofhow reproductive does selec- isolationecological resulting and mutation-order from intragenomic speciation conflict. when isolating mechanisms become developed” (6) would be favored in different populations under gence by the mutation-order process tion lead to speciation? What are the mechanisms selection pressures. Reproductive isolation evolves icgiven differences for the between sake of speciesconvenience were to caused a set by of many from pursuing the connection. Virtually no evolvingF2 hybrids under between the an mutation-order Oregon population mechanism. of monkey Male-fertile flowers (M. al scenarios are elaborated in (5). tionHowever, lead we to speciation?still have not What identified are the all mecha aspects- ofthere selection, is gene and flow. identifying I ignore reinforcement, the underlying a genes spe- (here, “physiological” is interpreted to mean mutation-order speciation. Divergence occurs (19). In contrast, ecological speciation of natural selection, what genes are affected, and because populations fix distinct mutations that naturalindividuals selection. closely For resembling example, each at the other... American” and research effort followed that tested the role of (left)guttatus and) having male-sterile a cytoplasmic (right) male flowers sterility of elementF2 hybrids and between nuclear Both models of speciation, nismsfor reproductive of natural isolation selection, remains what genes challenging. are af- cial type of natural selection thought to favor evolved reproductive isolation between popula- anyway because, by chance, the populations do can proceed with or without gene flow, anrestorer Oregon and population a closely related of monkey species flowers (M. nasutus (M.) guttatus having neither) hav- how do changes at these genes yield the habitat, would nevertheless be advantageous in both of Association“The amount for of the difference Advancement is one of very Science important 1939 adaptation in speciation. ecological and mutation-order, behavioral,fected, and mechanical,how do changes chemical, at these physiological, genes yield theirstronger environments. premating reproductive The relative importanceisolation once of speciationtions, as distinct symposium from [the geographical last major symposiumbarriers to not acquire the same mutations or fix them in the although it is easiest when gene flow ing(46 ,a47 cytoplasmic). Both flowers male shown sterility have elementM. guttatus and nuclearcytoplasm. restorer The are theoretically plausible, and criterion in settling whether two forms should be Models of Speciation by Selection flower on the left also has the nuclear restorer, whereas the one on andthe other habitat, incompatibilities behavioral, mechanical, that are the reproduc-chemical, thesepostzygotic two categories isolation of has mechanism evolved. forI also the ignore origin oninterbreeding). speciation before Subsequently, the biological species species were concept de- same order. Divergence is therefore stochastic is absent. and a closely related species (M. nasutus) having neither (46, only data can determine their rel- t took evolutionary biologists nearly 150 years, ries: ecological speciation and mutation-order ranked as species or varieties” (1). Under this 4the7). right, Both with flowers undeveloped shown anthers, have M. guttatus lacks the cytoplasm. restorer. [Photo The 7 The topic of natural selection in speciation is once ative importance in nature. The tivephysiological, barriers between and other new species? incompatibilities As a start,that the ofspeciation species inby nature polyploidy, is unknown.even though selection (fined)], an as extensive “groups comparativeof interbreeding and biogeographicnatural popu- but the process is distinct from genetic drift. It Experiments with laboratory popu- flowercredit: Andrea on the Case left (47also)] has the nuclear restorer, whereas the but at last we can agree with Darwin that the speciation. Ecological speciation refers to the view, speciation is defined as the accumulation of again receiving attention. The two most general key is to figure out by which manyare theorigin ways reproductive of by species, which barriers new“that species mystery between might of mysteriesnew arise spe by-” mightevolutionIn thisbe crucial review, of reproductive in I summarizethe early isolation stages. progress between in under- pop- studylationssufficiently showcasedthat many are instances reproductively differences in which between isolated derived popula- mor-from can occur in both small and large (though not hypotheseslations involving of Drosophila selection areand ecological yeast dem and- one on the right, with undeveloped anthers, lacks the restor- mechanism reproductive isolation I selectioncies?(1), really As cana doesstart, be occur groupedthe bymany means into ways two of naturalby broad which selection catego- new standingulations theor subsets general of features a single of population speciation by se- ad- phologicalothertions such to warrant andgroups” life their history(7). classificationFrom forms this ofpoint as fishes separateon, hadthe infinite) populations. Selection can be ecologi-mutation-orderonstrate the speciation. plausibility Ecological of ecological speciation er.isolation, [Photo includingcredit: Andrea premating Case isolation,(47)] hybrid first evolved (3). Once the earliest genetic differ-

species(2–5). Notmight all arise species by selection appearcan to be evolve grouped by lection.Speciationaptation I doto differentnotand differentiate Adaptation environments speciation from or Darwin ecologicalby sexual arisenstudytaxonomic overof speciation and species. over againwas Darwin the from study understood the same of the ancestral theevolu im-- cally based under mutation-order speciation, but is definedspeciation. as the evolutionIn those ofinstances reproductive when iso- sterility, and intrinsic hybrid inviability as well as ences have accumulated between populations by 12 lation between populations by divergent natural extrinsic, ecologically based pre- and postzygotic either process, subsequent mutations might be Biodiversityintoselection, two broad but Research the categories: evidence Centre and ecological suggests Zoology that Department,speciation most of selectiontoniches Dobzhansky (2 here, 8, because9). Natural natural selection selection is drives divergent, the typetionportance of ( reproductive). of The reproductive repeated, isolation barriers parallel (3). between originProgress of species up non- to ecology does not favor divergence as such. It measurable pre- and postmating re-

Universityand mutation-order of British Columbia, speciation. Vancouver, Ecological BC V6T spe 1Z4,- divergenceAppreciation of of mate the preferences,connection between by either adap eco-- parasiticthen1 in understanding lamprey in streams the link from between the same morpho migra-- can lead to the evolution of any type of repro-selectionproductive arising isolation from differencesevolved, between it was eco- greaterisolation. be- ( SpeciationOdsh, JYAlpha by sexual), and selectionsexual isolation is favored (ds2) in be one- population and not the other them do. The effort leading up to this conclusion acting in contrasting directions between environ- ( ), but the study of speciation after the pub- 2 8 9 14 Canada. E-mail:[email protected] logical or mutation-order mechanisms, in most tory, parasitic ancestor showed that “Again and logical environments ( , , , ). It predicts that ecological speciation if ecologically based diver- because of epistatic interactions with genetic ciationinvolved refers manyto experimental the evolution and of conceptual reproductive ad- tationments, and which speciation drivesbegan the fixation with Darwin of different when logicallication speciation of this workand focused adaptation mainly was on the largely evo- ductive isolation, with the exception of ecologi- tween lines subjected to different environments tween Drosophila species. Most of these genes10 reproductive isolation should evolve between gent selection drives divergence of mating background ( ). Hence, epistasis, including that isolationvances, including between apopulations revision ofor the subsets notion of of a aalleles, morphological each advantageous concept in of one species environment largely forgotten,lution of its species contributions differences, uncertain particularlyunder the of cally based pre- and postzygotic isolation. populationsthan between adapting lines to contrasting raised underenvironments homogeneouspreferences, show for example molecular by sensory signatures driveof (15 positive). producing selection, Dobzhansky-Muller incompatibilities single population by adaptation to different enOn- prevailed. In On the Origin of Species, Darwin new concept. Speciation resulting from intragenomic con- conditions (20, 21). Laboratory experiments proving natural10 selection’s role (3), provided that speciation itself, 80 years after publicationwww.sciencemag.org of but not in theSCIENCE other. FollowingVOL G. 323 S. Mani 6 FEBRUARY and morphological 2009 traits but also of behavioral and737 but not between populations adapting to similar In accordance with ( ), mutation-order spe- in hybrids between species (3), can result from vironmentsthe Origin of or the ecological Species, toniches a definition (2, 8, 9 based). Natu on- wroteB. C. Clarke that “I ( 10look), I atdefine the mutation-order term species, as specia- one otherThe phenotypic biological traits. species concept must surely flict such as meiotic drive or cytoplasmic male environments.on various The microbes basic idea has maintained been around under for homociation- is definedfixation as the occurred evolution before of reproductive complete reproductiveeither ecological or mutation-order speciation. ralreproductive selection is isolation divergent, instead acting of in morphological contrasting arbitrarilytion as the evolutiongiven for of the reproductive sake of convenience isolation by haveUnder made this it more Darwinian difficult perspective, to investigate linking any sterility (Fig. 1B) is likely to be mutation-order a whilegeneous (7), although conditions it was for tested many only generations recently. isolation have byisolation the fixation rather of than different afterward. advanta- The top-downSpeciation can be rapid under both speciation directionsdifferences between (6, 7). environments, which drives tothe a chanceset of individuals occurrence closely and fixation resembling of different each linkspeciation between with speciation adaptation and was natural relatively selection. straight- T. speciation because, by chance, the initial muta-Thedetected agents of divergentgenetic divergence selection are extrinsicconsistent and withgeous the mutationsapproach in separate involves populations identifying expe- (i) themodels, phenotypic because alleles are driven to fixation by the The fixation main questionof different today alleles, is how does each selec- advanta- other...”alleles between and “The populations amount of adapting difference to similaris one Dobzhanskyforward, requiring (13) onlysuggested a test of that whether the genes phenotyp- un- tions causing drive and those countering it are canmutation-order include abiotic and process biotic factors (22), such but as effects food onriencing re- similartraits under selection divergent pressures. selection, Whereas (ii) naturalthose selectiontraits in both cases. However, under resources, climate, habitat, and interspecies inter- different alleles are favored between populations the mutation-order process, the same alleles, if geoustion leadin toone speciation? environment What but are not the in mechanisms the other. veryselection important pressures.criterion Reproductive in settling isolation whether evolves two derlyingic differencesdifferences between between species populations were caused in or by- unlikely to be the same in separate populations. actionsproductive such as disease,isolation competition, have not andbeen behav- explored.under ecologicalassociated speciation, with reproductive the same alleles isolation,present, and would(iii) be favored in every population, at

Followingof natural selection,G. S. Mani whatand genes B. C. are Clarke affected, (10 and), I formsbecause should populations be ranked fix as distinctspecies mutations or varieties” that dinarynatural phenotypic selection. For traits example,were unlikely at the Americanto be the Speciation by sexual selection is mutation-order ioral interference.Two approaches Ecological investigate speciation the can mechanisms lead would be favoredthe genes in differentunderlying populations traits and under reproductiveleast in the iso early- stages of divergence. For this definehow do changes mutation-order at these genesspeciation yield as the habitat, evolu- (would1). Under nevertheless this view,be speciation advantageous is defined in both as the of basisAssociation of reproductive for the Advancement isolation. He of later Science changed 1939 speciation if divergence of mate preferences or to theof speciation evolution ofby anynatural type selection of reproductive in nature.mutation-order The lation. speciation. Step (iii) Divergence has been occurs challenging any- reason, under mutation-orderboth speciation is difficult when tionbehavioral, of reproductive mechanical, isolation chemical, by physiological, the chance accumulationtheir environments. of sufficiently The relative many importance differences of hisspeciation mind, but symposium at the time [the this last viewpoint, major symposium and the gamete recognition occurs by the fixation of bottom-up approach involves (i) genetic map- approaches but is needed to understand how se- occurrenceand other incompatibilities and fixation of that different are the alleles reproduc- be- betweenthese two populations categories ofto mechanismwarrant their for classifica the origin- generallyon speciation greater before difficulty the biological of speciesstudying concept repro- alternative advantageous mutations in different ping of reproductive isolation between closely lection has led to reproductive isolation. 738 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org tweentive barriers populations between adapting new species? to similar As a start, selec the- tionof species as separate in nature taxonomic is unknown. species. Darwin un- ductive(7)], an extensiveisolation than comparative morphology, and biogeographic must have populations, as by sexual conflict (16). Diver- related species, (ii) testing whether discovered tionmany pressures. ways by which Reproductive new species isolation might evolvesarise by derstoodIn this the review, importance I summarize of reproductive progress in barriers under- discouragedstudy showcased many instances from pursuing in which derived the connec mor-- gence in song and other learned components genes exhibit a genomic signature of positive Ecological Speciation becauseselection populations can be grouped fix intodistinct two broadmutations catego- that betweenstanding species the general (1), featuresbut the study of speciation of speciation by se- tion.phological Virtually and no life research history effort forms followed of fishes that had of behavior under purely social selection, not selection, and (iii) identifying the phenotype Evidence for ecological speciation has accumu- afterlection. the Ipublication do not differentiate of this work speciation focused by mainly sexual testedarisen the over role and of over adaptation again from in speciation. the same ancestral molded by selection for efficient signal trans- and source of fitness effects of alternative alleles lated from top-down studies of adaptation and 12 Biodiversity Research Centre and Zoology Department, onselection the evolution here because of species natural differences, selection drivesparticu the- type ( ). The repeated, parallel origin of non- mission (5), is the cultural equivalent of the at selected loci. The approach has been hugely reproductive isolation [reviewed in (2, 8, 9)]. University of British Columbia, Vancouver, BC V6T 1Z4, larlydivergence of morphological of mate preferences, traits but also by eitherof behav eco-- Modelsparasitic lampreyof Speciation in streams by Selection from the same migra- mutation-order process. Additional scenarios successful in identifying major genes implicated We now know of many real species that have, “ Canada. E-mail: [email protected] iorallogical and or other mutation-order phenotypic mechanisms,traits. in most Thetory, topic parasitic of natural ancestor selection showed thatin speciationAgain and is are elaborated in (5). in hybrid inviability (Hmr, Lhr, Nup96), sterility at least in part, evolved by divergent natural se- 26 27 www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 737 SPECIALSECTION there is gene flow, because gene flow increases the isolation repeatedly evolves between indepen- it is an evolved barrier to gene flow between possibility that favorable mutations occurring in dent populations adapting to contrasting environ- parental populations. Multiple traits are probably one population will spread to other populations, ments than between independent populations involved, including flowering time and tolerance preventing divergence (17, 18). Any process adapting to similar environments (20, 23). A of salt and drought. This type of reproductive resulting in low levels of gene flow, including major challenge in applying the test to natural isolation is context-dependent and is weakened in selection, facilitates subsequent divergence by the populations is to eliminate the possibility that each intermediate environments. On the other hand, mutation-order process (19). In contrast, ecological ecotype has originated just once and has spread to active selection favors the evolution of ever-greater speciation can proceed with or without gene flow, multiple locales. This is difficult because gene differences between populations, which may although it is easiest when gene flow is absent. flow of neutral markers between closely related strengthen the barrier to gene flow (20). Experiments with laboratory populations of but nearby populations can result in the false It is unclear how much reproductive isolation Drosophila and yeast demonstrate the plausibility appearance of multiple independent origins of typically evolves by ecologically based divergent of ecological speciation. In those instances when these populations when evaluated by phylogenies selection in nature. We can approximate an answer measurable pre- and postmating reproductive (3, 24). However, multiple origins are supported in from estimates of the combined contribution of isolation evolved, it was greater between lines several examples of parallel speciation, including active selection on traits and trait-based assortative subjected to different environments than between the sympatric benthic-limnetic species pairs of mating, as compared with other forms of re- lines raised under homogeneous conditions (20, 21). threespine stickleback in young lakes of British productive isolation (Fig. 2 and table S1). These Laboratory experiments on various microbes main- Columbia (25, 26), the repeated origin of diver- estimates are incomplete because individual tained under homogeneous conditions for many gent marine and stream populations of threespine studies may lack data on components of repro- generations have detected genetic divergence con- stickleback around the Northern Hemisphere (27), ductive isolation, separate components may not be sistent with the mutation-order process (22), but ecotypes of Timema walking stick insects living independent, and the strength of barriers between effects on reproductive isolation have not been on different host plants (28), Littorina marine snail species may not be symmetric (34). Nevertheless, explored. ecotypes inhabiting different zones of the intertidal compilation of the data shows that the amount of Two approaches investigate the mechanisms of (24), and mosquito fish inhabiting blue holes with reproductive isolation attributable to active selec- speciation by natural selection in nature. The and without fish predators in the Bahamas (29) tion and trait-based assortative mating is at least as bottom-up approach involves (i) genetic mapping (Fig. 1A). In these studies, it was shown that males strong, on average, as the amount from compo- of reproductive isolation between closely related and females are more likely to mate if they are of nents of reproductive isolation lacking identifiable species, (ii) testing whether discovered genes the same ecotype, regardless of relatedness as causes (Fig. 2). The unidentified component of exhibit a genomic signature of positive selection, indicated by phylogenetic affinity. speciation, if built by selection and not genetic and (iii) identifying the phenotype and source of Ecological speciation is also fitness effects of alternative alleles at selected loci. supported by examples of premat- The approach has been hugely successful in ing reproductive isolation in which Components – divergent selection plants (28), Littorina marine snail ecotypes of reproductive isolation lacking identifiable of mutation-order divergence. It may be that the between genes could generate reproductive iso- 20. W. R. Rice, E. E. Hostert, Evolution 47, 1637 identifying major genes implicated in hybrid individuals choose or preferentially 15 inhabiting different zones of the intertidal causes (Fig. 2). The unidentified component of mutation-order process is more difficult to de- lation (45). Mostly, I expect that he would be (1993). inviability (Hmr, Lhr, Nup96), sterility (Odsh, encounter mates on the basis of (24), and mosquito fish inhabiting blue speciation, if built by selection and not genetic tect, or perhaps we have not looked hard enough chuffed by mounting evidence for the role of 21. J. R. Dettman, C. Sirjusingh, L. M. Kohn, J. JYAlpha), and sexual isolation (ds2) between phenotypic traits that are under 10 B. Anderson, Nature 447, 585 (2007). Drosophila holes with and without fish predators in the drift, could be the result of either ecological or at species with only small ecological differences natural selection on phenotypic traits in the ori- species. Most of these genes show ecologically based divergent selec- 22. Z. D. Blount, C. Z. Borland, R. E. Lenski, molecular signatures of positive selection, proving tion. Examples include assortative Bahamas (29) (Fig. 1A). In these studies, it mutation-order mechanisms. (5). We still do not know much about the selec- gin of species. This is really what On the Origin Proc. Natl. Acad. Sci. U.S.A. 105, 7899 5 natural selection’srole(3), provided that fixation mating by host choice in insects, was shown that males and females are more These examples indicate a growing knowl- tive factors causing mutation-order speciation. of Species was all about. Between 1859 and the (2008). occurred before complete reproductive isolation body size and coloration in fish, 0 likely to mate if they are of the same eco- edge of the mechanisms of selection and its Evidence for mutation-order speciation present, the general acceptance of the biologi- 23. D. Schluter, L. M. Nagel, Am. Nat. 146, 292 rather than afterward. The top-down approach beak size in birds, pollinator pref- type, regardless of relatedness as indicated consequences for reproductive isolation. At comes from instances in which reproductive cal species concept altered the focus of specia- (1995). 24. R. K. Butlin, J. Galindo, J. W. Grahame, involves identifying (i) the phenotypic traits un- erences for specific phenotypic Components – unknown cause by phylogenetic affinity. this point, the most glaring deficiency is our isolation apparently evolved as a by-product tion studies. Yet, the discovery that reproductive der divergent selection, (ii) those traits associated floral traits, and variation in flower- Philos. Trans. R. Soc. London Ser. B Biol. Sci. — 15 Ecological speciation is also supported knowledge of the impact of selection on genes. of conflict resolution between genetic elements isolation can be brought about by ecological 363, 2997 (2008). with reproductive isolation, and (iii) the genes ing time traits inferred to be under underlying traits and reproductive isolation. Step divergent selection between environ- by examples of premating reproductive Optimistically, progress is being made with ge- within individuals (intragenomic conflict), such adaptation in ordinary phenotypic traits bridges 25. H. D. Rundle, L. Nagel, J. W. Boughman, D. (iii) has been challenging under both approaches ments [see examples in (8, 30, 31)]. Number of studies 10 isolation in which individuals choose or netic mapping to identify quantitative trait loci as cytoplasmic male sterility in hermaphroditic Darwin’s science of speciation and our own. Schluter, Science 287, 306 (2000). but is needed to understand how selection has led Ecologically based divergent preferentially encounter mates on the basis (QTLs) and genes or regulatory control regions plants (Fig. 1B), and genetic elements confer- The most obvious shortcoming of our cur- 26. E. B. Taylor, J. D. McPhail, Proc. R. Soc. to reproductive isolation. selection has also been directly 5 of phenotypic traits that are under ecologi- that affect individual phenotypic traits on which ring meiotic drive. Under both mechanisms, a rent understanding of speciation is that the London Ser. B Biol. Sci. 267, 2375 (2000). 27. J. S. McKinnon et al., Nature 429, 294 (2004). measured, as shown by reduced fit- cally based divergent selection. Examples components of reproductive isolation depend. mutation arises that can distort representation in threads connecting genes and selection are still Ecological Speciation 0 28. P. Nosil, S. P. Egan, D. J. Funk, Evolution 62, ness of each ecotype in the envi- Evidence for ecological speciation has accumu- ronment of the other [immigrant -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 include assortative mating by host choice Examples include the yup QTL, which affects gametes and spreads in a selfish manner, even few. We have many cases of ecological selec- 316 (2008). 29. R. B. Langerhans, M. E. Gifford, E. O. Joseph, lated from top-down studies of adaptation and inviability; reviewed in (31, 32)] Cumulative reproductive isolation in insects, body size and coloration in fish, flower color differences between the monkey though such an element reduces overall fit- tion generating reproductive isolation with little reproductive isolation [reviewed in (2, 8, 9)]. We and by reduced fitness of hybrids in beak size in birds, pollinator preferences for flowers,Mimulus cardinalis and M. lewisii (35). ness of the organism that bears it. This, in turn, knowledge of the genetic changes that allow Evolution 61, 2056 (2007). Fig. 2. Estimates of the magnitude of reproductive isolation now know of many real species that have, at least the parental environments [extrin- Fig. 2. Estimates of the magnitude of reproductive 30. S. Via, Trends Ecol. Evol. 16, 381 (2001). resulting from divergent selection components (top), compared specific phenotypic floral traits, and varia- Swapping alleles of this QTL between the spe- places selection on mutations in other genes that it. We have strong signatures of positive selec- in part, evolved by divergent natural selection sic postzygotic isolation (33)]. For isolation resulting from divergent selection components 31. D. B. Lowry, J. L. Modliszewski, K. M. with(top other), compared components with lacking other identifiable components causes (bottom lacking). tion in flowering time—traits inferred to be cies with repeated backcrossing resulted in shifts counter the selfish element’s effects and restore tion at genes for reproductive isolation without Wright, C. A. Wu, J. H. Willis, Philos. Trans. between environments. The connections between example, each of the coastal peren- Divergentidentifiable selection causes components (bottom include). Divergent those attributable selection to under divergent selection between environ- in pollinator preference and, hence, indirectly more equal genetic representation in gametes. enough knowledge of the mechanisms of se- Royal Soc. London Ser. B Biol. Sci. 363, 3009 selection on ordinary phenotypic traits and repro- nial and inland annual races of the active selection on traits (immigrant inviability and extrinsic ductive isolation are often strong and straight- monkey flower (Mimulus guttatus) components include those attributable to active ments [see examples in (8, 30, 31)]. affected premating isolation. Survival and salt Distorter and restorer mutations are unlikely to lection behind them. But we hardly have time (2008). postzygoticselection isolation) on traits and (immigrant to trait-based inviability assortative andmating extrinsic (habitat Ecologically based divergent selection tolerance of second-generation hybrids between be the same in different populations regardless to complain. So many new model systems for 32. P. Nosil, T. H. Vines, D. J. Funk, Evolution 59, forward. It follows that much of the genetic basis along the west coast of North preference, floral isolation, and breeding time). The unattributed of reproductive isolation should involve ordinary America has low fitness when postzygotic isolation) and to trait-based assortative has also been directly measured, as shown the sunflowers Helianthus annuus and H. peti- of environment; thus the process leads to diver- speciation are being developed that the filling of 705 (2005). componentsmating (habitat include preference, intrinsic hybrid floral inviability, isolation, sexual and breeding selection 33. H. D. Rundle, M. C. Whitlock, Evolution 55, genes that underlie differences in phenotypic transplanted to the habitat of the against hybrids, pollen competition, and reduced hybrid fecundity. by reduced fitness of each ecotype in the en- olaris transplanted to the salt marsh habitat of gence. The mismatch between the distorter in major gaps is imminent. By the time we reach time). The unattributed components include intrinsic 198 (2001). traits. But we still know little about the genetics other (31). This is an example of Data were taken from (32, 31) (table S1). A negative value hybrid inviability, sexual selection against hybrids, pollen vironment of the other [immigrant inviabil- their hybrid descendent species (H. paradoxus) one population and the restorer in the other can the bicentennial of the greatest book ever writ- 34. N. H. Martin, J. H. Willis, Evolution 61, 68 of ecological speciation. active selection on phenotypic dif- indicates that hybrids had higher fitness than the parental species competition, and reduced hybrid fecundity. Data were ity; reviewed in (31, 32)] and by reduced fit- mapped strongly to a QTL identified as the salt result in hybrid sterility or inviability and, thus, ten, I expect that we will have that much more (2007). One line of evidence comes from tests of ferences, and it also constitutes di- for at least one component of postzygotic isolation. One data value 35. H. D. Bradshaw, D. W. Schemske, Nature 426, parallel speciation, whereby greater reproductive rect reproductive isolation because oftaken–2.66 from was left(32 out, 31 of) (table the bottom S1). A panel. negative value indicates ness of hybrids in the parental environments tolerance gene CDPK3 (36). reproductive isolation (3, 41). Numerous exam- to celebrate. that hybrids had higher fitness than the parental species [extrinsic postzygotic isolation (33)]. For Another form of investigation involves the ples of selfish elements, such as those observed 176 (2003). for at least one component of postzygotic isolation. One 36. C. Lexer, Z. Lai, L. H. Rieseberg, New Phytol. SCIENCE 739example, each of the coastal perennial and analysis of genome scans of ecologically differ- in cytoplasmic male sterility of plants, support www.sciencemag.org VOL 323data 6 FEBRUARY value of –2.66 2009 was left out of the bottom panel. 161, 225 (2004). inland annual races of the monkey flower ent populations and species. These scans com- these hypotheses (42, 43). In addition, partial re- References and Notes 37. M. A. Beaumont, Trends Ecol. Evol. 20, 435 (Mimulus guttatus) along the west coast of pare allelic variation within and between popu- productive isolation generated by meiotic drive 1. C. Darwin, On the Origin of Species by Means (2005). lection between environments. The connections North America has low fitness when transplant- lations at many marker loci spaced throughout has been identified in Drosophila [reviewed of Natural Selection (J. Murray, London, 38. P. Nosil, D. J. Funk, D. Ortiz-Barrientos, Mol. 1859). Ecol. 18, 375 (2009). between selection on ordinary phenotypic traits ed to the habitat of the other (31). This is an ex- the genome (37). Markers that show excessive in (3, 41)]. Sexual conflict is also expected to 2. D. Schluter, The Ecology of Adaptive 39. S. M. Rogers, L. Bernatchez, Mol. Biol. Evol. and reproductive isolation are often strong and ample of active selection on phenotypic differ- differentiation between populations (outliers) lead to mutation-order speciation, but there are Radiation (Oxford Univ. Press, Oxford, 2000). 24, 1423 (2007). straightforward. It follows that much of the ge- ences, and it also constitutes direct reproductive may indicate selection on nearby genes. The few compelling examples (3). The contribu- 3. J. A. Coyne, H. A. Orr, Speciation (Sinauer 40. S. Via, J. West, Mol. Ecol. 17, 4334 (2008). netic basis of reproductive isolation should in- isolation because it is an evolved barrier to gene method is particularly informative when applied tion by these mechanisms to speciation is still Sunderland, MA, 2004). 41. H. A. Orr, J. P. Masly, N. Phadnis, J. Hered. volve ordinary genes that underlie differences in flow between parental populations. Multiple to populations with ongoing hybridization, be- uncertain, however. The alleles responsible for 4. P. R. Grant, B. R. Grant, How and Why 98, 103 (2007). Species Multiply: The Radiation of Darwin’s 42. D. A. Levin, Syst. Bot. 28, 5 (2003). phenotypic traits. But we still know little about traits are probably involved, including flowering cause outlier markers may identify points in the meiotic drive and cytoplasmic male sterility Finches (Princeton Univ. Press, Princeton, NJ, 43. L. H. Rieseberg, J. H. Willis, Science 317, 910 the genetics of ecological speciation. time and tolerance of salt and drought. This type genome that resist the homogenizing influence may be prevented from spreading to fixation be- 2008). (2007). One line of evidence comes from tests of of reproductive isolation is context-dependent of gene flow, perhaps indicating genomic re- cause selection on such elements is frequency- 5. T. Price, Speciation in Birds (Roberts & 44. D. C. Presgraves, Trends Genet. 24, 336 parallel speciation, whereby greater reproduc- and is weakened in intermediate environments. gions under divergent selection. However, sets dependent (43) and because restorer alleles arise Company, Greenwood Village, CO, 2008). (2008). tive isolation repeatedly evolves between in- On the other hand, active selection favors the of genes that diverged under a mutation-order and weaken selection on the distorter elements 6. T. G. Dobzhansky, Genetics and the Origin 45. D. C. Presgraves, Curr. Biol. 17, R125 (2007). of Species (Columbia Univ. Press, New York, 46. L. Fishman, J. H. Willis, Evolution 60, 1372 dependent populations adapting to contrasting evolution of ever-greater differences between process can produce the same pattern (17, 18), (44). Second, if divergent populations come 1937). (2006). environments than between independent popu- populations, which may strengthen the barrier which makes analysis of such studies more diffi- into secondary contact, the alleles within each 7. E. Mayr, Systematics and the Origin of 47. A. L. Case, J. H. Willis, Evolution 62, 1026 lations adapting to similar environments (20, to gene flow (20). cult. Clues to whether ecologically based diver- population causing cytoplasmic male sterility Species (Columbia Univ. Press, New York, (2008).

