Conserving Biodiversity Coldspots

Recent calls to direct conservation funding to the world'sbiodiversity hotspots may be bad investment advice

Peter Kareiva and Michelle Marvier

Peter Knrelva is Lead Scientistfor The Nature Conseivana/.His current focus is on the ex-amination of different consenm-tion strategies, although he alsomaintains a program of field re-search on Úie invasion of NorthAmerica hy exotic ladybird bee-tles. Knrdva is also affiliatedwith the Bren School of theUniversity ofGaliforiiia, SautaBarbara and the EnvironmentalStudies Institute at SantaClara University. MichelleMawier is a faculty member inthe Department of Biologi/ atSanta Clara University. Her re-search involves the studi/ ofparasitic plants and the envi-ronmental implications ofge-nelicaliy modißed organisms.Address for Kareiva: The Na-ture Consenmna/, 217 PineStreet, Suite 1100. Seattle, WA98101. ¡nternet:pkareiva@ttic,org

The numbers are chilling. Every year tropi-cal forests covering an area the size of

Poland are destroyed. With them, perhaps tenthousand species are wiped out annually, mostbefore they can be so much as cataloged. Someliken the current calamity to the last episodeof mass extinction—65 million years ago whena wayward asteroid killed off the dinosaursalong with about two-thirds of the species thenin existence.

Tlie rapid loss of tropical forests throughoutthe worlcl and the widely recognized "biodi-versity crisis" have spurred various non-governmental conservation organizations andinternational agencies to develop strategies forprotecting natural habitats. But the scale of thecrisis is so daunting that conservationists wide-ly accept the need for some sort of triage,whereby limited funds go to the places wherethe greatest good can be done. Experts haveexplored various ways to set priorities, and al-most without exception, rainforests get topbilling. The reason is simple: These tropicalecosystems harbor more unique species thanany other habitat or place. Identifying and pro-tecting such "biodiversity hotspots" has thusbecome the reigning scientific paradigmamong conservationists.

Biodiversity hotspots are regions with un-usually liigh concentrations of endemic species(species that are found nowhere else on Earth)that also have suffered severe habitat destruc-tion. Norman Myers, an environmentalist affil-iated with the University of Oxford, firstcoined tliis term in a scholarly paper he wrotein 1988. Now, virtually every textbook on con-servation biology contains a map of theworld's biodiversity hotspots. Although lushtropical rainforests first leap to mind, oceanicislands and Mediterranean ecosystems such asthose found in California, South Africa andAustralia are also considered hotspots becausethey, too, show exceptionally high rates ofplant endemism.

The hotspot concept has been extremely ef-fective at directing international funding andphilanthropy. Given this success, we think it

worth pausing to examine the scientific foun-dation of this conservation strategy and to con-sider what the consequences of this conceptmay be for the huge expanses of the planet thatit leaves out in the cold—places we might dubbiodiversity "coldspots."

Does it make scientific sense to downplaythe world's steppes, the Serengeti, the wildArctic and other relatively species-poor areasin favor of biodiversity hotspots? Clearlyspecies richness should be considered whendeciding where to invest conservation dollarsand effort. But there are other relevant factors.We believe that ecological theory, considera-tion of ecosystem services and sociopolitical re-alism all argue strongly against placing toomuch emphasis on biodiversity hotspots. Yet,lamentably, little consideration has been givento alternative frameworks for setting priorities.

Warming Up To the IdeaBefore probing the weak points of hotspot-based strategies, we should explain preciselywhat data and reasoning have been used to de-tine this approach to conservation. Hotspotsare meant to represent areas of high biologicalvalue that are under severe threat (the menaceusually being people's readiness to destroy thenatural habitats around them). It is hard to findfault with this general line of reasoning. How-ever, some potential problems become appar-ent after you realize that there are in fact manydifferent ways to quantify biological value. Thehotspot approach emphasizes species richness.More specifically, it calls for tallying up thenumber of endemic plant species in a region.Although this simple yardstick correlates withother measures of species richness at huge spa-tial scales (say, tropical versus temperate), sev-eral analyses over the past decade have re-vealed that the hotspots for different taxa donot coincide well with one another, thathotspots often miss rare species and that majoranimal groups could be lost by devoting toomuch attention to endemic plants.