23). A major challenge in applying the test to It is unclear how much reproductive isola- gent selection is involved are gained if outliers or meiotic drive (and the corresponding restorer 1942). 48. We thank M. Arnegard, R. Barrett, A. Case, natural populations is to eliminate the possibili- tion typically evolves by ecologically based at the same genomic locations turn up repeat- alleles) will spread between the populations by 8. D. Schluter, Trends Ecol. Evol. 16, 372 (2001). H. Hoekstra, M. Noor, P. Nosil, S. Otto, T. 9. H. D. Rundle, P. Nosil, Ecol. Lett. 8, 336 ty that each ecotype has originated just once and divergent selection in nature. We can approxi- edly in scans between populations that inhabit gene flow, eliminating that component of repro- Price, L. Rieseberg, S. Rogers, A. Schluter, (2005). S. Via, M. Whitlock, J. Willis, and a reviewer has spread to multiple locales. This is difficult mate an answer from estimates of the combined contrasting environments (38) and by identify- ductive isolation (43). Hence, for these mecha- 10. G. S. Mani, B. C. Clarke, Proc. R. Soc. for assistance and comments. This work was because gene flow of neutral markers between contribution of active selection on traits and ing phenotypic traits under divergent selection nisms to contribute to speciation, the fitness of London Ser. B Biol. Sci. 240, 29 (1990). supported by grants from the Natural Sciences

closely related but nearby populations can result trait-based assortative mating, as compared with that map to those locations in the genome (36, hybrids must be reduced to very low levels, or 11. M. Turelli, N. H. Barton, J. A. Coyne, Trends and Engineering Research Council of Canada in the false appearance of multiple independent other forms of reproductive isolation (Fig. 2 39, 40). As genomic resources increase for more other incompatibilities must arise that interact Ecol. Evol. 16, 330 (2001). and the Canada Foundation for Innovation. origins of these populations when evaluated by and table S1). These estimates are incomplete species, it will be possible to measure natural with these genes to prevent their spread after 12. C. L. Hubbs, Am. Nat. 74, 198 (1940). 13. T. Dobzhansky, Am. Nat. 74, 312 (1940). phylogenies (3, 24). However, multiple origins because individual studies may lack data on selection directly on genomic regions of inter- secondary contact. 14. Additional references are available as Supporting Online Material are supported in several examples of parallel components of reproductive isolation, separate est by transplanting otherwise relatively ho- supporting material on Science Online. www.sciencemag.org/content/full/323/5915/737/ speciation, including the sympatric benthic- components may not be independent, and the mogenous experimental populations containing Conclusions 15. J. A. Endler, Am. Nat. 139, S125 (1992). DC1 Tables S1 to S3 limnetic species pairs of threespine stickleback strength of barriers between species may not be alternative alleles into the environments of the Our understanding of the role of natural selec- 16. W. R. Rice et al., Proc. Natl. Acad. Sci. U.S.A. 102, 6527 (2005). References in young lakes of British Columbia (25, 26), the symmetric (34). Nevertheless, compilation of parent species (35). tion in speciation has come a long way since 10.1126/science.1160006 17. N. Barton, B. O. Bengtsson, Heredity 57, 357 repeated origin of divergent marine and stream the data shows that the amount of reproductive Darwin’s time. If he were here to witness, he (1986). populations of threespine stickleback around the isolation attributable to active selection and trait- Mutation-Order Speciation would most likely be staggered by the discover- 18. A. S. Kondrashov, Evolution 57, 151 (2003). Northern Hemisphere (27), ecotypes of Timema based assortative mating is at least as strong, Mounting evidence for divergent selection in ies of genes and molecular evolution and aston- 19. L. H. Rieseberg, personal communication walking stick insects living on different host on average, as the amount from components speciation does not diminish the potential role ished at the prospect that evolutionary conflict (2008). 28 29 SPECIALSECTION

32. P. Nosil, T. H. Vines, D. J. Funk, Evolution 59, 705 (2005). 40. S. Via, J. West, Mol. Ecol. 17, 4334 (2008). S. Rogers, A. Schluter, S. Via, M. Whitlock, J. Willis, and 33. H. D. Rundle, M. C. Whitlock, Evolution 55, 198 (2001). 41. H. A. Orr, J. P. Masly, N. Phadnis, J. Hered. 98, 103 a reviewer for assistance and comments. This work 34. N. H. Martin, J. H. Willis, Evolution 61, 68 (2007). (2007). was supported by grants from the Natural Sciences and 35. H. D. Bradshaw, D. W. Schemske, Nature 426, 176 (2003). 42. D. A. Levin, Syst. Bot. 28, 5 (2003). Engineering Research Council of Canada and the Canada 36. C. Lexer, Z. Lai, L. H. Rieseberg, New Phytol. 161, 225 43. L. H. Rieseberg, J. H. Willis, Science 317, 910 (2007). Foundation for Innovation. (2004). 44. D. C. Presgraves, Trends Genet. 24, 336 (2008). 37. M. A. Beaumont, Trends Ecol. Evol. 20, 435 (2005). 45. D. C. Presgraves, Curr. Biol. 17, R125 (2007). Supporting Online Material www.sciencemag.org/cgi/content/full/323/5915/737/DC1 38. P. Nosil, D. J. Funk, D. Ortiz-Barrientos, Mol. Ecol. 18, 46. L. Fishman, J. H. Willis, Evolution 60, 1372 (2006). Tables S1 to S3 375 (2009). 47. A. L. Case, J. H. Willis, Evolution 62, 1026 (2008). References 39. S. M. Rogers, L. Bernatchez, Mol. Biol. Evol. 24, 1423 48. We thank M. Arnegard, R. Barrett, A. Case, H. Hoekstra, (2007). M. Noor, P. Nosil, S. Otto, T. Price, L. Rieseberg, 10.1126/science.1160006

REVIEW values that place two bacterial isolates into the Speciation same or differentSPECIAL species. TheSECTION overall genetic relatednesspneumoniae of is isolates a major may human be measured pathogen, by theS. ABS. pseudopneumoniae S. pneumoniae extentmitis is of a DNA commensal hybridization bacteria between with them, a history and 3% divergent The32. P. Nosil, T. Bacterial H. Vines, D. J. Funk, Evolution 59 Species, 705 (2005). 40. S. Via, Challenge: J. West, Mol. Ecol. 17, 4334 (2008). S. Rogers, A. Schluter, S. Via, M. Whitlock, J. Willis, and 1.2% divergent those that show 70% or more DNA hybrid- V. ordalii 33. H. D. Rundle, M. C. Whitlock, Evolution 55, 198 (2001). 41. H. A. Orr, J. P. Masly, N. Phadnis, J. Hered. 98, 103 of taxonomica reviewer for uncertainty assistance and ( comments.11), and This S. workpseudo- V. rumoiensis 34. N. H. Martin, J. H. Willis, Evolution 61, 68 (2007). (2007). izationpneumoniaewas aresupported defined is a by recently grants as the from described same the Natural species organism Sciences (2, 10 and of). Making35. H. D. Bradshaw, D. W. Schemske, SenseNature 426, 176 of (2003). Genetic42. D. A. Levin, Syst. and Bot. 28, 5 (2003). SuchuncertainEngineering cutoffs status implyResearch that nonetheless that Council sequences of Canada corresponds and that the cluster Canada to s V. alginolyticus 36. C. Lexer, Z. Lai, L. H. Rieseberg, New Phytol. 161, 225 43. L. H. Rieseberg, J. H. Willis, Science 317, 910 (2007). togetherFoundation with for a Innovation. certain amount of similarity (2004). 44. D. C. Presgraves, Trends Genet. 24, 336 (2008). a distinct cluster in these data (12). There are Ecological Diversity mustSupporting be from Online the Material same species, and moreover 37. M. A. Beaumont, Trends Ecol. Evol. 20, 435 (2005). 45. D. C. Presgraves, Curr. Biol. 17, R125 (2007). striking differences in the amount of sequence Enterovibrio norvegicus thatwww.sciencemag.org/cgi/content/full/323/5915/737/DC1 this cutoff value is applicable to all groups V. aestuariaV. fischeri/logeinu 38. P. Nosil, D. J. Funk, D. Ortiz-Barrientos, Mol. Ecol. 18, 46. L. Fishman, J. H. Willis, Evolution 60, 1372 (2006). diversity observed within homologous house- 1 2,3,4 2 1 1 Tables S1 to S3 Christophe375 (2009). Fraser, * Eric J. Alm, Martin F. Polz,47. A.Brian L. Case, G. J. Spratt,H. Willis, EvolutionWilliam62, P. 1026 Hanage (2008). of bacteria or archaea. Recent MLSA studies, V. calviensis keepingReferences genes in these named species, ranging . fischeri 39. S. M. Rogers, L. Bernatchez, Mol. Biol. Evol. 24, 1423 48. We thank M. Arnegard, R. Barrett, A. Case, H. Hoekstra, which use the concatenated sequences of mul- S. mitis V from 1.2% for S. pneumoniae to 3.0% for S. . supersteus The(2007). Bacteria and Archaea are the most genetically diverseM. Noor, superkingdoms P. Nosil, S. Otto, of T. Price, life, L.and Rieseberg, techniques tiple10.1126/science.1160006 housekeeping genes to discern clustering 5% divergent V pseudopneumoniae and up to 5.0% for S. mitis. for exploring that diversity are only just becoming widespread. Taxonomists classify these patterns among populations of closely related The distance between two randomly selected V. panecida* organisms into species in much the same way as they classify eukaryotes, but differences in their taxa, suggest that species defined by taxono- — S.values mitisthat genotypes place twois similar bacterial to the isolates average into dis the- biologyREVIEWincluding horizontal gene transfer between distantly related taxa and variable rates of mists in many cases correspond to well-resolved homologous recombination—mean that we still do not understand what a bacterial species is. This sequencetancesame orbetween different clusters. S. pneumoniaespecies.However, The these and overall studiesS. pseudo genetic also- showpneumoniaerelatedness that there of genotypes isolates is no universal may (5.1%) be cutoff measured(2). This or descrip-implies by the is not merely a semantic question; evolutionary theory should be able to explain why species exist at all levels of the tree of life, and we need to be able to define species for practical torthatextent ofthe clustersof use DNA of thata hybridization fixed characterizes level betweenof sequence a species. them, diver Fur- and- S. oralis The Bacterial Species Challenge: Vibrio splendidus applications in industry, agriculture, and medicine. Recent studies have emphasized the need to thermore,gencethose that for inspection showdifferentiating 70% of or thespecies more clusters DNA would does hybrid- tend not 0.005 combine genetic diversity and distinct ecology in an attempt to define species in a coherent and alwaystoization either clearly are rejoin defined reveal S. pneumoniae as which the same level speciesand in the S. hierarchypseudo (2, 10).-

convincingMaking fashion. The Sense resulting data mayof help Genetic to discriminate among and the many theories of ispneumoniaeSuch more cutoffs fundamental, or imply break thanthat up sequences anyS. mitis other so (Fig. thatthat 1)clusternearly (7).

prokaryotic species that have been produced to date. everytogetherAs isolate an with example, was a certaina Fig.species 1A amount showsof its own. of the similarity relation- This is 0.005 Ecological Diversity shipsclearlymust beamong unsatisfactory. from multiple the same isolates species, of and three moreover closely he species debate in microbiology is not reflect only a tiny subset of those characters that relatedthat this streptococcal cutoff value is species. applicableStreptococcus to all groups Fig. 1. Multilocus sequence analysis of closely related species. (A) Radial associations of Vibrionaceae sequence clusters (13). Habitats (colored dots) were Christopheonlyhe species about Fraser, adebate1 human* Eric in desireJ. microbiology Alm, to2,3,4 catalogMartin is bac- not F. Polz,allowtraordinary2 Brian bacteria G.variety Spratt,to use of 1 differentmicrobialWilliam resources life, P. Hanage much in 1of the it pneumoniaeHabitatsof bacteria and oris aEcological archaea. major human Recent Differentiation pathogen, MLSAS. studies, mitis minimum evolution tree constructed using MEGA4, showing clusters among 97 estimated as differential distributions of groups of closely related strains among terialonly diversity about a in human a consistent desire manner, to catalog but environment,uncultured (1). and Beyond only capture this, taxa a small too fractionsimilar ofto isAwhich aclear commensal usenatural the concatenatedcriterion bacteria withto identify sequences a history clusters of of taxo- mul- of isolates of four Streptococcus species identified as indicated. The tree was built samples (size fractions enriched in different environmental resources). Clusters T using concatenates of six housekeeping loci, resulting in a total of 2751 positions associated with named species are evident, and in most cases species show a clear isThe also Bacteria abacterial fundamental and diversity Archaea argument in are a because theconsistent most of genetically what man it- thebe diverse distinguished true diversitysuperkingdoms and in circumscribed this of superkingdom life, andby techniques rRNA of life. se- nomicevolutionarytiple housekeeping uncertaintyimportance, (11 genes), and whichS. to pseudopneumoniae discern we might clustering want T patterns among populations of closely related in the final data set (2). Distances were calculated as the percentage of variant predilection for one of the habitats. The exception is V. splendidus, which breaks revealsner,for exploringbut aboutis also our that a fundamental ignorance diversity of are argument how only evolutionary just because becoming Morequences widespread. recently, have Taxonomistsrevealed molecular furthermethods classify diversity these [particularly through isto acall recently species, described is to find organismecological of uncertain features status that nucleotide sites. The mean distance within the clusters, calculated by MEGA4, is up into many closely related ecological populations. Asterisk denotes that trees oforganisms what it intoreveals species about in our much ignorance the same of wayhow as theymultilocus classify sequence eukaryotes, analysis but differences (MLSA) (in2) their and distinguishtaxa, suggest them that from species close defined relatives. by Among taxono- forces form, shape, and extinguish bacterial ge- DNA-DNA hybridization and ribosomal RNA that nonetheless corresponds to a distinct cluster in shown. To the right, the pneumococcal cluster is shown at larger scale, and based on additional loci indicate that the placement of V. panecida within V. biology—including horizontal gene transfer between distantly related taxa and variable rates of mists in many cases correspond to well-resolved neticevolutionary lineages, forces of the form, mechanisms shape, and of extinguish differen- (rRNA)metagenomic sequencing] studies have (1 helped), and tothis define diversity species, thesepathogens, data ( 12 the). There ability are to striking cause differences a distinctive in putative subclusters are indicated in dark gray, purple, and green. (B) Ecological splendidus may be an artifact of horizontal gene transfer at the Hsp60 locus. tiationbacterialhomologous between genetic recombination subpopulations lineages, of— themean sharing mechanisms that common we still of dobutneeds not these understand to be methods explained what have a by bacterial serious theory. limitations species Thus, is.practi Thisand- thediseasesequence amount has clusters. of historically sequence However, diversity been these used observed studies to withindefine also descent,differentiationis not merely and of abetween the semantic process subpopulations question; of adaptation evolutionary sharing to new theorycannotcal difficulties, should reliably be assign ablelack to aof large explaintheory, collection and why observations species of similar homologousspecies,show that but there pathogens housekeeping is no universal constitute genes cutoff inonly these or a descrip-minute named nichescommonexist at and all descent, changing levels ofand theenvironments. of treethe process of life, Animal andof adapta we spe- need- tostrainsof bevast able toamounts species to define of (e.g., as species yet rRNA unclassified for sequences practical microbial are too species,fractiontor of clusters rangingof overall that from bacterial characterizes 1.2% diversity. for S. apneumoniae species. Mapping Fur- ofto butvation; pathogens many constitute or most models only a minute of a population fraction of diversity.not, per se, The a very first interesting proposed observation; mechanism many or was makesstrict the reference accumulation to the of neutralunifying diversity. biological The ciestionapplications areto new defined niches in industry, by and their changing morphological agriculture, environments. and and medicine. be- conserveddiversity Recent studieshave to resolve all have fueled similar emphasized the controversy species). the rRNA need of how tose- 3.0%bacterialthermore, for S. diversity inspectionpseudopneumoniae onto of environmental the clustersand up to resources does 5.0% notfor overallreproducing bacterial with diversity. a small Mapping amount of of mutation bacterial basedmost models on artificial of a population selection reproducing experiments with with a principlesfirst proposed of mechanismselection and was niche based partitioning, on artificial havioralAnimalcombine traitsspecies genetic and are diversity by defined their and ability distinctby or their inability ecology morpho to- in anquenceone attempt defines surveys to a definebacterial have, however, species species in revealed 3( a– coherent8). the extra- and S.indicatesalways mitis. clearly The that distance closely reveal which betweenrelated level groups two in randomly the of hierarchy bacteria se- diversitywill eventually onto environmental produce populations resources consisting indicates smallbacteria amount grown of mutation for extended will eventually periods produce under theselection ecotype experiments has rightly with become bacteria popular grown as for a interbreed,logicalconvincing and but fashion.behavioral suchThe categories traits resulting and cannot by data their easily may abil help be- toordinary discriminate variety among of microbial the many life, theories much of of it lectedcanis more beS. ecologically fundamental mitis genotypes divergent. than is any similar For other example, to (Fig. the average 1) fine- (7). thatof clusters closely of related related groups organisms, of bacteria irrespective can be eco- of stablepopulations conditions consisting in chemostats, of clusters ofwhich related showed orga- frameworkextended periods within under which stable to conditions discuss bacterial in che- appliedityprokaryotic or inability to the species Bacteriato interbreed, that or have Archaea but been such (or produced categories indeed to to date.unculturedGenetic Clustering (1). Beyond this, taxa too similar to be distancescaleAs resource an between example, partitioningS. Fig.pneumoniae 1A hasshows beenand theS. observed relation- pseudo- logicallythe details divergent. of the evolutionary For example, forces fine-scale or ecologi re-- repeatednisms, irrespective selective of sweeps the details in which of the the evolution- whole evolution,mostats, which speciation, showed and repeated ecology. selective sweeps manycannot eukaryotic easily be microbes).applied to Instead,the Bacteria taxonomists or Ar- distinguishedDarwin commented and circumscribed that “all true byclassification rRNA se- pneumoniaeamongships among coastalgenotypes multiple Vibrio isolates(5.1%) populations ( of2). three This coexisting implies closely sourcecal differentiation. partitioning A has more been substantial observed observa among- genomeary forces was or ecological thought to differentiation. hitchhike to A fixation more inThe which ecotype the whole model genome (4, 16, 2 was4) predicts thought that to havechaea been (orhe indeed speciesforced to debate relymany on ineukaryotic biochemical microbiology microbes). tests is and not quencesisreflect genealogical” only have a tiny revealed [(9 subset), p. further404]. of those Taxonomists diversity characters through have that thatinrelated the the use water streptococcal of a column fixed level species. (1 of3). sequence PartitioningStreptococcus divergence was coastaltion is thatVibrio therepopulations is very little coexisting neutral in thediversity water substantialalong with observationan advantageous is that mutation there is very(periodic little commonhitchhike ancestry to fixation will along be preserved with an advantageous among bac- pneumoniae is a major human pathogen, S. mitis limitedInstead,only morphological taxonomists about a human have characteristics desire been forcedto catalog for this to pur-bac- rely multilocusthusallow used bacteria sequence sequence to use relatedness analysis different (MLSA) resourcesto define (2 in)cutoff and the fordiscovered differentiating because species strains would were collected tend to eitherfrom columnin many ( 13 populations). Partitioning of wasmicrobes, discovered from because which neutralselection) diversity (22). in Selective many populations sweeps of can microbes, purge terialmutation populations (periodicselection) within niches (22). Selective (which sweepsshould terial diversity in a consistent manner, but environment, and only capture a small fraction of is a commensal bacteria with a history of taxo- pose.onT biochemical Naturally, biochemicaltests and limited characters morphological have been metagenomicvalues that place studies two ( 1bacterial), and this isolates diversity into needs the rejoindistinct,S. pneumoniaeecologically andinformativeS. pseudopneumoniae samples, and, strainswe may were infer collected some fromfeatures distinct, of the ecologically selective almostfrom which all genetic we may diversity infer some in the features population of the becan monophyletic), purge almost all and genetic thus diversity predicts in that the pop-eco- is also a fundamental argument because of what it the true diversity in this superkingdom of life. nomic uncertainty (11), and S. pseudopneumoniae selectedcharacteristics for the conveniencefor this purpose. of taxonomists; Naturally, theybio- tosame be orexplained different by species. theory. The Thus, overall practical genetic dif- orthe break phylogenetic up S. mitis structureso that nearly of theevery ecologically isolate was informativelandscape. Neutral samples, diversity and the phylogenetic is the amount struc- of selectiveand thus constitute landscape. a candidate Neutral diversity mechanism is thefor typesulation are and coherent thus constitute self-contained a candidate gene mecha- pools. reveals about our ignorance of how evolutionary More recently, molecular methods [particularly is a recently described organism of uncertain status chemical characters have been selected for the ficulties,relatedness lack of ofisolates theory, may and be observations measured ofby vastthe adifferentiated species of its own.populations This is clearly was superimposedunsatisfactory. turepolymorphism of the ecologically that is evident differentiated in noncoding populations re- amountreducing of neutral polymorphism variation that(23). is evident in non- Asnism a result, for reducing it has neutralbeen suggested variation that (23). ecotypes 1forces form, shape, and extinguish bacterial ge- DNA-DNA hybridization and ribosomal RNA that nonetheless corresponds to a distinct cluster in convenienceDepartment of of Infectious taxonomists; Disease Epidemiology,they reflect Imperial only a amountsextent of ofDNA as yet hybridization unclassified between microbial them, diversity and on their habitats. Habitats were defined using wasgions superimposed or results in on synonymous their habitats. substitutions.Habitats were coding regions or results in synonymous sub- should be considered as putative or actual spe- netic lineages, of the mechanisms2 of differen- (rRNA) sequencing] have helped to define species, Habitatsthese data and (12). Ecological There are Differentiation striking differences in Niches and Ecotypes Collegetiny subset London, of London those W2 characters 1PG, UK. Departmentthat allow of bac Civil- havethose allthat fueled show the 70% controversy or more DNA of how hybridiza one de-- an empirical modeling approach. This analysis definedOne common using measure an empirical of neutral modeling diversity approach. is the Nichesstitutions. and One Ecotypes common measure of neutral cies, depending on the level of genetic differ- andtiation Environmental between subpopulationsEngineering, Massachusetts sharing Institute common of finesbut these a bacterial methods species have (3 serious–8). limitations and Athe clear amount natural of sequence criterion diversity to identify observed clusters within of This analysis revealed high levels of specialization diversity is the effective population size N , de- To extend this model, one can consider multiple teria to use different resources in 3the environ- tion are defined as the same species (2, 10). Such revealed high levels of specialization for some effective population size N , defined as the size To extend this model, one can consider multiplee entiation from the ancestral population. This Technology,descent, and Cambridge, of the process MA 02139, of adaptationUSA. Department to new of cannot reliably assign a large collection of similar evolutionaryhomologous housekeeping importance, which genes we in these might named want for some populations (e.g., V.e ordalii is only found fined as the size of a population evolving in the ecological niches characterized by the selective Biologicalment, and Engineering, only capture Massachusetts a small Institute fraction of Technol-of the cutoffs imply that sequences that cluster together populations (e.g., V. ordalii is only found as of a population evolving in the absence of se- ecological niches characterized by the selective model therefore has the advantage of providing niches and changing environments.4 Animal spe- Geneticstrains to Clustering species (e.g., rRNA sequences are too species, ranging from 1.2% for S. pneumoniae to ogy, Cambridge, MA 02139, USA. Broad Institute of MIT to call species, is to find ecological features that as single free-swimming cells), whereas others are absence of selection that would generate as much advantages they confer to specific genes. This is truecies diversity are defined in this by theirsuperkingdom morphological of life. and More be- withconserved a certain to resolveamount similarof similarity“ species). must rRNA be from se- single3.0% for free-swimmingS. pseudopneumoniae cells), whereasand up toothers 5.0% are for lection that would generate as much neutral di- advantages they confer to specific genes. This a mechanistic understanding of the evolution- and Harvard University, Cambridge, MA 02139, USA. Darwin commented that all true classification distinguish them from close relatives. Among more generalist (Fig. 1B) and can colonize a wide neutral diversity as is actually observed. Esti- the ecotype model, where genes adapted to recently,havioral traitsmolecular and by methods their ability [particularly or inabilityDNA- to thequence same surveys species,” have, and however, moreover revealed that this the cutoff extra- moreS. mitis generalist. The distance (Fig. 1 betweenB) and can two colonize randomlya wide se- versity as is actually observed. Estimates of Ne is the ecotype model, where genes adapted5 to9 ary processes, as well as an organizing principle *To whom correspondence should be addressed. E-mail: is genealogical [(9), p. 404]. Taxonomists have pathogens, the ability to cause a distinctive dis- variety of surfaces, including5 organic9 particles and mates of Ne for bacteria range from 10 to 10 specific niches cause selective sweeps within [email protected], hybridization but such and categories ribosomal cannot RNA easily (rRNA) be valueordinary is applicable variety of to microbial all groups life, of muchbacteria of or it varietylected S. of mitis surfaces,genotypes including is similar organic to the particles average for bacteria range from 10 to 10 (14–18). To specific– niches cause selective sweeps within for classifying species, that is based on experi- thus used sequence relatedness to define cutoff ease has historically been used to define species, zooplankton in the watercolumn (13). Most of the (14 18). To put this into context, the numbers of those niches but not in other niches. In this way sequencing]applied to the have Bacteriahelped or to Archaea define (or species, indeed but to archaea.uncultured Recent (1). Beyond MLSA this, studies, taxa toowhich similar use to the be anddistancezooplankton between inS. the pneumoniae water columnand (S.13 pseudo-). Most put this into context, the numbers of Vibrio cells those niches but not in other niches. In this mental observations of bacterial populations. predicted Vibrio populations are deeply divergent Vibrio cells per cubic meter of seawater in tem- the population will undergo adaptation and dif- thesemany methods eukaryotic have microbes). serious Instead,limitations taxonomistsand can- concatenateddistinguished sequences and circumscribedof multiple by housekeep rRNA se-- ofpneumoniae the predictedgenotypesVibrio (5.1%) populations (2). This are implies deeply fromper cubic each meter other, of and seawater in many in cases temperate are congruentcoastal perateway the coastal population regions range will from undergo 108 to adaptation 109 (19), ferentiationHowever, while these maintaining observations relatively of low repeated levels SCIENCE 741 8 9 nothave reliably been forced assign to a rely large on biochemicalcollection www.sciencemag.orgof tests similar and ingquences genes have to discern revealed clustering furtherVOL diversity patterns 323 6throughamong FEBRUARY divergentthat the 2009 use from of a fixed each level other, of sequence and in many divergence cases withregions named range species; from 10 however, to 10 V.(19 splendidus), which sugis a- andwhich differentiation suggests vast while census maintaining population relatively sizes selectiveof neutral diversity, sweeps were as selectivemade sweeps in chemostats, confined 20 strainslimitedto morphological species (e.g., characteristics rRNA sequences for this are pur-too populationsmultilocus sequenceof closely analysis related taxa, (MLSA) suggest (2) that and arefor congruent differentiating with species named wouldspecies; tend however, to either V. notablegests vast exceptioncensus andpopulation splits into sizes numerous (>10 closely). This low(>10 20 levels). Thisof observation neutral diversity,—a mismatch as selective of many whereasto each ecotypenatural environments regularly purge are the markedly populationun-

conservedpose. Naturally, to resolve biochemicalsimilarcharacters species). haverRNA been se- speciesmetagenomicdefined studies by ( 1taxonomists), and this diversity in many needs cases splendidusrejoin S. pneumoniae is a notableand exceptionS. pseudopneumoniae and splits into, relatedobservation—a groups withmismatch distinct ecological of many preferences, orders of sweepsorders of confinedmagnitude betweento each effective ecotype populationregularly stableof any and diversity diverse. that How might would have one accumulated detect the

quenceselected surveys for the conveniencehave, however, of taxonomists; revealed the they ex- correspondto be explained to well-resolved by theory. Thus,sequence practical clusters. dif- numerousor break up closelyS. mitis sorelated that nearly groups every with isolate distinct was presumablymagnitude indicatingbetween effective recent ecological population radiationsize purgesize and the censuspopulation population of any diversity size (true that of might most presence(Fig. 2A). of Crucially, selective whatsweeps neutral in diversity natural bacte we do-

However,ficulties, lack these of theory,studies andalso observations show that there of vastis ecologicala species of preferences, its own. This ispresumably clearly unsatisfactory.indicating fromand census a sympatric population ancestral size population (true of (most13). Thus,bac- havebacteriaaccumulated studied to date) (Fig.— 2wasA). originally Crucially, used what to rialobserve populations? is predicted The to most be associated conclusive withexamples adapt-

1Department of Infectious Disease Epidemiology, Imperial noamounts universal of as cutoff yet unclassified or descriptor microbial of clusters diversity that recent ecological radiation from a sympatric geneticteria studied clusterstothat date)—was correlate with originally ecology used can be to neutralcounter claims diversity of neutrality we do observe and instead is arguepredicted that comeive traits. not Thefrom ability bacteria of suchbut from selective RNA sweeps viruses, to College London, London W2 1PG, UK. 2Department of Civil characterizeshave all fueled a species. the controversy Furthermore, of how inspection one de- ancestralHabitats population and Ecological (13). DifferentiationThus, genetic clusters discerned.counter claims of neutrality and instead argue toall begenetic associated variation with was adaptive adaptive traits. (20, 21The). whichlimit the mutate effective at much population higher sizerates has than been DNA- rec- and Environmental Engineering, Massachusetts Institute of – 3 offines the a bacterialclusters doesspecies not (3always8). clearly reveal thatA clear correlate natural with criterionecology to can identify be discerned. clusters of thatWhat all genetic do the variation genetic data was tell adaptive us about (20 mech-, 21). abilityHowever, of there such are selective several sweepsdifferent to mechanisms limit the basedognized life for forms. some timeIt has (17been, 23 ), established and this model from Technology, Cambridge, MA 02139, USA. Department of which level in the hierarchy is more fundamen- evolutionaryWhat do the importance, genetic data which tell us we about might mech want- However, there are several different mecha- effective population size has been recognized sequences collected over many years that the Biological Engineering, Massachusetts Institute of Technol- anisms of population differentiation and the that can explain this mismatch (Fig. 2). has been substantially developed by Cohan and Genetic Clustering ogy, Cambridge, MA 02139, USA. 4Broad Institute of MIT tal than any other (Fig. 1) (7). anismsto call species, of population is to finddifferentiation ecological features and that the evolutionarynisms that can historyexplain of thethis microbesmismatch in (Fig. question? 2). for Whatever some time mechanisms (17, 23), areand driving this model the differ- has populationcolleagues (structure4, 16, 24). of Because the human it links influenza patterns ofvi- and Harvard University, Cambridge, MA 02139, USA. DarwinAs an commented example, Fig. that 1“Aall shows true classificationthe relation- evolutionarydistinguish them history from of close the microbes relatives. in Among ques- ThatWhatever bacteria are mechanisms organized into are genetic driving clusters the is beenentiation substantially of bacteria developed into clusters, by they Cohan must and re- rusgenetic is predominantly differentiation withdriven adaptation, by repeated and makesselec- *To whom correspondence should be addressed. E-mail: shipsis genealogical among multiple” [(9), p. isolates 404]. Taxonomists of three closely have tion?pathogens, That thebacteria ability are to organized cause a distinctive into genetic dis- differentiation of bacteria into clusters, they colleagues (4, 16, 24). Because it links patterns tive sweeps (25) and that the resulting effective [email protected] relatedthus used streptococcal sequence relatedness species. toStreptococcus define cutoff clustersease has is historically not, per se, been a very used interesting to define species, obser- must restrict the accumulation of neutral of genetic differentiation with adaptation, and population size N (<100) is very much smaller 742 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org e 30 31 www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 741 Speciation A metapopulation may evolve, differentiate, and not eliminate recombination between closely re- tween a “clonal” organism in which clusters are “ ” adaptA metapopulation without global may selective evolve, sweeps. differentiate, Diversity and latednot eliminate species. Because recombination of such between interspecies closely recom- re- predictedtween a clonal to divergeorganism (the in green which line) clusters and are a lostadapt by without a local global selective selective sweep sweeps. in one patch Diversity may bination,lated species. any Because given isolate of such within interspecies a species recom- is “predictedsexual” organism to diverge (the (the blue green line) in line) which and they a “ ” belost rescued by a local and selective reintroduced sweep from in one other patch patches. may almostbination, certain any to given contain isolate at least within some a genetic species ma- is aresexual predictedorganism to be (the held blue together line) in by which recombi- they Thebe rescued ecotype and model, reintroduced with its from predicted other patches. mono- terialalmost that certain is characteristic to contain at of least other some closely genetic related ma- nation.are predicted For “sexual to be” heldspecies, together the divergence by recombi- of “ ” phyleticThe ecotype relationship model, between with its nichepredicted and mono- geno- species.terial that Hence, is characteristic whereas it of was other once closely thought related that clustersnation. For requiressexual a processspecies, that the reduces divergence the rate of type,phyletic may relationship therefore not between be an appropriate niche and model geno- bacteriaspecies. do Hence, not form whereas species it wasin the once eukaryotic thought sense that ofclusters recombination requires abetween process them that— reducesfor example, the rate a — oftype, speciation may therefore in complex not be ecosystems. an appropriate model becausebacteria do they not do form not species recombine in the eukaryotic at all (32), sense one periodof recombination of allopatry between or ecological them differentiation.for example, a of speciation in complex ecosystems. currentbecause view they dois that not recombinethey do not at form all (32 species), one Theperiod speciation of allopatry point or is ecologicalthe amount differentiation. of divergence Choosing Between Models becausecurrent viewthey recombine is that they too do much not (5 form). species betweenThe speciation clusters point that is the needs amount to accumulate of divergence to Choosing Between Models It has proven difficult to discriminate between becauseIn asexual they recombine clonal organisms, too much even (5). in the ab- preventbetween them clusters from that returning needs to to a single accumulate cluster to if modelsIt has proven of population difficult differentiation to discriminate that focusbetween on senceIn asexual of any clonal selective organisms, pressure, even clusters in the will ab- theprevent barriers them to from recombination returning toa are single removed. cluster A if ecotypesmodels of or population metapopulations. differentiation For example, that focus the on spontaneouslysence of any selective split into pressure, multiple clusters lineages will or recentthe barriers study to hypothesized recombination that are tworemoved. related A ecotypicecotypes orstructure metapopulations. of a soil Bacillus For example,has been the “spontaneouslydaughter” clusters split (15 into). However, multiple under lineages certain or Campylobacterrecent study hypothesizedspecies are currently that two undergoing related “ ” modeledecotypic structure to predict of a a priori soil Bacillus which sequencehas been circumstancesdaughter clusters recombination (15). However, can underprevent certain this, thisCampylobacter process of mergingspecies areinto currently a single species undergoing as a clustersmodeled were to predict ecotypes, a priori and hence which which sequence ones andcircumstances we can hence recombination divide the bacteria can preventinto “sexual this,” resultthis process of changes of merging in their into environment a single species (33). as a “ ” shouldclusters be were associated ecotypes, with and specific hence which ecological ones and we“nonsexual can hence” species. divide the This bacteria effect, into describedsexual at resultThe of above changes insights in their were environment reached using (33). mod- “ ” propertiesshould be (associated16). Some with clusters specific are associatedecological greaterand nonsexual length elsewherespecies. This(15), effect, is summarized described inat els basedThe above on the insights assumption were that reached genetic using variation mod- withproperties certain (16 phenotypic). Some traits,clusters such are as associated a propen- Fig.greater 3, which length shows elsewhere the rate (15 at), which is summarized two clusters in isels neutral. based on Although the assumption this is obviously that genetic not variation always sitywith to certain grow onphenotypic shady north-facing traits, such slopes as a propen- or sun- divergeFig. 3, which over shows time— thethat rate is, at the which increase two clusters in the anis neutral. appropriate Although assumption, this is obviously it is plausible not that always the — nysity south-facingto grow on shady slopes. north-facing However, slopes this or model sun- meandiverge genetic over timedistancethat between is, the them. increase If this inbe- the numberan appropriate of loci assumption, explicitlyinvolved it is plausible in adaptive that the fittedny south-facing no better (and slopes. in fact However, slightly worse) this modelthan a comesmean genetic negative, distance then the between two clusters them. If will this stop be- ecologicalnumber of differentiation loci explicitly will involved be small, in andadaptive thus versionfitted no of better the model (and in with fact several slightly subpopulations worse) than a divergingcomes negative, and instead then converge. the two clusters The three will exam- stop thatecological in an unstable differentiation landscape, will be genomic small, and barriers thus andversion diversity of the generated model with only several by neutral subpopulations drift. This plesdiverging shown and in instead Fig. 3 converge. differ only The in three the rate exam- of tothat recombination in an unstable will landscape, depend more genomic on the barriers accu- versionand diversity of the generated model was only dismissed by neutral because drift. of This its homologousples shown in recombination Fig. 3 differ between only in the the clusters, rate of mulationto recombination of differences will depend at neutral more onloci the than accu- at associationversion of the with model a very was low dismissed estimate because of popula- of its allhomologous other parameters recombination being held between constant. the clusters, As re- adaptivemulation loci. of differences The models at also neutral assumed loci a than homo- at SPECIALSECTION tionassociation size (14 with). However, a very low estimates estimate of of effective popula- combinationall other parameters increases, being we held see a constant. distinction As be- re- geneousadaptive loci. distribution The models of polymorphisms also assumed a homo- across tion size (14). However, estimates of effective combination increases, we see a distinction be- geneous distribution of polymorphisms across population size Ne are often grossly disconnected population size N are often grossly disconnected reference to the unifying biological principles of A E1 E2 B Metapopulation structure, in which the popu- usingfrom censusestablished populatione ecological sizes. criteria, It has proven a hypoth very- Recombination rate relative to mutation from census population sizes. It has proven very Recombination rate relative to mutation selection and niche partitioning, the ecotype has new and empty lation is divided into patches and where indi- esischallengingthat can be to tested. find models that successfully 0.10 clonal colonization 3 rightly become popular as a framework within viduals disperse between patches, can generate explainchallengingThis problem low estimatedto findof low models power values to thatof detectN successfullywhile selection pro- 3 0.10 clonal e 0.53 threshold which to discuss bacterial evolution, speciation, very low effective population sizes if patches (or,vidingexplain more better low accurately, estimated predictions to values rejectthan ofmodels neutrality)Ne while based ispro- on a viding better predictions than models based on 0.53 threshold and ecology. turn over (i.e., if patches are only intermittently verysimple general neutral problem drift. in The population analysis genetics of Bacillus that 2 2.00 sexual