Why then just count plants? Because conser-vationists need some quantitative way to

344 American Scienfist, Volume 91

Rowell/LorbiK

Figure 1. Faced with limited resources, many conservationists are concentrating their efforts on biodiversity "hotspots," comparatively small ar-eas that each harbor a large number of species that can be found nowhere else. The authors argue, however, that such a strategy is problemat-ic, in part because it ignores the requirements of large carnivores, like this polar bear living near the shores of Hudson Bay. These creatures re-quire vast territories, which may not be particularly rich in species.

gauge species richness, which is likely to be acomponent of anyone's figuring of biologicalvalue, and the available data tend to be morecomplete for plants than for anin:ials. So count-ing plants is certainly a reasonable approach.One just needs to recognize the limitations ofworking with this rather one-dimensional rep-resentation of biodiversity.

The degree of threat to an area is even hard-er to quantify, because there are many kinds ofperns facing species and many different waysto evaluate the hazards. Moreover, there is no apriori reason for selecting one measure of dan-ger over another. The hotspot approach focus-es on the percentage of primary vegetation thathas been destroyed. To qualify as a hotspot, aregion must have suffered a loss of at least 70percent of the original vegetation. Again thisdefinition is reasonable, but it has its limits—inparticular because the fraction of natLiral plantlife that is already gone is more a statementabout land use in the past than a direct indica-tion of future threat. Thomas D. Sisk of North-ern Arizona University and his colleagues

have argued that the danger may be more ac-curately reflected in the current rate of humanpopulation growth nearby or in the fraction ofspecies in the area that are currently regardedas being at risk.

Clearly, the détails of how hotspots are de-nned are subject to disagreement, but the gener-al concept still appears compelling: Hotspotsseem to be a way of directuig conservation ef-forts toward the most cost-effective actions. In-deed, the notion of cost-effectiveness is a topselling point. As Myers and four colleagues ar-gued three years ago in an influential Nature ar-ticle, "44% of all species of vascular plants and35%. of all species in four vertebrate groups areconfined to 25 hotspots comprising only 1.4%of the land surface of the Earth." Their basic pre-scription is to collect as long a list of species aspossible in as small an area as possible and tar-get conservation first where people have al-ready destroyed a great deal of habitat. Al-though these authors (and some others whohave taken up the baton) do acknowledge theneed for complementary strategies, the hotspot

viTVW.americanscientist.org 2003 July-August 345

Figure 2. Biodiversity hotspots, as defined by Myers et al. 2000 (red), span the globe, but they fall mostly in the tropics. The influence of Myers'shotspot concept is evident in the distribution of the office locations for the three largest nongovernmental conservation organizations (greendots)—Conservation International, The Nature Conservancy and the World Wildlife Fund—shown here with the many office locations in theunited States and Western Europe excluded. (Background colors from satellite imagery, courtesy of ARC Science.)

concept has grown so popular in recent yearswithin the larger conservation community thatit now risks eclipsing all other approaches.

Nol-So-Hot ConsequencesAt first glance it may seem self-evident thatconservation investment should be funneledinto the regions or countries with the most bio-diversity. But the hotspot methodology is logi-cal only if the exclusive goal of conservation is

\ Yellowstone( "v ^i National

=- ' ^ ' Park

greaterYellowstoneecosystem

Figure 3. State of Montana and the country of Ecuador cover similar areas, but the latteris home to far more species. In weighing how to address conservation concerns in thesetwo places, should all available resources be focused on preserving Ecuador's richernatural capital? No, the authors ai-gue; it would be better to allow a small fraction ofEcuador's species to perish if this sacrifice would permit the protection of the samenumber of species in Montana, where they would amount to a larger fraction of theoriginal flora and fauna. Conservation efforts in Montana would also belp to protect thegreater Yellowstone ecosystem (dnslied line), the only place in the lower 48 states wherelarge mammals such as grizzly bears, wolves, bison and elk still roam freely in the wild.

to protect the largest possible number ofspecies in the smallest possible area. Usinghotspots to set priorities comes into question assoon as one considers a broader range of objec-tives, such as maintaining fimctioning ecosys-tems throughout the world, providing thegreatest variety of distinct plant and animallineages for future evolutionary break-throughs, preserving spectacular wild land-scapes that inspire the human spirit or protect-ing nature in a way that provides for thewell-being of people living alongside.

A hypothetical example reveals some of theunfortunate side effects that arise from empha-sizing hotspots above all else. Consider two ar-eas of roughly equal size, the country ofEcuador and the state of Montana. Ecuador is arenowned biodiversity hotspot, harboring2,466 vertebrate species and 19,362 vascularplant species. In contrast, Montana is a biodi-versity coldspot, with only 12 percent ofEcuador's species richness. Clearly, if one mea-sures success as protecting the largest numberof species in the smallest possible area, itmakes sense to ignore Montana and to concen-trate solely on Ecuador. But assume for the mo-ment that we desire some level of conservationeffort in both places. Suppose we set a goal ofensuring protection for 20,000 total speciesfrom these two areas. We could attain that out-come by preserving 18,000 species in Ecuadorand 2,000 species in Montana, or, alternatively,by safeguarding 19,000 species in Ecuador and1,000 species in Montana. If all that matters isthe total number of species protected, thesetwo strategies are equivalent.