) 2.00 sexual The ecotype model (4, 16, 24) predicts that * * simple neutral drift. The analysis of Bacillus 2 able to support bacterial growth, and if a small doespartly not did negate this bythe predictingimportance more of adaptation ecotypes in in -9 )

partly did this by predicting more ecotypes in -9 common ancestry will be preserved among bac- * number of bacteria are dispersed to colonize evolution,the model than but rather were observed suggests using that more established work * colonization the model than were observed using established terial populations within niches (which should be * empty patches) (Fig. 2B) (27). This structure isecological needed ifcriteria, we want a hypothesis model-based that methods can be 1 istance between istance between 1 monophyletic), and thus predicts that ecotypes tested.ecological criteria, a hypothesis that can be well describes the situation for parasites, which to discriminate among different biologically istance between tested. are coherent self-contained gene pools. As a re- can colonize a host but are then forced to move plausibleThis problem explanations of low power of genetic to detect data. selection In sult, it has been suggested that ecotypes should This problem of low power to detect selection 0 on because the host develops immunity or dies Table(or, more 1 we accurately, proposeto a reject scheme neutrality) for performing is a very 0 be considered as putative or actual species, de- (or, more accurately, to reject neutrality) is a very * old and void (17). It also describes any situation where bac- analysesgeneral problem that could in population be used genetics to test, develop, that does pending on the level of genetic differentiation general problem in population genetics that does teria use a limited resource intensively for short andnot negatevalidate the different importance competing of adaptation models in more evo- from the ancestral population. This model there- not negate the importance of adaptation in evo- -1 * bursts, followed by dispersal to new resource systematically.lution, but rather suggests that more work is -1 lution, but rather suggests that more work is rate in d of change fore has the advantage of providing a mechanistic stable ecotype needed if we want model-based methods to dis- "Speciation point" for clusters per generation (x10 clusters per generation patches (e.g., colonization of organic particles needed if we want model-based methods to dis- rate in d of change sexual"Speciation species point" for understanding of the evolutionary processes, as concept (x10 clusters per generation in seawater by Vibrio populations). This meta- Homologouscriminate among Recombination different biologically plausible -2 sexual species well as an organizing principle for classifying CD explanationscriminate among of genetic different data. biologically In Table 1 plausible we pro- -2

population model is fundamentally different One specific challenge to models that invoke Relative r species, that is based on experimental observa- poseexplanations a scheme of for genetic performing data. Inanalyses Table 1that we could pro- Relative r tions of bacterial populations. 20 from the ecotype model because it does not ecotypicpose a scheme structure for performinginvolves a feature analyses ofthat bacterial could Bacteria Phage be used to test, develop, and validate different -3 However, these observations of repeated se- predict an association between neutral diversity evolution—homologousbe used to test, develop, and recombination—that validate different -3 competing models more systematically. 0% 10% 20% 30% lective sweeps were made in chemostats, whereas 15 and adaptive traits. wecompeting have not models yet discussed. more systematically. Bacterial reproduc- 0% 10% 20% 30% Mean genetic distance between clusters natural environments are markedly unstable and The relevance of the metapopulation model tionHomologous does not involve Recombinationthe obligate reassortment Mean genetic distance between clusters diverse. How would one detect the presence of 10 to the species question is that, although highly ofHomologous genetic material Recombination observed in most higher or- Fig. 3. The dynamics of cluster divergence. The figure summarizes some key results from (15)ina One specific challenge to models that invoke Fig. 3. The dynamics of cluster divergence. The figure summarizes some key results from (15)ina selective sweeps in natural bacterial populations? idealized and simplified, it may capture some of ganisms.One specific However, challenge recombination to models does that occur invoke in phase-space plot of the genetic dynamics of two populations, with recombination occurring be- ecotypic structure involves a feature of bacterial phase-space plot of the genetic dynamics of two populations, with recombination occurring be- Population size tween them at a rate that is varied for the three different simulations. The y axis shows the rate of The most conclusive examples come not from 5 the effects of complexity and instability of actu- bacteriaevolutionecotypicand structure— homologousarchaea involves (29) and recombination a featuretypically of involves bacterial—that — — changetween them of genetic at a rate distance that is between varied for the the clusters three as different a function simulations. of the genetic The y distanceaxis shows itself the (x rateaxis). of bacteria but from RNA viruses, which mutate at al ecosystems on population structure. Selective theweevolution havereplacement not yethomologous discussed. of a short Bacterial recombinationpiece of reproduction DNA withthat Whenchange the of rategenetic of change distance is between positive, the the clusters populations as a function will diverge of the genetically; genetic distance when negative, itself (x axis). they much higher rates than DNA-based life forms. It 0 sweeps are predicted to be inevitable in simple, thedoeswe have homologous not not involve yet discussed. segment the obligate Bacterial from reassortment another reproductionstrain. of converge.When the rate The of direction change of is positive, change forthe eachpopulations scenario will is diverge shown genetically; by arrows color-coded when negative, to each they has been established from sequences collected 0 20 40 60 80 100 stable environments but not in complex meta- Recombinationdoes not involve becomes the obligate less reassortment probable with of genetic material observed in most higher orga- scenario.converge. For The low direction recombination of change rates, for the each populations scenario is are shown effectively by arrows clonal color-coded and always to diverge each over many years that the population structure of Time populations [a point partly addressed in (28)]. A increasinggenetic material sequence observed divergence in most between higherorga- the nisms. However, recombination does occur in (greenscenario. line). For As low the recombination recombination rates, rate the increases, populations the cohesive are effectively effects ofclonal recombination and always slow diverge the metapopulation may evolve, differentiate, and donornisms. and However, the recipient recombination (30, 31), which does occurreduces in the human influenza virus is predominantly driven Fig. 2. Different models of microbial evolution that lead to low values of N .(A) The ecotype bacteria and archaea (29) and typically involves rate(green of line). divergence, As the until recombination a threshold rate ispassed increases, (red the line) cohesive and the effects populations of recombination become effectively slow the 25 e bacteria and archaea (29) and typically involves by repeated selective sweeps ( ) and that the model of bacterial population differentiation. The tree shows a single bacterial lineage that dif- adapt without global selective sweeps. Diver- butthe does replacement not eliminate of a short recombination piece of DNA between with sexualrate of in divergence, the sense until that athe threshold populations is passed no longer (red line) diverge. and For the recombination populations become rates above effectively this N the replacement of a short piece of DNA with resulting effective population size e (<100) is ferentiates into two sublineages (E1 and E2) that differ in some aspect of their ecology. Periodic sity lost by a local selective sweep in one patch closelythe homologous related species. segment Because from anotherof such strain.inter- level,sexual the in fate the ofsense the thattwo populations the populations will depend no longer on how diverge. genetically For recombination distinct they are rates at theabove outset. this very much smaller than observed for bacteria. selection (a selective sweep) occurs at the points marked by asterisks and eliminates almost all of may be rescued and reintroduced from other speciesRecombinationthe homologous recombination, becomes segmentany less fromgiven probable anotherisolate with within strain. in- Iflevel, they the are fate within of the the two“speciation populations point, will” then depend recombination on how genetically will cause distinct them they to merge. are at the If they outset. are “ ” The use of longitudinal ecological and genetic the diversity that has arisen since the last episode of periodic selection, which is shown by the patches. The ecotype model, with its predicted acreasingRecombination species sequence is almost becomes divergence certain less to between probable contain the withat donor least in- fartherIf they are away within than the thisspeciation“speciation point, point,then” they recombination will continue will to cause diverge them from to each merge. other. If they These are “ ” data to distinguish between competing models dashed branches (diversity purged by periodic selection) or solid branches (existing diversity) on monophyletic relationship between niche and someandcreasing the genetic recipient sequence material (30 divergence, 31 ), that which is between characteristic reduces the but donor does of curvesfarther are away derived than usingthis speciation the model point, describedthey in will (15). continue to diverge from each other. These of evolution has a long pedigree in eukaryotic the tree. As the two populations are ecologically distinct (i.e., ecotypes), periodic selection in one genotype, may therefore not be an appropriate otherand the closely recipient related (30, species.31), which Hence, reduces whereas but does it curves are derived using the model described in (15). biology (26). On the basis of these analogies, sublineage does not influence diversity in the other sublineage and vice versa. Each ecotype can model of speciation in complex ecosystems. was once thought that bacteria do not form spe- any inference of a population structure driven by therefore diverge to become separate species. Reproduced from (24) with permission. (B) A meta- 744 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org 744 cies in the eukaryotic sense because they6 FEBRUARY do not 2009 VOL 323 SCIENCE www.sciencemag.org selective sweeps would require good longitudi- population. Patches of varying size (gray circles) are vacant (empty) or may be colonized by a Choosing Between Models recombine at all (32), one current view is that process that reduces the rate of recombination Illegitimate Recombination and Gene single genotype randomly acquired from another patch. Strains may diversify within a patch (as nal data from natural bacterial populations, as It has proven difficult to discriminate between they do not form species because they recom- between them—for example, a period of allopa- Content Variation well as observations of episodic crashes in shown by different colors representing distinct genotypes), which may colonize empty patches as models of population differentiation that fo- bine too much (5). try or ecological differentiation. The speciation Illegitimate recombination or gene acquisition described above. A characteristic of this sort of metapopulation is patch turnover, in which patches diversity causally associated with genetic changes cus on ecotypes or metapopulations. For ex- In asexual clonal organisms, even in the point is the amount of divergence between clus- is another unusual feature of bacteria. In this and not associated with changes in ecological occasionally become unable to support colonization and their inhabitants are removed (solid gray circles). (C) A neutral model with small population size. Different genotypes (different colors) arise ample, the ecotypic structure of a soil Bacil- absence of any selective pressure, clusters will ters that needs to accumulate to prevent them case, genes or clusters of genes are acquired covariates. by mutation or recombination and increase or decrease in the population by random drift. For lus has been modeled to predict a priori which spontaneously split into multiple lineages or from returning to a single cluster if the barriers that typically have no homolog(s) in the recipi-

some purposes, this simple model is an adequate effective description of the more complex pro- sequence clusters were ecotypes, and hence “daughter” clusters (15). However, under cer- to recombination are removed. A recent study ent strain. The importance of this phenomenon Bottlenecks, Metapopulations, and Local Extinctions cesses represented in (A), (B), and (D), and of other more complex evolutionary models not de- which ones should be associated with specific tain circumstances recombination can prevent hypothesized that two related Campylobacter is evident in the clear and ubiquitous signature scribed in this review. (D) Predator-prey dynamics and population bottlenecks. Regular population ecological properties (16). Some clusters are this, and we can hence divide the bacteria into species are currently undergoing this process of such events in the growing body of genomic The essential element of the ecotype model with bottlenecks can drastically shrink the effective population size. In this case, bacteria-phage associated with certain phenotypic traits, such “sexual” and “nonsexual” species. This effect, of merging into a single species as a result of data. These are identified by differences in the respect to limiting neutral diversity is not niche predator-prey dynamics are simulated with a classical Lotka-Volterra model, which can generate as a propensity to grow on shady north-facing described at greater length elsewhere (15), is changes in their environment (33). characteristics of the acquired DNA and that of adaptation per se, but rather the effective bottle- oscillations in population size of any amplitude. Population sizes and time axes are in arbitrary slopes or sunny south-facing slopes. However, summarized in Fig. 3, which shows the rate The above insights were reached using mod- the host strain, for example, in base composition neck caused by the replacement of the whole units for illustrative purposes only. this model fitted no better (and in fact slightly at which two clusters diverge over time—that els based on the assumption that genetic varia- or codon usage; in most cases, the donor of the population by descendants from a single indi- worse) than a version of the model with several is, the increase in the mean genetic distance tion is neutral. Although this is obviously not DNA in question is unknown. Gene acquisition vidual and the resulting extinction of all other lineages (Fig. 2A). Other mechanisms that induce (27). This structure well describes the situation does not predict an association between neutral subpopulations and diversity generated only between them. If this becomes negative, then always an appropriate assumption, it is plau- leads to genomes being punctuated by stretches Bottlenecks, Metapopulations, and Local or involve regular population bottlenecks will thanfor parasites, observed which for bacteria. can colonize The use a host of longitu but are- diversity and adaptive traits. by neutral drift. This version of the model was the two clusters will stop diverging and instead sible that the number of loci explicitly involved of foreign DNA. The largest of these (which Extinctions also restrict neutral diversity. Metapopulation dinalthen forced ecological to moveand on genetic because data the to host distinguish develops The relevance of the metapopulation model to dismissed because of its association with a very converge. The three examples shown in Fig. 3 in adaptive ecological differentiation will be may be many kilobases in length) were initially structure, in which the population is divided into betweenimmunity competing or dies ( 17models). It alsoof evolution describes has any a Thethe species essential question element is that, of althoughthe ecotype highly model ide- low estimate of population size (14). However, differ only in the rate of homologous recom- small, and thus that in an unstable landscape, termed “pathogenicity islands,” becausethe new patches and where individuals disperse between longsituation pedigree where in bacteria eukaryotic use a biology limited ( resource26). On withalized respect and simplified, to limiting it mayneutral capture diversity some is of not the estimates of effective population size Ne are of- bination between the clusters, all other param- genomic barriers to recombination will depend functions encoded by the imports were often patches, can generate very low effective popu- theintensively basis of for these short analogies, bursts, followed any inference by dispersal of a nicheeffects adaptation of complexity per se, and but instability rather the of effecactual- ten grossly disconnected from census popula- eters being held constant. As recombination in- more on the accumulation of differences at neu- involved in virulence, but a better term is “ge- lation sizes if patches turn over (i.e., if patches populationto new resourcestructure patches driven (e.g., by selective colonization sweeps of tiveecosystems bottleneck on population caused by structure.the replacement Selective of tion sizes. It has proven very challenging to find creases, we see a distinction between a “clonal” tral loci than at adaptive loci. The models also nomic islands” as the phenomenon is far from are only intermittently able to support bacterial wouldorganic require particles good in longitudinal seawater bydataVibrio frompopula- natu- thesweeps whole arepopulation predicted to by be descendantsinevitable in simple,from a models that successfully explain low estimated organism in which clusters are predicted to di- assumed a homogeneous distribution of poly- limited to pathogens (36, 37). Although it is growth, and if a small number of bacteria are raltions). bacterial This metapopulationpopulations, as well model as isobservations fundamen- singlestable environments individual and but the not resulting in complexextinction meta- values of Ne while providing better predictions verge (the green line) and a “sexual” organism morphisms across the genome, and violation hard to quantify the selective impact of import- dispersed to colonize empty patches) (Fig. 2B) oftally episodic different crashes from thein diversity ecotype modelcausally because associ it- ofpopulations all other [alineages point partly (Fig. 2 addressedA). Other in mecha (28)].- than models based on simple neutral drift. The (the blue line) in which they are predicted to be of this may alter the tempo and mode of these ing any given gene(s) into a new background, ated with genetic changes and not associated nisms that induce or involve regular population analysis of Bacillus partly did this by predicting held together by recombination. For “sexual” processes (34, 35). the occasional ability to gain a new adaptation with changes in ecological covariates. bottlenecks will also restrict neutral diversity. more ecotypes in the model than were observed species, the divergence of clusters requires a in this fashion—such as a new metabolic capa- www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 743 32 33 SPECIALSECTION the genome, and violation of this may alter the may be present in all sequenced genomes of a which individualsSPECIAL are similarSECTION enough, or inter- tempo and mode of these processes (34, 35). named species (41). We may consider genes breed enough, that individual variant genes the genome, and violation of this may alter the withinmay a be named present species in all as sequenced being characteristic genomes of of a whichcompete individuals directly arefor reproductivesimilar enough, success. or inter- Prac- Illegitimatetempo and Recombination mode of these and processes (34, 35). differentnamed levels species of ( ecological41). We may specificity, consider rang-genes breedtical advances enough, building that individual on this or variant other theoretical genes Gene Content Variation ingwithin from a highly named speciesconserved as being core characteristic functions that of competeconcepts directly will only for reproductive come when success. these Prac- are de- IllegitimateIllegitimate recombination Recombination or gene and acquisition is aredifferent essential levels for growth of ecological in all specificity, environments rang- to ticalveloped advances into building explicit on models this or other and theoretical model-based anotherGene unusual Content feature Variation of bacteria. In this case, lociing that from are involved highly conserved with adaptation core functions to a specific that conceptsalgorithms will that only are come tested when and refined these are on a de- wide genesIllegitimate or clusters recombinationof genes are acquired or gene acquisition that typ- is habitat.are essential Some narrow for growth niche-specific in all environments genes may to velopedrange of into data. explicit Alternatively, models it and may model-based be sensible to ically haveanother no unusual homolog(s) feature in of the bacteria. recipient In strain. this case,beloci distributed that are involved across species, with adaptation being to transferred a specific algorithmssuggest an that ad are hoc tested application and refined of on principles a wide to The importancegenes or clusters of this ofphenomenon genes are acquired is evident that in typ-betweenhabitat. them Some by narrow mobile niche-specific elements. The genes evolu- may rangedifferent of data. genera Alternatively, on the basis it mayof their be sensible specific to char- the clearically and have ubiquitous no homolog(s) signature in theof such recipient events strain. tionarybe distributed fate of such across genes species, may being hence transferred be only suggestacteristics, an ad including hoc application the extent of principles of variation to in The importance of this phenomenon is evident in between them by mobile elements. The evolu- different genera on the basis of their specific char- in the growing body of genomic data. These are loosely coupled with that of any particular gene content and recombination. In any case, no the clear and ubiquitous signature of such events tionary fate of such genes may hence be only acteristics, including the extent of variation in species or strain in which they are found, and biologist would deny the importance of ecology identifiedin the by growing differences body in of the genomic characteristics data. These of are loosely coupled with that of any particular gene content and recombination. In any case, no the acquiredidentified DNA by anddifferences that of in the the host characteristics strain, for of theyspecies are maintained or strain in throughwhich they selection are found, by and the biologistto what would we observe, deny the but importance it may not of ecology be easy to example,thein acquired base composition DNA and that or of co- the host strain, for they are maintained through selection by the to what we observe,incorporate but it it may in a not fashion be easy that to is bility or a new mode of trans- ticular species or strain in which these alternatives. One clear result from all of niches within (for example) the nasopharynx or don usage;example, in most in base cases, composition the donor ofor co- incorporateconvenient it for in ataxonomists. fashion that None- is 21. R. Milkman, Trends Biochem. Sci. 1, N152 mission for a pathogen—may they are found, and they are main- the studies discussed here is that the underlying gut is difficult, and studies of the relationships (1976). the DNAdon usage; in question in most cases,is unknown. the donor of convenienttheless, population for taxonomists. geneticists None- may be of enormous importance in tained through selection by the theoretical questions concerning species will between bacterial populations and ecology may 22. K. C. Atwood, L. K. Schneider, F. J. Ryan, Genethe acquisition DNA in leads question to genomes is unknown. theless,have little population choice geneticists but to tackle may the terms of speciation. habitat to which each host strain is not be answered in the absence of more detailed be more fruitful for some environmental spe- Proc. Natl. Acad. Sci. U.S.A. 37, 146 (1951). beingGene punctuated acquisition by leads stretches to genomes of havequestion little of choice defining but bacterialto tackle species the 23. B. R. Levin, Genetics 99, 1 (1981). Perhaps even more striking adapted. In the case of very mobile genetic-environmental mapping. Moreover, cies where the categorization of niches is a more foreignbeing DNA. punctuated The largest by of stretches these of questionor, at ofthe defining very least,bacterial populations. species 24. F. M. Cohan, E. B. Perry, Curr. Biol. 17, R373 is the amount of variation in elements—for example, plasmids some guidelines for the types of ecological tractable enterprise. Hopefully, we will soon ob- (whichforeign may be DNA. many The kilobases largest of in theseEcological or,Whether at the we very are least, estimating populations. effective (2007). gene content revealed by mul- factorEcological b encoding resistance to antibiotics studies that will be most informative are emerg- tain richer data sets that map bacterial diversity 25. A. Rambaut et al., Nature 453, 615 (2008). length)(which were initially may be termed many kilobases“patho- in Whetherpopulation we sizeare estimating from neutral effective diversity tiple genomes from the same factor b or heavy metals—the ecological ing. Most important, the ecological data collect- onto ecology and provide a way to distinguish 26. R. A. Fisher, E. B. Ford, Heredity 1, 143 genicitylength) islands, were” because initially termed the new“patho- populationor choosing size from an appropriate neutral diversity set of species, which ” implies that specificity determined by these ed must be relevant to the niche boundaries of among various models of population differen- (1947). genicity islands, because the new or choosing an appropriate set of 27. M. Slatkin, Theor. Popul. Biol. 12, 253 functionsgene encoded acquisition by the occurs imports at were a strainsaccessory to test loci for may positive have selection no link at the populations studied. And if genetic groups tiation and speciation, including those based on functions encoded by the imports were strains to test for positive selection at (1977). often involved in virulence, but a bet- a locus of interest, species definitions surprisingly high frequency. It ato locus the ofsequence interest,we species observe definitions using do not map exclusively onto sampling catego- ecotypes or metapopulations. 28. J. Majewski, F. M. Cohan, Genetics 152, 1459 ter termoften is “ involvedgenomic in islands virulence,” as but the a bet- are implicit in much of the analytical teris now term commonplace is “genomic islandsto speak” as the arehousekeeping implicit in much genes of (Fig. the analytical4). ries (as is likely to be the case), more complex (1999). phenomenon is far from limited to toolkit of population genetics. phenomenonof the “core” isgenome, far from which limited to toolkit of population genetics. statistical models will be needed to identify and 29. E. J. Feil, B. G. Spratt, Annu. Rev. Microbiol. pathogenspathogensencodes (36, 37 ( fundamental36). Although, 37). Although it is func hard it is- hard IdentifyingDistinguishingDistinguishing Mechanisms among among mecha-and mecha- describe the underlying niche structure. Lon- References and Notes 55, 561 (2001). 30. J. Majewski, F. M. Cohan, Genetics 153, 1525 to quantifytions the shared selective by all impact members of im- nismsDelineating of population Species differentiation gitudinal studies that measure the dynamics of 1. J. Handelsman, Microbiol. Mol. Biol. Rev. 68, to quantify the selective impact of im- nisms of population differentiation (1999). 669 (2004). portingportingof any a species given any given gene(s)(and, gene(s)it intoshould ainto newgo a new Ecological factor a ininWhat bacteria bacteria do ultimately we ultimately want comes from comes bacterial down down to to ecological associations over time will also be 31. J. Majewski, P. Zawadzki, P. Pickerill, F. M. Ecological factor a 2. W. P. Hanage, C. Fraser, B. G. Spratt, Philos. background,background,without the saying, occasional the occasional other ability related ability to to testingtestingspecies? the the ability Do ability we of need of different different theoretical models models helpful to determine how transient natural habi- Cohan, C. G. Dowson, J. Bacteriol. 182, 1016 Trans. R. Soc. London Ser. B 361, 1917 — —Fig.Fig. 4. Differences 4. Differences between between core core and and auxiliary auxiliary genes. genes. This schematic schematic (2000). gain a newgainspecies), aadaptation new onto adaptation which in this in is fashion this bolted fashion totoconsistency explain explain highly highly even variable at variable the expense patterns patterns tats are, and thus how likely bottlenecks are to (2006). illustratesillustrates the the relationships relationships between between three three species species in in “ecotype space, space,”” 32. S. T. Cowan, in Microbial Classification, such assuchthe a “auxiliary” new as a metabolic new metabolicor “accessory” capability capability withinwithinof taxonomic and and between between practicality, genetic-ecological genetic-ecological incorpo- result. Finally, whole-genome sequences from 3. M. Achtman, M. Wagner, Nat. Rev. Microbiol. 12th Symposium of the Society for General orgenome, a new mode composed of transmission of genes for ashownshown here here in two in two dimensions, dimensions, and and a a mobile mobile gene gene common common to all all three. three. clustersrating both (Fig. “clonal” 1). It is and still “sexual” unclear entire populations of environmental bacteria 6, 431 (2008). or a new mode of transmission for a clusters (Fig. 1). It is still unclear Microbiology, G. C. Ainsworth, P. H. A. — TheThe areas areas occupied occupied by by the the species species are are shown shown as as solid solid lines lines in red, red, blue, blue, 4. F. M. Cohan, Annu. Rev. Microbiol. 56, 457 pathogenpathogenand— operonsmay bemay that of be enormous may of enormous or may im- im- whetherwhetherpopulations these these into patterns patterns a single are are main- theo main-- will be useful in dissecting the roles of the aux- Sneath, Eds. (Cambridge Univ. Press, andand green. green. The The part part of ofthe the ecological ecological space space where where the the shared shared mobile gene gene (2002). portancenot be present in terms in all of speciation.isolates. It tainedretical byframework? gene flow One or selection, unifying iliary and core genome in ecological differentia- Cambridge, 1962), pp. 433–455. portance in terms of speciation. is selected in each species is shown by a dashed purple line and overlaps tained by gene flow or selection, 5. W. F. Doolittle, R. T. Papke, Genome Biol. 7, Perhaps even more striking is theis selected in each species is shown by a dashed purple line and overlaps and what the effect of population 33. S. K. Sheppard, N. D. McCarthy, D. Falush, Perhapsseemseven likely more that strikingsuch auxilia is the- all three species ranges. Examples of (circular) genomes from each species andtheoretical what conceptthe effect is to of consider population tion. If after this process it emerges that some 116 (2006). all three species ranges. Examples of (circular) genomes from each species M. C. J. Maiden, Science 320, 237 (2008). amountamountry ofgenes variation ofhelp variation to in determine gene in gene content the content structurestructurespecies is.as is. Thethe The arena joint joint distributionwithin distribution which of of model or models are consistently validated for 6. D. Gevers et al., Nat. Rev. Microbiol. 3, 733 withwith and and without without the the purple purple mobile mobile element element are are also also illustrated. illustrated. Note Note that that 34. A. C. Retchless, J. G. Lawrence, Science 317, revealedspecific by ecological multiple genomes properties from geneticindividuals and ecological are similar data enough, can be different study systems, these would inevitably (2005). revealed by multiple genomes from for each species, the locus is not selected for all isolates, and its evolu- genetic and ecological data can be 1093 (2007). the same species, which implies thatfor each species, the locus is not selected for all isolates, and its evolu- used, as described above for Vibrio 7. M. F. Polz, D. E. Hunt, S. P. Preheim, D. M. the sameof the species, organism. which For implies example, that tionary fate is uncoupled from that of each host species, because if one used,or interbreed as described enough, above that for indiVibrio- form a good basis for identifying fundamental 35. K. Vetsigian, N. Goldenfeld, Proc. Natl. Acad. tionary fate is uncoupled from that of each host species, because if one Weinreich, Philos. Trans. R. Soc. London Ser. genea group acquisition of related occursLeptospiril at a- sur- undergoes a selective sweep or goes extinct, the mobile gene may be speciesvidual ( variant13), to genes define compete populations di- levels of clustering, or species. Sci. U.S.A. 102, 7332 (2005). gene acquisition occurs at a sur- undergoes a selective sweep or goes extinct, the mobile gene may be species (13), to define populations B 361, 2009 (2006). prisinglylum has recently high frequency. been hypoth It is- now reintroduced from one of the other species. Examples of such distributed withoutrectly for making reproductive a strong success.theoret- In the foregoing we have emphasized eco- 36. A. Tuanyok et al., BMC Genomics 9, 566 prisingly high frequency. It is now without making a strong theoret- 8. D. B. Rusch et al., PLoS Biol. 5, e77 (2007). commonplace to speak of the “core”reintroduced from one of the other species. Examples of such distributed ical commitment to either of these (2008). esized to adapt to different“ ” loci include drug resistance determinants in pathogens (e.g., b-lactamase Practical advances building on type and metapopulation models, but there are 9. C. Darwin, The Origin of Species (Penguin commonplace to speak of the core loci include drug resistance determinants in pathogens (e.g., b-lactamase ical commitment to either of these 37. U. Dobrindt, B. Hochhut, U. Hentschel, J. genome,areas of an which acid encodesmine drainage fundamental genes) and heavy metal resistance in environmental organisms. These alternatives.this or other One theoretical clear result concepts from others that deserve consideration—notably the Classics, London, 1985). genome, which encodes fundamental alternatives. One clear result from Hacker, Nat. Rev. Microbiol. 2, 414 (2004). functions shared by all members of agenes)genes and may heavy be transferred metal resistance among strains in environmental and species by organisms. conjugative Theseplas- all of the studies discussed here is 10. E. Stackebrandt et al., Int. J. Syst. Evol. system by shuffling of chro- will only come when these are epidemic clonal model (42) and the impact of 38. P. Wilmes, S. L. Simmons, V. J. Denef, J. functionsspecies shared (and, by allit should members go of without a genesmids may or beother transferred mobile elements among(including strains and transducing species by phage). conjugative plas- thatall the of underlying the studies theoretical discussed ques- here is Microbiol. 52, 1043 (2002). mosome segments enriched in developed into explicit models phage epidemics causing classic Lotka-Volterra F. Banfield, FEMS Microbiol. Rev. 33, 109 species (and, it should go without mids or other mobile elements (including transducing phage). that the underlying theoretical ques- 11. R. Facklam, Clin. Microbiol. Rev. 15, 613 saying,noncore other genes related (38, 3 species),9). We onto tionsand concerning model-based species algorithms will not that be boom-bust dynamics (43) illustrated in Fig. (2009). (2002). saying,which other is related bolted thespecies),“auxiliary onto” or “accessory” habitat to which each host strain is adapted. In answered in thetions absence concerning of more speciesdetailed genetic-will not be 39. V. J. Denef et al., Environ. Microbiol. 11, 313 should, however, be aware are tested and refined on a wide 2D—and it is possible, even likely, that more 12. J. C. Arbique et al., J. Clin. Microbiol. 42, whichgenome, is bolted composed the “auxiliary of genes” or and“accessory operons” thathabitatthe case to which of very each mobile host elements strain is— adapted.for exam- In environmentalanswered in the mapping. absence Moreover, of more detailed some guide- genetic- (2008). that changes in core genes may also lead to tic of different levels of ecological specificity, range of data. Alternatively, it may be sensible than one of these mechanisms may be relevant 4686 (2004). — 40. E. Bantinaki et al., Genetics 176, 441 (2007). genome,may composed or may not of be presentgenes andin all operons isolates. It that seems theple, case plasmids of very encoding mobile resistance elements to antibioticsfor exam- or linesenvironmental for the types mapping. of ecological Moreover, studies some that will guide- 13. D. E. Hunt et al., Science 320, 1081 (2008). ecological differentiation, a phenomenon well ranging from— highly conserved core functions to suggest an ad hoc application of principles to any given problem in speciation and cluster 41. R. A. Welch et al., Proc. Natl. Acad. Sci. may orlikely may thatnotbe such present auxiliary in all genes isolates. help toIt seemsdetermine ple,heavy plasmids metals encodingthe ecological resistance specificity to antibiotics deter- or belines most for informative the types of are ecological emerging. studies Most impor- that will 14. See supplementary information in (16) for a documented in experimental studies of bacteria that are essential for growth in all environments to different genera on the basis of their specific formation. Distinguishing among these mecha- U.S.A. 99, 17020 (2002). the specific ecological properties of the organism. mined by these— accessory loci may have no link tant, the ecological data collected must be rel- statistical comparison of the ecotype model likely thatgrowing such in auxiliary structured genes environments help to determine (40). heavyto loci metals that arethe involved ecological with specificity adaptation deter- to a characteristics,be most informative including are the emerging. extent ofMost variation impor- nisms is the bacterial species challenge (Table 42. J. Maynard Smith, N. H. Smith, M. O’Rourke, with an effective neutral model and an the specificFor example, ecological a group properties of related of theLeptospirillum organism. minedto the by sequence these accessory clusters we loci observe may have using no house- link evanttant, theto the ecological niche boundaries data collected of the populations must be rel- B. G. Spratt, Proc. Natl. Acad. Sci. U.S.A. 90, Estimates vary, depending on the genomes specific habitat. Some narrow niche-specific in gene content and recombination. In any case, 1), described in 1991 by John Maynard Smith implicit estimate of N . has recently been hypothesized to adapt to dif- keeping genes (Fig. 4). studied. And if genetic groups do not map ex- e 4384 (1993). For example,that are available, a group ofbut related as littleLeptospirillum as 40% of genes to thegenes sequence may be clusters distributed we observeacross species, using house-being noevant biologist to the would niche deny boundaries the importance of the populations of ecol- as follows: “Ecotypic structure, hitch-hiking, 15. C. Fraser, W. P. Hanage, B. G. Spratt, Science ferent areas of an acid mine drainage system by clusively onto sampling categories (as is likely 43. K. H. Hoffmann et al., FEMS Microbiol. Lett. has recentlymay be been present hypothesized in all sequenced to adapt genomes to dif- of a keepingtransferred genes between (Fig. 4). them by mobile elements. ogystudied. to what And weif observe, genetic but groups it may do not not be mapeasy ex- and localized recombination can explain the 315, 476 (2007). shuffling of chromosome segments enriched in Identifying Mechanisms and to be the case), more complex statistical models 273, 224 (2007). ferent areas of an acid mine drainage system by clusively onto sampling categories (as is likely 16. A. Koeppel et al., Proc. Natl. Acad. Sci. noncorenamed species genes ( 38 (4,139). ). We We may should, consider however, genes be DelineatingThe evolutionary Speciesfate of such genes may hence willto incorporate be needed toit in identify a fashion and describethat is convenient the under- observed patterns of variation. The difficulty, of 44. We thank T. Connor and S. Deeny for shufflingwithin of chromosomea named species segments as being enriched characteris in -Identifyingbe only loosely Mechanisms coupled and with that of any par- forto be taxonomists. the case), more Nonetheless, complex population statistical models ge- course, is that the model is sufficiently flexible U.S.A. 105, 2504 (2008). aware that changes in core genes may also lead to What do we want from bacterial species? Do we lying niche structure. Longitudinal studies that 17. J. Maynard Smith, Proc. R. Soc. London Ser. useful discussions. Supported by University Speciation Delineating Species noncoreecological genes (38 differentiation,, 39). We should, a phenomenon however, be well need theoretical consistency even at the expense measureneticistswill be needed themay dynamics have to identifylittle of choice ecological and describebut to associations tackle thethe under- to explain almost anything. To test the hypoth- B 245, 37 (1991). Research Fellowships from the Royal Society (C.F. and W.P.H.), a program grant from the awaredocumented thatTable changes 1. A proposedin in experimentalcore genes strategy may studies for also developing of lead bacteria to What andof validating do taxonomic we want models practicality, from of bacterial bacterial incorporating species? evolution Do both that we overquestionlying18. H. time Ochman, niche of will defining structure. A. also C. Wilson, be bacterial helpful Longitudinalin Escherichia tospecies determine coli and studiesor, at how the that esis of ecotypic structure, we need to know the 18. H. Ochman, A. C. Wilson, in Escherichia ecological differentiation, a phenomenon well need“ theoretical” “ consistency” even at the expense verymeasureSalmonella least, the populations. typhimurium dynamics :Whether Cellular of ecological and we Molecular are associationsestimat Biology-, distribution of electrophoretic types [i.e., geno- coli and Salmonella typhimurium: Cellular Wellcome Trust (B.G.S.), grants from the growingmight eventually in structured be usedenvironments to classify (40 genetic). diversityclonal dataand andsexual providepopulations a firm foundation into a single for a transient natural habitats are, and thus how likely U.S. Department of Energy Genomes to Life documentedbacterial in species experimental concept. studies of bacteria of taxonomic practicality, incorporating both ingover effectiveF. time C. Neidhart, will population Ed.also (ASM be Press,size helpful from Washington, to neutral determine DC, diver 1987),- how types] in different habitats” (17). and Molecular Biology, F. C. Neidhart, Ed. Estimates vary, depending on the genomes theoretical framework? One unifying theoretical bottleneckspp. 1649– are1654. to result. Finally, whole-genome program (M.F.P. and E.J.A.), and the NSF/ “ ” “ ” sity or choosing an appropriate set of strains to Much research on bacterial species to date (ASM Press, Washington, DC, 1987), pp. growingthat in are structured available, environments but as little as (40 40%). of genes clonalconceptand is tosexual considerpopulations species as the into arena a withinsingle sequencestransient19. J. R. Thompson natural fromet entire habitats al., Appl. populations Environ. are, and Microbiol. thus of environ- how70, 4103 likely 1649–1654. National Institute of Environmental Health 1. Collect samples according to systematic ecological stratification. Focus on longitudinal studies, Estimates vary, depending on the genomes theoretical framework? One unifying theoretical testbottlenecks for(2004). positive are selection to result. at Finally, a locus whole-genomeof interest, has come from studies on pathogens, where the 19. J. R. Thompson et al., Appl. Environ. Sciences Woods Hole Centre for Oceans that are available,geographical but as little studies, as 40%and measurement of genes ofconcept physical and is to chemical consider gradients species affecting as the arena bacterial within speciessequences20. M. Kimura, definitions fromTrends entire Biochem. are implicit populations Sci. 1, N152in (1976).much of environ-of the correct identification of species is crucial for ac- Microbiol. 70, 4103 (2004). and Human Health, the NSF Biological growth. Consider biotic factorswww.sciencemag.org such as the presence of other competingSCIENCE bacteriaVOL or 323 parasitic 6 phage. FEBRUARYanalytical21. R.2009 Milkman, toolkitTrends of Biochem.population Sci. 1 genetics., N152 (1976). 745 curate clinical diagnoses. However, for patho- 20. M. Kimura, Trends Biochem. Sci. 1, N152 Oceanography Program, and the Moore 2. For each isolate, sequence as much as possible and affordable (16S rRNA, MLSA, auxiliary 22. K. C. Atwood, L. K. Schneider, F. J. Ryan, Proc. Natl. (1976). Foundation (M.F.P.). DistinguishingAcad. Sci. U.S.A. 37 among, 146 (1951). mechanisms of pop- gens the identification of the multiple ecological genes, full genomes, etc.).www.sciencemag.org SCIENCE VOL 323 6 FEBRUARYulation23. B. 2009 R. differentiationLevin, Genetics 99, 1in (1981). bacteria ultimately 745 3. Use empirical classification algorithms that use genetic and ecological data to jointly map isolates. comes24. F. M. down Cohan, toE. B. testing Perry, Curr. the Biol. ability17, R373 of (2007).differ- 4. To guide model formulation, use population genetic tests on observed clusters, focusing on ent25. models A. Rambaut toet explain al., Nature highly453, 615 variable (2008). patterns 26. R. A. Fisher, E. B. Ford, Heredity 1, 143 (1947). tests for selection, population structure, and gene flow. within and between genetic-ecological clusters 5. Generate evolutionary models and simulate populations. 27. M. Slatkin, Theor. Popul. Biol. 12, 253 (1977). (28.Fig. J. 1 Majewski,). It is still F. M. unclear Cohan, Genetics whether152 these, 1459 patterns (1999). 6. Test, then reject or adapt, evolutionary models according to agreement between simulations are29. maintained E. J. Feil, B. G. Spratt, by geneAnnu. Rev. flow Microbiol. or selection,55, 561 (2001). and and real populations; if necessary, return to step 1. 30. J. Majewski, F. M. Cohan, Genetics 153, 1525 (1999). what the effect of population structure is. The 7. For successful models, develop model-based methods for interpreting pure genetic data 31. J. Majewski, P. Zawadzki, P. Pickerill, F. M. Cohan, joint distribution of genetic and ecological data (without ecological covariates) and test on new data. C. G. Dowson, J. Bacteriol. 182, 1016 (2000). can32. be S. T. used, Cowan, as in describedMicrobial Classification, above for 12th Vibrio Symposium spe- 8. If one or more validated models emerge, use these to classify genetic data and to develop cies of(1 the3), Societyto define for General populations Microbiology without, G. C. Ainsworth,making bacterial species concepts. P. H. A. Sneath, Eds. (Cambridge Univ. Press, Cambridge, a strong1962), theoretical pp. 433–455. commitment to either of 33. S. K. Sheppard, N. D. McCarthy, D. Falush, M. C. J. 34 mental bacteria will be useful in dissecting the References and Notes Maiden, Science 320, 237 (2008). 35 roles of the auxiliary and core genome in 1. J. Handelsman, Microbiol. Mol. Biol. Rev. 68, 669 (2004). 34. A. C. Retchless, J. G. Lawrence, Science 317, 1093 (2007). ecological differentiation. If after this process it 2. W. P. Hanage, C. Fraser, B. G. Spratt, Philos. Trans. R. 35. K. Vetsigian, N. Goldenfeld, Proc. Natl. Acad. Sci. U.S.A. Soc. London Ser. B 361, 1917 (2006). 102, 7332 (2005). emerges that some model or models are consistently 3. M. Achtman, M. Wagner, Nat. Rev. Microbiol. 6, 431 (2008). 36. A. Tuanyok et al., BMC Genomics 9, 566 (2008). validated for different study systems, these would 4. F. M. Cohan, Annu. Rev. Microbiol. 56, 457 (2002). 37. U. Dobrindt, B. Hochhut, U. Hentschel, J. Hacker, inevitably form a good basis for identifying 5. W. F. Doolittle, R. T. Papke, Genome Biol. 7, 116 (2006). Nat. Rev. Microbiol. 2, 414 (2004). fundamental levels of clustering, or species. 6. D. Gevers et al., Nat. Rev. Microbiol. 3, 733 (2005). 38. P. Wilmes, S. L. Simmons, V. J. Denef, J. F. Banfield, 7. M. F. Polz, D. E. Hunt, S. P. Preheim, D. M. Weinreich, In the foregoing we have emphasized ecotype FEMS Microbiol. Rev. 33, 109 (2009). Philos. Trans. R. Soc. London Ser. B 361, 2009 (2006). 39. V. J. Denef et al., Environ. Microbiol. 11, 313 (2008). and metapopulation models, but there are others 8. D. B. Rusch et al., PLoS Biol. 5, e77 (2007). 40. E. Bantinaki et al., Genetics 176, 441 (2007). that deserve consideration—notably the epidemic 9. C. Darwin, The Origin of Species (Penguin Classics, 41. R. A. Welch et al., Proc. Natl. Acad. Sci. U.S.A. 99, 17020 clonal model (42) and the impact of phage epi- London, 1985). (2002). 10. E. Stackebrandt et al., Int. J. Syst. Evol. Microbiol. 52, 42. J. Maynard Smith, N. H. Smith, M. O’Rourke, B. G. Spratt, demics causing classic Lotka-Volterra boom-bust 1043 (2002). 43 — Proc. Natl. Acad. Sci. U.S.A. 90, 4384 (1993). dynamics ( ) illustrated in Fig. 2D and it is 11. R. Facklam, Clin. Microbiol. Rev. 15, 613 (2002). 43. K. H. Hoffmann et al., FEMS Microbiol. Lett. 273, 224 (2007). possible, even likely, that more than one of these 12. J. C. Arbique et al., J. Clin. Microbiol. 42, 4686 (2004). 44. We thank T. Connor and S. Deeny for useful discussions. mechanisms may be relevant to any given 13. D. E. Hunt et al., Science 320, 1081 (2008). Supported by University Research Fellowships from the problem in speciation and cluster formation. Dis- 14. See supplementary information in (16) for a statistical Royal Society (C.F. and W.P.H.), a program grant from the comparison of the ecotype model with an effective Wellcome Trust (B.G.S.), grants from the U.S. Department tinguishing among these mechanisms is the neutral model and an implicit estimate of Ne. of Energy Genomes to Life program (M.F.P. and E.J.A.), bacterial species challenge (Table 1), described in 15. C. Fraser, W. P. Hanage, B. G. Spratt, Science 315, 476 and the NSF/National Institute of Environmental Health 1991 by John Maynard Smith as follows: “Eco- (2007). Sciences Woods Hole Centre for Oceans and Human typic structure, hitch-hiking, and localized recom- 16. A. Koeppel et al., Proc. Natl. Acad. Sci. U.S.A. 105, 2504 Health, the NSF Biological Oceanography Program, and (2008). the Moore Foundation (M.F.P.). bination can explain the observed patterns of 17. J. Maynard Smith, Proc. R. Soc. London Ser. B 245, 37 variation. The difficulty, of course, is that the (1991). 10.1126/science.1159388 model is sufficiently flexible to explain almost anything. To test the hypothesis of ecotypic structure, we need to know the distribution of REVIEW electrophoretic types [i.e., genotypes] in different habitats” (17). Much research on bacterial species to date has Is Genetic Evolution Predictable? come from studies on pathogens, where the cor- rect identification of species is crucial for accu- David L. Stern1* and Virginie Orgogozo2* rate clinical diagnoses. However, for pathogens the identification of the multiple ecological Ever since the integration of Mendelian genetics into evolutionary biology in the early 20th century, niches within (for example) the nasopharynx or evolutionary geneticists have for the most part treated genes and mutations as generic entities. However, gut is difficult, and studies of the relationships recent observations indicate that all genes are not equal in the eyes of evolution. Evolutionarily relevant between bacterial populations and ecology may mutations tend to accumulate in hotspot genes and at specific positions within genes. Genetic evolution is be more fruitful for some environmental species constrained by gene function, the structure of genetic networks, and population biology. The genetic basis where the categorization of niches is a more of evolution may be predictable to some extent, and further understanding of this predictability requires tractable enterprise. Hopefully, we will soon incorporation of the specific functions and characteristics of genes into evolutionary theory. obtain richer data sets that map bacterial diversity onto ecology and provide a way to distinguish ne hundred and fifty years ago, Charles tion in populations. Both Darwin and Wallace among various models of population differentia- Darwin and Alfred Russell Wallace pro- recognized the importance of heritable variation tion and speciation, including those based on Oposed that biological diversity results to evolutionary theory, but neither man knew the ecotypes or metapopulations. from natural selection acting on heritable varia- true cause of inheritance. Early in the 20th cen-

746 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org REPORTS

rainforest. Originally extending for 1,300,000 km2 (LGM, 21 kybp); (iii) absence of genetic patterns of the putative Bahia refugium (19). The results also REPORTS along the Brazilian coast and reaching into isolation-by-distance in unstable areas, given that agree generally with forest models published Paraguay and Argentina, this biome has been colonization has been too recent to permit restora- previously (14), although the central refugium 27% closer, respectively, to queen sounds than highest attainable position in the colony’s hier- 9. J. A. Thomas, G. W. Elmes, J. C. Wardlaw, Proc. R. Soc. reduced to less than 8% of its range (8). Today’s tion of equilibrium between migration and genetic extends farther south in the frog-based models. to those of workers (Fig. 2B) [mean normalized archy. Selection for accurate acoustical mimicry London Ser. B 265, 1895 (1998). 10. G. W. Elmes, J. C. Wardlaw, K. Schönrogge, J. A. Thomas, fragments harbor one of the largest percentages drift (15); and (iv) strong phylogeographic structure Such differences are expected because the forest euclidean distances between individual butter- may have been stronger in pupae, which lack the Entomol. Exp. Appl. 110, 53 (2004). of endemic species in the world, with many between refugia, reflecting assemblage-wide, long- and its associated species may differ slightly in flies’ and ants’ sounds are as follows: pupa-queen main secretory organs of M. rebeli larvae and of- 11. Materials and methods are available as supporting species and even genera of vertebrates still being term population persistence in isolated areas. their climatic tolerances and realized niches. In 2.47 T (SE) 0.10, pupa-worker 3.03 T 0.15, t = fer only weak rewards to tending workers. material on Science Online. describedREPORTS (8, 9). Our ultimate goal is to pinpoint Distribution models developed under current H. albomarginatus and H. faber, the extension of – 12. P. J. DeVries, R. B. Cocroft, J. A. Thomas, Biol. J. Linn. 3.14, df 87, distancepupa-queens < distancepupa-workers, The young stages of other Maculinea species regions for inventory work and habitat protection climatic2 conditions accurately predict distribu- the predicted São Paulo refugium westward into Soc. 49, 229 (1993). rainforest. Originally extending for 1,300,000 km (LGM, 21 kybp); (iii) absence of genetic patterns of the putative Bahia refugium (19). The results also P < 0.001; larva-queen 2.52 T 0.11, larva-worker make similar pulsed sounds to M. rebeli (12): before we lose a substantial fraction of described tions of each of the target species along the the neighboring Cerrado biome reflects model T – 13. F. Roces, J. Tautz, J. Acoust. Soc. Am. 109, 3080 (2001). along the Brazilian coast and reaching into isolation-by-distance in unstable areas, given that agree generally with forest models published 3.21 0.12, t = 4.32, df 237, distancelarva-queens < All differ substantially from those of other studied 14. R. Hickling, R. L. Brown, J. Acoust. Soc. Am. 108, 1920 and undocumentedParaguay diversity. and Argentina, The approach this biome dif- has beenAtlanticcolonization rainforest has domain been too [area-under-the-curve recent to permit restora- previouslyoverprediction (14), although (fig. S2) the (14 central). refugium distancelarva-workers, P < 0.001]. The distributions Lycaenidae, most of which are commensals or (2000). fers from previousreduced methods to less than by directly 8% of its modeling range (8). Today(AUC)’s tion values of equilibrium (16) 0.968, between 0.989, migration and and 0.994;genetic extendsModels farther of south habitat in the stability frog-based through models. fluctuating in Fig. 2 also satisfy the concept that the perfect mutualists or have no known relationship with ants 15. H. Markl, B. Hölldobler, Behav. Ecol. Sociobiol. 4, 183 historical processes,fragments asopposed harbor one to of observed the largest bio- percentagesmaximumdrift (15 Kappa); and (iv) (17 strong) 0.81, phylogeographic 0.925, and structure0.94 in Suchclimates differences correctly are expected predict because patterns the of forest phylogeog- mimic should have maximal overlap with queen (12, 23–27). None of the latter mimics the acous- (1978). of endemic species in the world, with many between refugia, reflecting assemblage-wide, long- and its associated species may differ slightly in 16. H. Markl, Z. Vgl. Physiol. 60, 103 (1968). diversity patterns (10), with the aim of informing H. albomarginatus, H. faber, and H. semilineatus, raphy in the Brazilian Atlantic rainforest (Fig. 2 acoustics and minimal overlap with those of tics of associated ants in obvious ways, although species and even genera of vertebrates still being term population persistence in isolated areas. their climatic tolerances and realized niches. In 17. H. Markl, Science 149, 1392 (1965). conservation. described (8, 9). Our ultimate goal is to pinpointrespectivelyDistribution (fig. S2)]. models Stability developed maps, under depicting current H.and albomarginatus figs. S3 to S5).and H. In faber all, species, the extension high of levels of workers. the sound of one strongly mutualistic species attracts 18. D. A. Grasso, A. Mori, F. Le Moli, M. Giovannotti, We use molecularregions for genetic inventory data work from and multiple, habitat protectionthe intersectionclimatic conditions of distribution accurately models predict for distribu- each thedivergence predicted São and Paulo population refugium structure westward are into observed – A. Fanfani, Ital. J. Zool. 65, 167 (1998). Playing recordings of Maculinea pupal calls workers (23 26). Thus, the use of acoustics to sig- largely codistributedbefore we species lose a substantial to test whether fraction spa- of describedtaxon undertions of current, each of 6 the kybp, target and species 21 kybp along cli- the theacross neighboring refugia Cerrado (Tamura-Nei biome reflects corrected model distances 19. T. C. Scott-Phillips, J. Evol. Biol. 21, 387 (2008). to the same naïve cultures of Myrmica schencki nal superior status to ants is unlikely to be a basal and undocumented diversity. The approach dif- Atlantic rainforest domain [area-under-the-curve overprediction (fig. S2) (14). 20. K. G. Schurian, K. Fiedler, Nachr. Entomol. Vereins Apollo tial modeling of species-specific Late Quaternary mates, predict for all species a large central re- (20): 4 to 7% between Bahia and Pernambuco workers resulted in enhanced benevolent responses trait in the Lycaenidae, although we might expect 14, 339 (1994). refugia shedsfers light from on previous historical methods processes by directly and modelingfugium(AUC) throughout values the (16) Late 0.968, Quaternary 0.989, and (“Bahia 0.994; refugia,Models 1% of habitat between stability the through nearby fluctuating Bahia and São similar to those elicited by queen ant sounds. We it in Phengaris, the sister genus to Maculinea. 21. C. J. Hill, J. Aust. Entomol. Soc. 32, 283 (1993). hence improveshistorical prediction processes, of genetic as opposed endemism to observedrefugium bio- maximum”) (Fig. Kappa 2). A(17) second, 0.81, 0.925, much and smaller 0.94 in climatesPaulo refugiacorrectly in predictH. faber patterns). Similarly, of phylogeog- in all taxa found no significant differences toward Maculinea Beyond the Lycaenidae, ~10,000 species of 22. D. R. Nash, T. D. Als, R. Maile, G. R. Jones, J. J. Boomsma, diversity patterns (10), with the aim of informing H. albomarginatus, H. faber, and H. semilineatus, raphy in the Brazilian Atlantic rainforest (Fig. 2 and diversity inconservation. tropical Brazil (11). We focus on refugiumrespectively is predicted (fig. S2)]. in Stability the northeasternmost maps, depicting andthere figs. are S3to multiple, S5). In all divergent species, high clades levels within of the pupal and Myrmica queen calls in any of the four ant social parasites may exist (5), particularly Science 319, 88 (2008). “ ” 23. P. J. DeVries, Am. Mus. Nov. 3025, 1 (1991). three common speciesWe use molecular of tree genetic frogs data that from are multiple,portionthe of intersection the forest of ( distributionPernambuco models refugium for each). divergenceBahia region, and population agreeing structure with are model-based observed pre- behaviors scored, and pupal calls elicited six times among other Lepidoptera, Coleoptera, Diptera, 24. K. Fiedler, B. Hölldobler, P. Seufert, Experientia 52, 14 widely distributedlargely along codistributed the Brazilian species to Atlantic test whetherIn spa-H.taxon faber, under a third, current, southeastern 6 kybp, and refugium 21 kybp cli- of acrossdictions refugia of a (Tamura-Nei large refugium corrected in distancesthis area. In H. more instances of royal on-guard attendance than and inquiline ants (1, 6). If acoustics plays the (1996). forest: Hypsiboastialalbomarginatus modeling of species-specific, H. semilineatus Late Quaternary, intermediatemates, predict size is for also allspecies predicted a large (“São central Paulo re- (20faber): 4, to divergent 7% between clades Bahia are and also Pernambuco represented in the occurred when worker sounds were played (Fig. 3 role that we suggest in reinforcing an ant’s hi- 25. M. A. Travassos, N. E. Pierce, Anim. Behav. 60, 13 and H. faber.refugia Given sheds their light life on historyhistorical traits, processesrefugium and fugium”). This throughout is not surprising, the Late Quaternary given that (“Bahia this refugia,São Paulo 1% between region, the matching nearby Bahiapredictions and São of a mid- ” and table S1) (P < 0.001). Recordings of M. rebeli erarchical status, it seems likely that this cue has (2000). amphibians arehence useful improves indicators prediction of of environ- genetic endemismspeciesrefugium occupies) (Fig.a broader 2). A environmental second, much smaller niche. Paulosized refugia refugium in H. in faber this). Similarly, area. All in taxa all taxa show low larvae induced lower worker responses and, de- evolved in other social parasites to infiltrate and 26. N. E. Pierce et al., Annu. Rev. Entomol. 47, 733 (2002). and diversity in tropical Brazil (11). We focus on refugium is predicted in the northeasternmost there are multiple, divergent clades within the 27. J. C. Downey, A. C. Allyn, Bull. Mus. Entomol. 14, mental changesthree through common time species (12). of Whereas tree frogsH. thatIn are contrastportion to of the the central forest (“ andPernambuco northern refugium regions,”). Bahiagenetic region, diversity agreeing across with the model-based southernmost pre- range of spite eliciting 2.3 times more on-guard attendances exploit their societies. 1 (1973). albomarginatuswidelyand H. distributed semilineatus alongoccur the Brazilian in low AtlanticpopulationsIn H. faber south, a of third, the southeastern Bahia or refugiumSão Paulo of dictionsthe forest, of a an large area refugium predicted in tothis be area. less In stableH. by the than worker calls, did not differ significantly from 28. We thank N. Elfferich and P. J. DeVries for introducing and mid altitudesforest: andHypsiboas are mostly albomarginatus restricted, H. to semilineatus the refugia, intermediate appear much size less is also stable, predicted despite (“São the Paulomore faberpalaeomodels., divergent clades Furthermore, are also represented mitochondrial in the DNA responses toward worker sounds (Fig. 3 and table References and Notes us to ant-butterfly acoustics; G. W. Elmes, J. C. Wardlaw, evergreen or semideciduousand H. faber. Given components their life of history the traits,extensiverefugium (preclearing)”). This is not range surprising, of the given forest that this in São(mtDNA) Paulo region, lineages matching found predictions in this region of a mid- are shared 1. B. Hölldobler, E. O. Wilson, The Ants (Springer, Berlin, V. La Morgia, M. B. Bonsall, and referees for S1). We did not play Maculinea calls to queen Atlantic Forestamphibians in eastern are Brazil, usefulH. indicators faber has of a environ-southernspecies and occupies southeastern a broader Brazil. environmental We hypothe- niche. sizedwith refugium adjacent in refugia this area. (one All in taxaH. show albomarginatus low 1990). comments and advice; and M. Charles for designing mental changes through time (12). Whereas H. In contrast to the central and northern regions, genetic diversity across the southernmost range of ants but predict that they would provoke rivalry the acoustical equipment. broader altitudinal range and also inhabits mixed size that these areas received a significant influx and H. semilineatus, two in H. faber). 2. A. Lenoir, P. D'Ettorre, C. Errard, A. Hefetz, Annu. Rev. albomarginatus and H. semilineatus occur in low populations south of the Bahia or São Paulo the forest, an area predicted to be less stable by the similar to that observed when live Maculinea Entomol. 46, 573 (2001). Supporting Online Material and deciduousand areas, mid occupying altitudes and interior are mostly and restricted coast- toof the migrantsrefugia fromappearadjacent, much less stable, large despite refugial the morepop- palaeomodels.Metrics of Furthermore, genetic diversity mitochondrial confirm DNA the above 3. B. Hölldobler, J. Comp. Physiol. 184, 129 (1999). pupae were artificially enclosed with Myrmica www.sciencemag.org/cgi/content/full/323/5915/782/DC1 al sites in the Atlantic Forest south to Paraguay and ulations after the LGM. These palaeomodel re- patterns (Table 1). In H. albomarginatus and H. 4. E. Ruiz, M. H. Martínez, M. Martínez, J. M. Hernández, evergreen or semideciduous components of the extensive (preclearing) range of the forest in (mtDNA) lineages found in this region are shared schencki queens (11) (fig. S1). Materials and Methods Ann. Soc. Entomol. Fr. 42, 99 (2006). Argentina (figs.Atlantic S1 and Forest S2) ( in13 eastern). The comparativeBrazil, H. faber hassults a aresouthern congruent and southeastern with the fossil Brazil. pollen We hypothe- record, withsemilineatus adjacent refugia, genetic (one diversity in H. albomarginatus (21) is an order of SOM Text We suggest that regional host specificity in 5. J. A. Thomas, J. Settele, Nature 432, 283 (2004). phylogeographicbroader approach altitudinal is arange powerful and also test inhabits of mixedwhichsize documents that these areasa replacement received a significant of forests influx by andmagnitudeH. semilineatus larger, twoin the in H. central faber). (Bahia) refugium Maculinea populations is mediated first through 6. J. A. Thomas, K. Schönrogge, G. W. Elmes, in Insect Figs. S1 and S2 and deciduous areas, occupying interior and coast- of migrants from adjacent, large refugial pop- Metrics of genetic diversity confirm the above Table S1 assemblage-scale responses to former environ- grasslands in the southern Atlantic forest during relative to the less stable (southern) portion of the chemical mimicry (6, 22); but once the intruder Evolutionary Ecology, M. D. Fellowes, G. Holloway, J. Rolff, mental changeal and sites thereby in the Atlantic provides Forest a south means to Paraguay for the and LGMulations (14, after18) and the LGM. suggests These the palaeomodel occurrence re- of patternsforest. (Table Diversity 1). In ofH.H. albomarginatus faber in thisand southernH. area Eds. (CABI, Wallingford, UK, 2005), pp. 479–518. References Argentina (figs. S1 and S2) (13). The comparative sults are congruent with the fossil pollen record, semilineatus, genetic diversity (21) is an order of is admitted and accepted as a member of a host Audio S1 to S4 7. K. Schönrogge et al., J. Chem. Ecol. 30, 91 (2004). critical assessmentphylogeographic of the scenarios approach produced is a powerful by testsmall of forestwhich refugia documents in the a replacement southernmost of forests range by of magnitudeis higher larger than in the the othercentral species (Bahia) refugium because of the society, it mimics adult ant acoustics (particu- 8. T. Akino, J. J. Knapp, J. A. Thomas, G. W. Elmes, Proc. R. 22 July 2008; accepted 28 November 2008 modeling of speciesassemblage-scale’ distributions responses under to palaeo- former environ- grasslands in the southern Atlantic forest during relative to the less stable (southern) portion of the larly queens) to advance its seniority towardREPORTS the Soc. London Ser. B 266, 1419 (1999). 10.1126/science.1163583 climates (7). mental change and thereby provides a means for the LGM (14, 18) and suggests the occurrence of forest. Diversity of H. faber in this southern area is higher than the other species because of the The palaeomodelingcritical assessment method of the intersects scenarios pre- produced by small forest refugia inSpecies the southernmost rangeMap of of predicted Current & modeling of species’ distributions under palaeo- 27% closer, respectively, to queen sounds than highest attainable position in the colony’s hier- 9. J. A. Thomas, G. W. Elmes, J. C. Wardlaw, Proc. R. Soc. dicted species’ distributions under current condi- occurrence stable areas historical climate climates (7). to those of workers (Fig. 2B) [mean normalized archy. Selection for accurate acoustical mimicry London Ser. B 265, 1895 (1998). tropicalin persistence biomes and have hence usefully diversity, employed they post lead hoc to niches.tions andIn climaticH. albomarginatus extremes of the and Late H. Quaternaryfaber, the data (putative refugia) data 10. G. W. Elmes, J. C. Wardlaw, K. Schönrogge, J. A. Thomas, The palaeomodeling method intersects pre- Species Map of predicted Current &

palaeoclimatephylogeographic models predictions of species for both and habitatsindividual to extension of the predicted São Paulo refugium Diversity euclidean distances between individual butter- may have been stronger in pupae, which lack the Stability Predicts Genetic Diversity in (6000 years beforedicted present, species or’ distributions 6 kybp, and under 21 kybp) current condi- Modeling occurrence stable areas historical climate Entomol. Exp. Appl. 110, 53 (2004). Distribution providespecies insights and assemblages about processes (codistributed shaping genetic taxa; westward into the neighboring Cerrado biome data (putative refugia) data flies’ and ants’ sounds are as follows: pupa-queen main secretory organs of M. rebeli larvae and of- 11. Materials and methods are available as supporting to predict areastions of and stability climatic extremes (regions ofin the whichLate Quaternary

andFig. species1). Field diversity sampling (5, 7 ).is Buildingdriven by on the them, model we reflects model overprediction (fig. S2) (14). Diversity T T material on Science Online. species are predicted(6000 years to before occupy present, irrespective or 6 kybp, and of 21 kybp) Modeling 2.47 (SE) 0.10, pupa-worker 3.03 0.15, t = fer only weak rewards to tending workers. the Brazilian Atlantic Forest Hotspot Distribution Spatially explicit hypotheses Re: 12. P. J. DeVries, R. B. Cocroft, J. A. Thomas, Biol. J. Linn. firstpredictions map theto cover palaeodistribution both predicted ofrefugia endemic and timeModels period) of and tohabitat predict unstable stability areas areas of through stability(7, 14). fluctuat (regions Because- in which –3.14, df 87, distancepupa-queens < distancepupa-workers, The young stages of other Maculinea species species are predicted to occupy irrespective of T Soc. 49, 229 (1993). 1 2 3 speciesunstable to (recently identify colonized) temporally areas, stable particularly (refugial) ingthe stabilityclimates mapscorrectly raisepredict specific patterns hypotheses of phylo about- Spatially explicit hypotheses Re: P < 0.001; larva-queen 2.52 0.11, larva-worker make similar pulsed sounds to M. rebeli (12): Ana13. F. Carolina Roces, J. Tautz, Carnaval,J. Acoust.* Soc.Michael Am. 109 J., 3080 Hickerson, (2001). Célio F. B. Haddad, time period) and unstable areas (7, 14). Because 3.21 T 0.12, t = –4.32, df 237, distance < All differ substantially from those of other studied 4 1 andemphasizing unstable previously (recently colonized)undersampled regions areas. for If geographyregional differences in the Brazilian in persistence Atlantic and rainforest hence di- larva-queens Miguel14. R. Hickling, T. Rodrigues, R. L. Brown, CraigJ. Acoust. Moritz Soc. Am. 108, 1920 the stability maps raise specific hypotheses about spatial-temporal patterns species occurrence, which are then validated with versity, they lead to phylogeographic predictions distribution of congruence distancelarva-workers, P < 0.001]. The distributions Lycaenidae, most of which are commensals or (2000). the approach correctly predicts current patterns (Fig. 2 and figs.regional S3 to differences S5). In inall persistence species, andhigh hence di- Hypothesis of colonization Formulation genetic diversity spatial-temporal patterns across taxa in Fig. 2 also satisfy the concept that the perfect mutualists or have no known relationship with ants 15. H. Markl, B. Hölldobler, Behav. Ecol. Sociobiol. 4, 183 multispeciesof biodiversity molecular at the data. regional Going scale, beyond species the levelsfor both of individualdivergenceversity, species and they population lead and to phylogeographicassemblages structure (co-are predictions distribution of and/or vicariance congruence

Biodiversity hotspots, representing regions with high species endemism and conservation threat, Hypothesis of colonization (1978). Formulation genetic diversity across taxa mimic should have maximal overlap with queen (12, 23–27). None of the latter mimics the acous- traditionalshould consistently species-by-species show (i) approach,higher genetic the mo- di- observeddistributed across taxa;for Fig.refugia both 1). individual Field(Tamura-Nei sampling species and corrected is assemblages driven (co- and/or vicariance have16. H. been Markl, mappedZ. Vgl. Physiol. globally.60, 103 Yet, (1968). biodiversity distribution data from within hotspots are too sparse acoustics and minimal overlap with those of tics of associated ants in obvious ways, although lecularversity analyseswithin and contrast among thepopulations fit of assemblage- in refugia distancesby the model (20): predictionsdistributed 4 to 7% taxa;between to cover Fig. 1). Bahia both Field predictedand sampling Per- is driven for17. effective H. Markl, Science conservation149, 1392 in (1965). the face of rapid environmental change. Using frogs as indicators, by the model predictions to cover both predicted levelrelative data to tounstable the spatially areas, because explicit demographicof long-term nambucorefugia and refugia, unstable 1% (recentlybetween the colonized) nearby Bahia areas, Sampling across predicted stable and unstable areas workers. the sound of one strongly mutualistic species attracts ecological18. D. A. Grasso, niche A. modelsMori, F. Le under Moli, M. paleoclimates, Giovannotti, and simultaneous Bayesian analyses of multispecies refugia and unstable (recently colonized) areas, Sampling across predicted stable and unstable areas – A. Fanfani, Ital. J. Zool. 65, 167 (1998). scenarios suggested by the climate-based models. particularly emphasizing previously undersam- Playing recordings of Maculinea pupal calls workers (23 26). Thus, the use of acoustics to sig- molecular data, we compare alternative hypotheses of assemblage-scale response to late persistence and population structure; (ii) genetic and São Paulo particularlyrefugia in emphasizingH. faber). Similarly, previously in undersam- 19. T. C. Scott-Phillips, J. Evol. Biol. 21, 387 (2008). ’ to the same naïve cultures of Myrmica schencki nal superior status to ants is unlikely to be a basal signatureWe apply of thispopulation approach expansion to one of inthe unstable world s allpled taxa areas. there If are thepled multiple, approach areas. If thedivergent correctly approach clades predicts correctly with cur- predicts- cur- Quaternary20. K. G. Schurian, climate K. Fiedler, change.Nachr. This Entomol. reveals Vereins a Apollohotspot within the Brazilian Atlantic forest hotspot. workers resulted in enhanced benevolent responses trait in the Lycaenidae, although we might expect mostareas, species-rich, reflecting yetmultispecies notoriously colonization endangered from and inrent the patterns Bahia region, ofrent biodiversity patterns agreeing of biodiversity at with the regional model-based at the scale, regional scale, Model We show14, 339 that (1994). the southern Atlantic forest was climatically unstable relative to the central region, Model Genetic tests tests of of Assemblage-scaleAssemblage-scale

understudied ecosystems: the Brazilian Atlantic species should consistently show (i) higher genetic Validation species should consistently show (i) higher genetic Validation Descriptive similar to those elicited by queen ant sounds. We it in Phengaris, the sister genus to Maculinea. 21. C. J. Hill, J. Aust. Entomol. Soc. 32, 283 (1993). adjacent refugial regions after the Last Glacial predictions of a large refugium in this area. In Descriptive stability/expansion, hypothesis testing which served as a large climatic refugium for neotropical species in the late Pleistocene. This sets phylogeography stability/expansion, hypothesis testing diversity withindiversity and among within populations and among populations in refugia in refugia phylogeography divergence times (HABC) found no significant differences toward Maculinea Beyond the Lycaenidae, ~10,000 species of new22. D. priorities R. Nash, T. for D. Als, conservation R. Maile, G. R.in Jones, Brazil J. J. Boomsma, and establishes a validated approach to biodiversity Maximum (LGM, 21 kybp); (iii) absence of H. faber, divergent clades are also represented divergence times (HABC) 1 relative to unstable areas, because of long-term per- pupal and Myrmica queen calls in any of the four ant social parasites may exist (5), particularly Science 319, 88 (2008). geneticMuseum patterns of Vertebrate of isolation-by-distance Zoology, University of California, in un- inrelative the São to unstablePaulo region, areas, becausematching of predictions long-term per- of prediction23. P. J. DeVries, in otherAm. Mus. understudied, Nov. 3025, 1 (1991).species-rich regions. – 2 sistence and population structure; (ii) genetic sig- behaviors scored, and pupal calls elicited six times among other Lepidoptera, Coleoptera, Diptera, Berkeley, CA 94720 3160, USA. Biology Department, sistence and population structure; (ii) genetic sig- 24. K. Fiedler, B. Hölldobler, P. Seufert, Experientia 52, 14 Queensstable College,areas, givenCity University that colonization of New York, Flushing, has been NY a mid-sized refugiumnature of in population this area. expansion All taxa in show unstable areas, Fig. 1. Proposed method of biodiversity prediction. Three stages are involved: biodiversity dis- nature of population expansion in unstable areas, more instances of royal on-guard attendance than and inquiline ants (1, 6). If acoustics plays the (1996). too recent to3 permit restoration of equilibrium low genetic diversityreflecting multispecies across the colonization southernmost from adjacentFig. 1.tributionProposed modeling method (top), of model-based biodiversity hypothesis prediction. formulation Three stages (middle), are hypothesis involved: testing biodiversity and dis- ate Quaternary climate fluctuations helped refugia models have been dismissed because of 11367, USA. Departamento de Zoologia, Instituto de occurred when worker sounds were played (Fig. 3 role that we suggest in reinforcing an ant’s hi- 25. M. A.ate Travassos, Quaternary N. E. Pierce, climateAnim. Behav. fluctuations60, 13 being described (8, 9). Our ultimate goal is to Biociências,between migration UNESP, Rio and Claro, genetic SP 3526-4100, drift (15); Brazil. and rangereflecting of the multispecies forest,refugial an regions colonizationarea afterpredicted the from Last to Glacial adjacentbe less Maximum tributionmodel modeling validation (top), (bottom). model-based hypothesis formulation (middle), hypothesis testing and (2000).to shape present-day diversity in temper- conflicting evidence (2, 3) or circularity in iden- 4 refugial regions after the Last Glacial Maximum model validation (bottom). and table S1) (P < 0.001). Recordings of M. rebeli erarchical status, it seems likely that this cue has helped to shape present-day diversity in pinpoint regions for inventory work and habitat (iv)Departamento strong phylogeographic de Zoologia, Instituto structure de Biociências, between stable by the palaeomodels. Furthermore, mito- 26. N.ate E. Pierce andet boreal al., Annu. systems Rev. Entomol. (1),47 providing, 733 (2002). a tifying putative refugia (4), but historical pro- Universidade de São Paulo, SP 055008-090, Brazil. larvae induced lower worker responses and, de- evolved in other social parasites to infiltrate and L temperate and boreal systems (1), pro- protection before we lose a substantial fraction refugia, reflecting assemblage-wide, long-term chondrial 786DNA (mtDNA) lineages found in this6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org general27. J. C. Downey,context A. for C. Allyn, understandingBull. Mus. Entomol. current14, pat- cesses must be invoked to explain regions of spite eliciting 2.3 times more on-guard attendances exploit their societies. L1 (1973). *To whom correspondence should be addressed. E-mail: viding a general context for understanding cur- of described and undocumented diversity. The population persistence in isolated areas. region are shared with adjacent refugia (one in scale, long-term persistence of populations in and E). Using Bayes factor (25), we also detect terns of endemism. In the tropics, Pleistocene high endemism (5, 6). Recent studies from sub- [email protected] 786 6 FEBRUARY 2009 VOL 323 SCIENCE www.sciencemag.org than worker calls, did not differ significantly from rent28. Wepatterns thank N. of Elfferich endemism. and P. J.In DeVries the tropics, for introducing Pleis- approach differs from previous methods by di- Distribution models developed under current H. albomarginatus and H. semilineatus, two in isolated refugial areas, as opposed to post-LGM evidence for stability in both areas under the no- responses toward worker sounds (Fig. 3 and table References and Notes us to ant-butterfly acoustics; G. W. Elmes, J. C. Wardlaw, 1. B. Hölldobler, E. O. Wilson, The Ants (Springer, Berlin, toceneV. La refugia Morgia, M.models B. Bonsall, have and been referees dismissed for be- rectly modeling historical processes, as opposed climatic conditions accurately predict distribu- H. faber). colonization of refugial regions. migration model [B(Z2 = 0, Z2 > 0) = 4.89], as S1). We did not play Maculinea calls to queen 1990). causecommentsof conflicting and advice; evidence and M. Charles (2, 3) for or designingwww.sciencemag.orgcircularity to observed biodiversitySCIENCE patternsVOL 323 (10), with 6 FEBRUARY the tions 2009 of each of the target species along the At-785 Metrics of genetic diversity confirm the To test for assemblage-wide colonization well as under a postisolation migration model ants but predict that they would provoke rivalry 2. A. Lenoir, P. D'Ettorre, C. Errard, A. Hefetz, Annu. Rev. the acoustical equipment. in identifying putative refugia (4), but historical aim of informing conservation. lantic rainforest domain [area-under-the-curve above patterns (Table 1). In H. albomarginatus of predicted unstable areas, we group mtDNA [B(Z2 = 0, Z2 > 0) = 4.84]. similar to that observed when live Maculinea Entomol. 46, 573 (2001). processesSupporting must Online be Material invoked to explain regions We use molecular genetic data from multiple, (AUC) values (16) 0.968, 0.989, and 0.994; and H. semilineatus, genetic diversity (21) is an sequences from the southernmost refugial sites Relative to nuclear loci, mtDNA data are 3. B. Hölldobler, J. Comp. Physiol. 184, 129 (1999). pupae were artificially enclosed with Myrmica www.sciencemag.org/cgi/content/full/323/5915/782/DC1 4. E. Ruiz, M. H. Martínez, M. Martínez, J. M. Hernández, of high endemism (5, 6). Recent studies from largely codistributed species to test whether spa- maximum Kappa (17) 0.81,0.925, and 0.94 in H. order of magnitude larger in the central (Bahia) [population 1 (Fig. 3A)] and from localities in more variable and readily collected and often schencki queens (11) (fig. S1). Materials and Methods Ann. Soc. Entomol. Fr. 42, 99 (2006). subtropical biomes have usefully employed post tial modeling of species-specific Late Quaterna- albomarginatus, H. faber, and H. semilineatus, refugium relative to the less stable (southern) unstable areas south of the refugium [population provide key insights into biological response We suggest that regional host specificity in SOM Text 5. J. A. Thomas, J. Settele, Nature 432, 283 (2004). hoc palaeoclimate models of species and habi- ry refugia sheds light on historical processes and respectively (fig. S2)]. Stability maps, depict- portion of the forest. Diversity of H. faber in 2 (Fig. 3A)] to contrast two alternative historical to environmental modification (1). Although Maculinea populations is mediated first through 6. J. A. Thomas, K. Schönrogge, G. W. Elmes, in Insect Figs. S1 and S2 tatsTable to S1 provide insights about processes shaping hence improves prediction of genetic endemism ing the intersection of distribution models for this southern area is higher than the other spe- models across the three codistributed species, single-locus inference can be imprecise in the chemical mimicry (6, 22); but once the intruder Evolutionary Ecology, M. D. Fellowes, G. Holloway, J. Rolff, Eds. (CABI, Wallingford, UK, 2005), pp. 479–518. geneticReferences and species diversity (5, 7). Building and diversity in tropical Brazil (11). We focus each taxon under current, 6 kybp, and 21 kybp cies because of the presence of two lineages that while allowing the taxon-specific demographic face of coalescent variance and the possibility of is admitted and accepted as a member of a host Audio S1 to S4 7. K. Schönrogge et al., J. Chem. Ecol. 30, 91 (2004). on them, we first map the palaeodistribution of on three common species of tree frogs that are climates, predict for all species a large central co-occur in the adjacent refugia. In all species, parameters to vary. In H , the long-term persis- selection (26), our method benefits from a mul- society, it mimics adult ant acoustics (particu- 1 8. T. Akino, J. J. Knapp, J. A. Thomas, G. W. Elmes, Proc. R. endemic22 July 2008; species accepted to 28 identifyNovember 2008temporally stable widely distributed along the Brazilian Atlantic refugium throughout the Late Quaternary (“Ba- average net nucleotide differences across locali- tence model, two contemporary populations split titaxon approach, while explicitly accounting larly queens) to advance its seniority toward the Soc. London Ser. B 266, 1419 (1999). 10.1126/science.1163583 (refugial) and unstable (recently colonized) forest: Hypsiboas albomarginatus, H. semi- hia refugium”) (Fig. 2). A second, much smaller ties (22) reflects high geographic structure with- from an ancestral population prior to the LGM for the stochasticity of a single-locus coalescent regions for species occurrence, which are then lineatus, and H. faber. Given their life history refugium is predicted in the northeasternmost in refugia (2.6 to 6.2% divergence). In contrast, (120,000 to 1.2 million years before present, or across taxa. Combining data sets from several

tropicalvalidated biomes with multispecies have usefully molecular employed data. post Go hoc- traits, amphibians are useful indicators of envi- portion of the forest (“Pernambuco refugium”). sites located outside (south of) the refugia are Mybp, Fig. 3A). In H2, the recent colonization codistributed groups into a single hierarchical

Stability Predicts Genetic Diversity in palaeoclimateing beyond the models traditional of species species-by-species and habitats to ronmental changes through time (12). Whereas In H. faber, a third, southeastern refugium of in- genetically more similar to each other, although model, population 2 is modeled as being colo- Bayesian analysis allowed us to estimate con-

provideapproach, insights the molecular about processes analyses shaping contrast genetic the H. albomarginatus and H. semilineatus occur in termediate size is alsopredicted (“São Paulo ref- to a lesser extent in H. faber (0.1 to 1.6%). Sig- nized from refugial population 1 subsequent to gruence across species,while borrowing strength

andfit speciesof assemblage-level diversity (5, 7data). Building to the spatially on them, ex we- low and mid altitudes and are mostly restricted ugium”). This is not surprising, given that this natures of population expansion (23) are found the LGM (0 to 20 kybp; Fig. 3A). The results from the full comparative phylogeographicsam- the Brazilian Atlantic Forest Hotspot plicitfirst map demographic the palaeodistribution scenarios suggested of endemicby the to the evergreen or semideciduous components species occupies a broader environmental niche. in the unstable area for H. albomarginatus and indicate that all three species colonized the ple (24). This can translate into higher analytical

climate-basedspecies to identify models. temporally stable (refugial) of the Atlantic Forest in eastern Brazil, H. faber In contrast to the central and northern regions, H. faber, as well as in the Bahia refugium area southern (unstable) areas after the LGM (Z2 = 3, power and be more informative than qualitative Ana Carolina Carnaval,1* Michael J. Hickerson,2 Célio F. B. Haddad,3 andWe unstable apply this (recently approach colonized) to one of regionsthe world’s for has a broader altitudinal range and also inhabits populations south of the Bahia or São Paulo for H. faber and H. semilineatus. The lack of the number of species evolved under H ), even comparisons of species-specific analyses. By Miguel T. Rodrigues,4 Craig Moritz1 2 mostspecies species-rich, occurrence, whichyet notoriously are then validated endangered with mixed and deciduous areas, occupying interior refugia appear much less stable, despite the signature of population expansion in the south- when allowing for postisolation migration (Fig. capturing the historical signal that emerges from andmultispecies understudied molecular ecosystems: data. Going the beyond Brazilian the and coastal sites in the Atlantic Forest south to more extensive (preclearing) range of the for- ernmost localities of H. semilineatus may reflect 3, B and C). When Bayes factor is used (25), larger, combined multispecies molecular data Biodiversity hotspots, representing regions with high species endemism and conservation threat, Atlantictraditional rainforest. species-by-species Originally approach, extending the mo- for Paraguay and Argentina (figs. S1 and S2) (13). est in southern and southeastern Brazil. We hy- low statistical power because of the exception- there is strong support for recent colonization in sets, HABC will offer the possibility of looking have been mapped globally. Yet, biodiversity distribution data from within hotspots are too sparse 1,300,000lecular analyses km2 along contrast the the Brazilian fit of assemblage- coast and The comparative phylogeographic approach is pothesize that these areas received a significant ally low levels of diversity observed in this spe- all three species (Z = 3) under the no-migration at patterns of historical community assembly in for effective conservation in the face of rapid environmental change. Using frogs as indicators, 2 reachinglevel data into to Paraguay the spatially and Argentina, explicit demographic this biome a powerful test of assemblage-scale responses influx of migrants from adjacent, large refugial cies. As predicted, isolation by distance is not model [B(Z = 3, Z < 3) = 35.16], and moderate codistributed nonmodel organisms for which ecological niche models under paleoclimates, and simultaneous Bayesian analyses of multispecies 2 2 hasscenarios been reduced suggested to byless the than climate-based 8% of its range models. (8). to former environmental change and thereby populations after the LGM. These palaeomodel observed in unstable regions, but is detected support under a postisolation migration model barcode-type DNA sequence information (e.g., molecular data, we compare alternative hypotheses of assemblage-scale response to late Today’sWe apply fragments this approach harbor one to oneof the of largest the world per’s- provides a means for critical assessment of the results are congruent with the fossil pollen re- within refugial areas for H. albomarginatus and [B(Z = 3, Z < 3) = 5.70]. mtDNA data) can be feasibly collected. Quaternary climate change. This reveals a hotspot within the Brazilian Atlantic forest hotspot. 2 2 centagesmost species-rich, of endemic yet notoriouslyspecies in the endangered world, with and scenarios produced by modeling of species’ dis- cord, which documents a replacement of forests H. faber. Using the same framework to test for long- Collectively, the results identify the central We show that the southern Atlantic forest was climatically unstable relative to the central region, manyunderstudied species ecosystems:and even genera the of Brazilian vertebrates Atlantic still tributions under palaeoclimates (7). by grasslands in the southern Atlantic forest dur- The hierarchical approximate Bayesian com- term persistence of refugial populations, we region as a hotspot within the Atlantic rainforest which served as a large climatic refugium for neotropical species in the late Pleistocene. This sets The palaeomodeling method intersects ing the LGM (14, 18) and suggests the occur- putation (HABC) method (24) allows us to use compare mtDNA sequences between the pre- hotspot and a refuge for biodiversity during new priorities for conservation in Brazil and establishes a validated approach to biodiversity 1 prediction in other understudied, species-rich regions. Museum of Vertebrate Zoology, University of California, predicted species’ distributions under current rence of small forest refugia in the southernmost data from all three species at once to test for dicted Pernambuco refugium [population 1 (Fig. climatic extremes of the Late Pleistocene. Berkeley, CA 94720–3160, USA. 2Biology Department, conditions and climatic extremes of the Late range of the putative Bahia refugium (19). The assemblage-wide responses to Late Quaternary 3A)] and adjacent (northern) populations from This is not to say that southern areas entirely Queens College, City University of New York, Flushing, NY Quaternary (6000 years before present, or 6 results also agree generally with forest models climate change. These analyses support both the Bahia refugium [population 2 (Fig. 3A)] to lacked forested habitats in the late Pleistocene: 11367, USA. 3Departamento de Zoologia, Instituto de ate Quaternary climate fluctuations helped refugia models have been dismissed because of Biociências, UNESP, Rio Claro, SP 3526-4100, Brazil. kybp, and 21 kybp) to predict areas of stability published previously (14), although the central model-driven hypotheses of (i) simultaneous, contrast alternative historical models H1 and H2. The existence of species and genera endemic to shape present-day diversity in temper- conflicting evidence (2, 3) or circularity in iden- 4Departamento de Zoologia, Instituto de Biociências, (regions in which species are predicted to oc- refugium extends farther south in the frog-based multispecies colonization of unstable areas from In this case, the HABC results infer long-term to the southern forests (27), as well as some Late and boreal systems (1), providing a tifying putative refugia (4), but historical pro- Universidade de São Paulo, SP 055008-090, Brazil. cupy irrespective of time period) and unstable models. Such differences are expected because adjacent refugial populations since the LGM, persistence of populations in isolated refugia for palaeoecological and genetic evidence (28), general context for understanding current pat- cesses must be invoked to explain regions of *To whom correspondence should be addressed. E-mail: areas (7, 14). Because the stability maps raise the forest and its associated species may differ as opposed to long-term persistence of popu- all three species (Z2 = 0, i.e., Z1 = 3), even when offer evidence to the contrary. Rather, the terns of endemism. In the tropics, Pleistocene high endemism (5, 6). Recent studies from sub- [email protected] specific hypotheses about regional differences slightly in their climatic tolerances and realized lations in unstable areas, and (ii) assemblage- allowing for postisolation migration (Fig. 3, D phylogeographically validated palaeomodels 36 37 www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 785 REPORTS

population expansion in the southernmost local- the LGM (0 to 20 kybp; Fig. 3A). The results stochasticity of a single-locus coalescent across ities of H. semilineatus may reflect low statistical indicate that all three species colonized the taxa. Combining data sets from several codis- power because of the exceptionally low levels of southern (unstable) areas after the LGM (Z2 = 3, tributed groups into a single hierarchical Bayesian diversity observed in this species. As predicted, the number of species evolved under H2), even analysis allowed us to estimate congruence across isolation by distance is not observed in unstable when allowing for postisolation migration (Fig. species, while borrowing strength from the full regions, but is detected within refugial areas for 3, B and C). When Bayes factor is used (25), comparative phylogeographic sample (24). This H. albomarginatus and H. faber. there is strong support for recent colonization in can translate into higher analytical power and be The hierarchical approximate Bayesian com- all three species (Z2 = 3) under the no-migration more informative than qualitative comparisons of putation (HABC) method (24) allows us to use model [B(Z2 = 3, Z2 < 3) = 35.16], and moderate species-specific analyses. By capturing the his- data from all three species at once to test for support under a postisolation migration model torical signal that emerges from larger, combined assemblage-wide responses to Late Quaternary cli- [B(Z2 = 3, Z2 <3)=5.70]. multispecies molecular data sets, HABC will mate change. These analyses support both model- Using the same framework to test for long- offer the possibility of looking at patterns of driven hypotheses of (i) simultaneous, multispecies term persistence of refugial populations, we com- historical community assembly in codistributed colonization of unstable areas from adjacent re- pare mtDNA sequences between the predicted nonmodel organisms for which barcode-type fugial populations since the LGM, as opposed to Pernambuco refugium [population 1 (Fig. 3A)] and DNA sequence information (e.g., mtDNA data) long-term persistence of populations in unstable adjacent (northern) populations from the Bahia can be feasibly collected. areas, and (ii) assemblage-scale, long-term per- refugium [population 2 (Fig. 3A)] to contrast al- Collectively, the results identify the central sistence of populations in isolated refugial areas, ternative historical models H1 and H2. In this case, region as a hotspot within the Atlantic rain- as opposed to post-LGM colonization of refugial the HABC results infer long-term persistence of forest hotspot and a refuge for biodiversity regions. populations in isolated refugia for all three spe- during climatic extremes of the Late Pleistocene. To test for assemblage-wide colonization cies (Z2 = 0, i.e., Z1 = 3), even when allowing for This is not to say that southern areas entirely of predicted unstable areas, we group mtDNA postisolation migration (Fig. 3, D and E). Using lacked forested habitats in the late Pleistocene: sequences from the southernmost refugial sites Bayes factor (25), we also detect evidence for The existence of species and genera endemic to REPORTS [population 1 (Fig. 3A)] and from localities in stability in both areas under the no-migration the southern forests (27), as well as some palaeo- unstable areas south of the refugium [population model [B(Z2 = 0, Z2 > 0) = 4.89], as well as under ecological and genetic evidence (28), offer evidence presence of two lineages that co-occur in the ad- 6.2% divergence). In contrast, sites located population expansion (23) are found in the 2 (Fig. 3A)] to contrast two alternative historical a postisolation migration model [B(Z2 = 0, Z2 > to the contrary. Rather, the phylogeographically jacent refugia. In all species, average net nucle- outside (south of) the refugia are genetically unstable area for H. albomarginatus and H. faber, models across the three codistributed species, 0) = 4.84]. validated palaeomodels presented here show that otide differences across localities (22) reflects more similar to each other, although to a lesser as well as in the Bahia refugium area for H. faber while allowing the taxon-specific demographic Relative to nuclear loci, mtDNA data are the central region had much higher stability parameters to vary. In H1, the long-term persist- more variable and readily collected and often relative to the south. Forest lizards (14, 29) and high geographic structure within refugia (2.6 to extent in H. faber (0.1 to 1.6%). Signatures of and H. semilineatus. The lack of signatureREPORTS of REPORTS ence model, two contemporary populations split provide key insights into biological response to birds (30) also show high diversity in the central from an ancestral population prior to the LGM environmental modification (1). Although single- portion of the biome relative to southern areas, Fig.presence 2. Genetic ofpresence two lineagesdiversity of two lineages that in putative co-occur that co-occur in the inA ad- the ad-6.2%6.2% divergence). divergence). In contrast,B sites sites located locatedpopulationpopulation expansionC expansion (23) are (23 found) are in found the in the (120,000 to 1.2 million years before present, or locus inference can be imprecise in the face of and provide evidence for population expansion in refugial (stable) versus unstable areas jacent refugia.jacent In refugia. all species, In all species, average average net nucle- net nucle-outsideoutside (south (south of) of) the the refugia are are genetically geneticallyunstableunstable area for areaH. foralbomarginatusH. albomarginatusand H. faberand, H. faber, Mybp, Fig. 3A). In H2, the recent colonization coalescent variance and the possibility of selec- southern regions. This reassures us that the pro- inotide the differences Brazilianotide differences Atlantic across across localities rainforest. localities (22) ( reflects22) reflectsmoremore similar similar to to each each other, although although to to a lesser a lesseras wellas wellas in theas in Bahia the refugiumBahia refugium area for H. area faber for H. faber model, population 2 is modeled as being colo- tion (26), our method benefits from a multitaxon cesses uncovered by the amphibian data may be (highTop) geographic Species-specifichigh geographic structure stability structure within maps; within refugia refugia (2.6 (2.6 to toextentextent in inH.H. faber faber (0.1(0.1 to to 1.6%). 1.6%). Signatures Signatures of ofand andH. semilineatusH. semilineatus. The lack. The of lack signature of signature of of nized from refugial population 1 subsequent to approach, while explicitly accounting for the generalized to and help to explain patterns of modeled refugia in black. (A) H. Pernambuco albomarginatus,(B) H. semilineatus, refugium Fig. 2. GeneticFig. 2. diversityGenetic diversity in putative in putative A B C Fig. 3. HABC analyses. (C) H. faberrefugial. Note (stable) the absence versus ofunstable large areas A B C Bahia refugium (A) Simulated models refugialstable regions (stable)in the in versus Brazilian the southern unstable Atlantic portion areas rainforest. inof the the forest Brazilian(Top (south) Species-specific Atlantic of the Bahiarainforest. stability and maps; * * H1 (long-term persist- (SãoTop) Paulo Species-specificmodeled refugia) refugia relative stability in black. to maps; the (A) H. Pernambuco ence) and H2 (recent modeledcentral andalbomarginatus refugia northern in areas. black.,(B) H. Asterisks ( semilineatusA) H. , refugium Pernambuco colonization). In both albomarginatus(C) H. faber,(B). NoteH. semilineatus the absence of, large São Paulo refugiumrefugium cases, each species was denote refugia inferred beyond the Bahia refugium (currentC) H. faber rangesstable. Note regions of the the absence in target the southern ofspecies. large portion modeled as two con- of the forest (south of the Bahia and * 5.4% * Symbolsstable regions indicate inlocalities the southern sampled portion for Bahia refugium temporary populations São Paulo refugia) relative to the * with mutation-drift pa- molecularof the forestcentral analysis. (south and Scale of northern the bar, Bahia areas.400 andkm. Asterisks * São Paulo refugium q q (Bottom)São Paulo Thedenote refugia) 50% refugia majority-rule relative inferred to beyond con- the the rameters 1 and 2 that central and northern areas. Asterisks split from an ancestral sensus Bayesiancurrent ranges phylogenetic of the target trees, species. 7% 5.4% São Paulo refugium rooteddenote with refugiaSymbols sequences inferred indicate from localities beyond the sampled oth- the for 7.8% population at a time t in ercurrent two rangescongenericmolecular of analysis.the species target Scale studied species. bar, 400 km. 5.4% the past. Ancestral pop- (rootSymbols not indicate shown).(Bottom) localities Thick The 50% internodes sampled majority-rule de- for con- ulation sizes are repre- sensus Bayesian phylogenetic trees, 5.3 – molecular analysis. Scale bar, 400 km. 7% sented by (qt) and (qt) ; note cladesrooted with with posterior sequences probability from the oth- 5.8% 7.8% 4% 1 2 greater(Bottom) than Theer 90%. two 50% congeneric Percentages majority-rule species indicate con- studied ybp, years before pres- Tamura-Neisensus Bayesian(root corrected not phylogeneticshown). distances Thick between internodes trees, de- 7% ent. (B to E) Hyperposte- 5.3 – 7.8% cladesrooted ( with20note). sequences clades with from posterior the probabilityoth- 5.8% 4% rior (bars) and hyperprior er two congenericgreater than 90%. species Percentages studied indicate 5.6% (dashed) densities of Z2 (root not shown).Tamura-Nei Thick corrected internodes distances de- between (number of species evolved clades (20). 5.3 – note clades with posterior probability 5.6% 5.8% 4% under H2) given data from greater than 90%. Percentages indicate three codistributed frog Tamura-Nei corrected distances between species. (B) and (C) Mod- clades (20). els of refugial sites (pop- 5.6% ulation 1) and unstable, southern areas (popula- tion 2). (D) and (E) Models of Pernambuco refugium (population 1) and Bahia refugium (population 2). (B) and (D) Postisolation migration not included in model; (C) and (E) postisolation migration included in model.

At a broader level, the congruence between An Online Reference, version 5.2, 15 July (2000). 788 SCIENCE model-based demographic hypotheses6 FEBRUARY and 20092008 (American VOL 323 Museum of Naturalwww.sciencemag.org History, 32. We thank U. Caramaschi and H. Zaher for Table 1. Population genetic summary metrics used in model validation. n, q, and average D values of the former were obtained not only from the total New York, 2008); electronic database providing access to collections MNRJ and Table 1. Population genetic summary metrics used in model validation. n, q, and average D valuesa of the former were obtained not only from the total joint, multispecies analyses of mtDNA diversity Sample size; S, number of segregating sites. The diversity parameter q and mean number of samples,a but also from all possible combinations of spatially accessible at http://research.amnh.org/ MZUSP; O. Peixoto, M. Gomes, A. Muri, R. Sample size; S, number of segregating sites. The diversity parameter q and mean number of samples, but also from all possible combinations of spatially shows that palaeoclimatic niche models and herpetology/amphibia/index.php. Kautskyi, S. Lima, E. Santos, J. S. Filho, J. D across localities are given per base pair (bp). Hs test (23) is used to detect contiguous localities distributed within the geographic extension of the unstable a Da across localities are given per base pair (bp). Hs test (23) is used to detect contiguous localities distributed within the geographic extension of the unstable assemblage-scale molecular genetic analyses V. Filho, G. Barros, J. Queiroz, R. Araújo, L. area. Parentheses encompass minimum and maximum values from subsamples. 14. A. C. Carnaval, C. Moritz, J. Biogeogr. 35, populationpopulation expansion. expansion. BA, Bahia; BA, Bahia; SP, São SP, PauloSão Paulo refugia. refugia. Because Because predicted predicted area. Parentheses encompass minimum and maximum values from subsamples. can be used to forecast spatial patterns of 1187 (2008). Japp, H. Japp, J. Giovanelli, J. Alexandrino, PPvaluesvalues in in bold bold highlight highlight statistical statistical significance significance at 0.05 at probability 0.05 probability level. level. refugia wererefugia often were larger often than larger predicted than predicted unstable unstable (recently (recently colonized) colonized) areas, areas,nn,,SS, diversity in poorly explored, highly threatened 15. M. Slatkin, Evolution 47, 264 (1993). L. Toledo, O. Araújo, G. Egito, J. Zina, D. 16. J. A. Hanley, B. J. McNeil, Radiology 143, 29 Loebmann, D. Pavan, R. Amaro, V. Verdade, n n SS q q q Mean Da D Hs Hs Mantel’s corr. coef.’ ecosystems. In a world of ever-accelerating Table 1. PopulationSpecies genetic summaryArea metrics used in model validation. n, , and average Da valuesMean of the formera were obtained notMantel only froms corr. the coef. total (1982). F. Curcio, M. Dixo, and J. Cassimiro for field Species Area (min.; max.) (min.; max.) (min.; max.) (min.; max.) (P value) (P value) environmental changes, this approach can help Sample size; S, number of segregating sites. The diversity(min.; max.) parameter(min.;q and mean max.) number(min.; of max.) samples, but(min.; also max.) from all possible(P value) combinations(P value) of spatially 17. J. Cohen, Ed. Psych. Meas. 20, 37 (1960). work assistance; W. Monahan and R. Hijmans D acrossH. localities albomarginatus are given perStable base pair (BA) (bp). Hs test36 (23) is used to207 detect contiguous0.076 localities distributed0.062 within– the20.546 geographic extension0.499 of the unstable to guide research and conservation in other for discussions about the modeling work; L. H.a albomarginatus Stable (BA) 36 207 0.076 0.062 –20.546 0.499 18. H. Behling, R. R. B. Negrelle, Quat. Res. 56, population expansion.(970 bp) BA, Bahia; SP, São Paulo refugia.(13; 23) Because predicted(81; 155) area.(0.034; Parentheses 0.072) encompass(0.020; 0.082) minimum(0.141) and maximum values(0.001) from subsamples. global hotspots or similarly complex tropical 383 (2001). Smith and D. Turong for DNA-sequencing (970 bp) Unstable (13; 23)27 (81;22 155) P values(0.034;0.003 in 0.072) bold highlight(0.020;0.001 statistical 0.082) significance–11.498(0.141) at 0.05– probability0.140 (0.001) level. 19. H. Behling, H. W. Arz, J. Pätzold, G. Wefer, assistance; R. Pereira, R. Damasceno, S. refugia were often larger than predicted unstable (recently colonized) areas, n, S, – – ecosystems. Unstable(south of BA) 27 22 0.003 0.001 (0.004)11.498 (0.580) 0.140 Palaeogeogr. Palaeoclimatol. Palaeoecol. Rovito, J. Kolbe, S. Singhal, R. Puschendorf, H. semilineatus (southStable of BA) (BA) n 28 S71 0.031q 0.036Mean D –17.778(0.004)Hs 0.054Mantel(0.580)’s corr. coef. 179, 227 (2002). and A. Pounds for discussions about earlier Species Area a References and Notes H. semilineatus(718 bp) Stable (BA) (min.;28(6; max.) 13) (min.;(14;71 max.) 58) (0.009;(min.;0.031 0.034) max.) (0.007;(min.; 0.041)0.036 max.) (0.029)(–P17.778value) (0.460) (P0.054value) 1. G. Hewitt, Nature 405, 907 (2000). [CrossRef] 20. K. Tamura, M. Nei, Mol. Biol. Evol. 9, 678 versions of the manuscript. Funding was (718 bp) Unstable (6; 13)15 (14;9 58) (0.009;0.003 0.034) (0.007;0.004 0.041) 0.114(0.029) 0.436 (0.460) 2. F. E. Mayle, D. J. Beerling, W. D. Gosling, M. (1993). provided by the NSF (awards DBI 0512013 H. albomarginatus Stable (BA) 36 207 0.076 0.062 –20.546 0.499 to A.C.C., DEB 0743648 to M.J.H., DEB Unstable(south of BA) 15 9 0.003 0.004 (0.357)0.114 (0.248) 0.436 B. Bush, Philos. Trans. R. Soc. London B Biol. 21. F. Tajima, Genetics 143, 1457 (1996). (970 bp)H. faber Stable (BA) (13; 23)28 (81;94 155) (0.034;0.018 0.072) (0.020;0.026 0.082)–38.111(0.141) 0.803 (0.001) 22. M. Nei, Molecular Evolutionary Genetics 416250 and DEB 0817035 to C.M.), Fundação (south of BA) (0.357) (0.248) Sci. 359, 499 (2004). (771 bp) Unstable 27(13; 23) (42;22 80) (0.012;0.003 0.022) (0.001; 0.044)0.001 (0.003)–11.498 (0.0003) –0.140 3. E. P. Lessa, J. A. Cook, J. L. Patton, Proc. (Columbia Univ. Press, New York, 1987). de Amparo à Pesquisa do Estado de São Paulo H. faber Stable (BA) 28 94 0.018 0.026 – –38.111 0.803 (southStable of BA) (SP) 15 48 0.023 0.028 5.981(0.004) 0.305 (0.580) Natl. Acad. Sci. U.S.A. 100, 10331 (2003). 23. J. C. Fay, C. I. Wu, Genetics 155, 1405 (2000). and Conselho Nacional de Desenvolvimento (771 bp) (13; 23) (42; 80) (0.012; 0.022) (0.001; 0.044) (0.115)(0.003) (0.221) (0.0003) 24. M. J. Hickerson, C. P. Meyer, BMC Evol. Biol. Científico e Tecnológico (grants to C.F.B.H. H. semilineatus Stable (BA) 28 71 0.031 0.036 –17.778 0.054 4. B. W. Nelson, C. A. C. Ferreira, M. F. da StableUnstable (SP) 15 18 4840 0.0150.023 0.0160.028 –13.255–5.981 0.0001 0.305 8, 322 (2008). and M.T.R.). Sequences are deposited in (718 bp) (6; 13) (14; 58) (0.009; 0.034) (0.007; 0.041) (0.029) (0.460) Silva, M. L. Kawasaki, Nature 345, 714 (south of SP) (0.014)(0.115) (0.456) (0.221) (1990). 25. R. E. Kass, A. E. Raftery, J. Am. Stat. Assoc. GenBank (FJ502639-FJ502822). A.C.C. and Unstable 15 9 0.003 0.004 0.114 0.436 Unstable 18 40 0.015 0.016 –13.255 0.0001 5. C. H. Graham, C. Moritz, S. E. Williams, 90, 773 (1995). C.M. designed the study; A.C.C., C.F.B.H., (south of BA) (0.357) (0.248) 26. J. W. O. Ballard, M. Kreitman, Genetics 138, and M.T.R. collected the data; A.C.C., C.M. (south of SP) (0.014) (0.456) Proc. Natl. Acad. Sci. U.S.A. 103, 632 (2006). H. faber Stable (BA) www.sciencemag.org28 SCIENCE94 VOL 3230.018 6 FEBRUARY0.026 2009 –38.111 0.803787 6. W. Jetz, C. Rahbek, R. K. Colwell, Ecol. Lett. 757 (1994). and M.J.H. analyzed the data; A.C.C. wrote (771 bp) (13; 23) (42; 80) (0.012; 0.022) (0.001; 0.044) (0.003) (0.0003) 7, 1180 (2004). 27. C. F. B. Haddad, L. F. Toledo, C. P. A. Prado, the paper. All authors discussed the results Anfibios da Mata Atlântica (Atlantic Forest and commented on earlier versions of Stable (SP) 15 48 0.023 0.028 –5.981 0.305 7. A. Hugall, C. Moritz, A. Moussalli, J. Stanisic, presented here show that the centralwww.sciencemag.org region much more distantlySCIENCE relatedVOLgroups 323 of Atlantic 6 FEBRUARY estation 2009in this region relative to the more ex- 787 Proc. Natl. Acad. Sci. U.S.A. 99, 6112 (2002). Amphibians) (Editora Neotropica, São Paulo, the manuscript. (0.115) (0.221) had much higher stability relative to the south. forest endemics. tensive forests in São Paulo and southern Brazil 8. L. P. C. Morellato, C. F. B. Haddad, Brazil, 2008). – Supporting Online Material Unstable 18 40 0.015 0.016 13.255 0.0001 Biotropica 32, 786 (2000). 28. M.-P. Ledru et al., Divers. Distrib. 13, 761 Forest lizards (14, 29) and birds (30) also show Because collection efforts, molecular studies, (9, 31). Not only could much unique diversity www.sciencemag.org/cgi/content/ (south of SP) (0.014) (0.456) (2007). high diversity in the central portion of the biome and conservation priorities have been heavily be lost, but ongoing habitat destruction could 9. M. T. Rodrigues, Conserv. Biol. 19, 659 full/323/5915/785/DC1 29. K. C. M. Pellegrino et al., Biol. J. Linn. Soc. relative to southern areas, and provide evidence biased toward southern and southeastern Brazil quickly erase the signature of the historical (2005). Materials and Methods London 85, 13 (2005). 10. C. Kremen et al., Science 320, 222 (2008). Figs. S1 to S6 for population expansion in southern regions. (8, 9, 31), we predict that genetic diversity and processes that led to it, preventing a full under- 11. Materials and methods are available as 30. G. S. Cabanne, F. M. d’Horta, E. H. R. Sari, www.sciencemag.org SCIENCE VOL 323 6 FEBRUARY 2009 787 Tables S1 and S2 This reassures us that the processes uncovered narrow endemism in the central corridor of the standing of the mechanisms underlyinglocal en- supporting material on Science Online. F. R. Santos, C. Y. Miyaki, Mol. Phylogenet. References by the amphibian data may be generalized to biome have been substantially underestimated. demism and, therefore, impeding more effective 12. M. B. Araújo et al., Ecography 31, 8 (2008). Evol. 49, 760 (2008). 8 October 2008; accepted 9 December 2008 and help to explain patterns of diversity in other, This is serious, given the higher rate of defor- conservation measures. 13. D. R. Frost, Amphibian Species of the World: 31. J. M. C. da Silva, M. Tabarelli, Nature 404, 72 10.1126/science.1166955 38 39 Our New Darwin shirt — yours free when you become a AAAS member.

Join Today!

Special offer: $99 Professional $50 Postdoc/Student* Become a AAAS member today, at this special introductory offer and receive a free Darwin shirt (a $22.50 value free). Your membership gives you all of the personal membership benefits of AAAS including 51 issues of Science magazine, career development tools, and much more — all while supporting our international, public policy, and educational programs — the ones that advance science and serve society.

Order online at promo.aaas.org/darwin24

Visit promo.aaas.org/giftdarwin to order now *International orders will receive Science Digital edition - other membership categories are available online.

J-2412_PlusYou_HumanRights_Fullpg.indd 1 5/21/2009 10:54:19 AM darwin_fullpg_tradeshow3.indd 1 5/20/09 9:16:56 AM