In reality these two choices would have vast-ly different consequences on the groimd. Bothwould leave Ecuador with the buJi: of its biodi-

346 American Scientist, Volume 91

versity intact (82 percent C5r 87 percent, if, for ar-gument's sake, one considers just vertebratesand vascular plants) and, presumably, with rea-sonably well-functioning ecosystems. But shift-ing from the first to the second strategy cuts thefraction of species protected in Montana se-verely (from 74 percent to 37 percent).Youwouldn't have to be a scientist to notice the dif-ference between saving three-quarters versustwo-fifths of the species in the state. But eventhe best-designed scientific monitoring pro-grams would be hard pressed to register thedifference between having 87 percent or 82 per-cent of the species under protection in Ecuador.

This example illuminates a major flaw withapproaches to conservation that are solelybased on hotspots. If we measure success sim-ply by tallying up total species protected, werisk the folly of allowing major ecosystems todegrade beyond repair simply because they donot provide lengthy species lists. For instance,the Yellowstone ecosystem, which includesparts of Montana, Idaho and Wyoming, isspecies poor, relatively speaking. Yet tliis regionharbors the last assemblage of large mammalsand carnivores in the lower 48 states (grizzlybears, wolves, bison, elk and so forth). And be-cause the Yellowstone ecosystem contains theworld's first national park, it is obviously animportant locale for continued conservation,despite the relative paucity of species that existthere. So by itself, the niunber of species savedis an inadequate barometer of success. Otherdimensions of the problem need to be consid-ered as well, including what environmentahstsoften describe as "ecosystem services."

Earth's Life-Support SystemSociety depends on the products and servicesecosystems provide to a far greater extent thanmost of us realize. When asked what naturegives them, most people could name a fewfoods, drugs and building materials derivedfrom wild species. But they would probablytake for granted many other products of na-ture, including clean air, abundant fresh water,fertile soil and a benign climate. In short, theywould fail to recognize the critical role that di-verse communities of species play in fosteringa healthy and predictable environment.

Would not hotspot-based conservationstrategies automatically take such effects intoaccount? Hardly. Investing conservation effortsonly in hotspots could lead us to ignore andpotentially lose some of our most valuableecosystems simply because they harbor fewplant species. Consider, for example, the factthat hotspot analyses so often point to tropicalforests as areas of highest priority. Tliese forestsdo indeed provide important ecosystem ser-vices, such as climate regulation and nutrientcycling, which Robert Costanza, an economistat the University of Maryland, and his col-

www, americanscientist.org

leagues recently valued at about $2,000 perhectare per year. But hotspot analyses overlookthe value of wetlands—and it's easy to xmder-stand why: A typical Spartina marsh has no en-demic plants and no more than 20 or 30 plantspecies total. Still, tidal marshes offer consider-able ecosystem services, such as flood regula-tion, waste treatment and fisheries production,with an estimated annual value of nearly$10,000 per hectare per year. Clearly, marshesare precious resources, not just for people butalso for the enormous variety of wildlife thatdepends on clean water. Yet by any sort ofhotspot reckoning, these marshes would comeout at the very bottom of the pile.

This difficulty would not arise if the statedconservation goal were to preserve the func-tioning of the planet's ecosystems to the maxi-mum extent possible. Tliis objective would notnecessarily be incompatible with a desire tosave species: Many empirical studies haveshown that ecosystem services, such as pro-duction of plant biomass, retention of nutri-ents, resistance to drought, polhnation of cropsand decomposition of organic matter, declinewith major losses in biodiversity. But one strik-ing feature of the relation between ecosystemservices and biodiversity is that it is not linear.Rather, the benefits of biodiversity are quicklyrealized with an initial accumulation of speciesand thereafter remain constant, so that protect-ing more species does not forever translate intomore or better ecosystem services.

Figure 4. Wetlands, such as this one near Mobile, Al-abama, provide valuable ecosystem services. Yet theseplaces may host only a small number of plant species. Atypical Spnrtina marsh, for example, harbors fewer than30. (Courtesy of the U.S. Army Corps of Engineers.)

2003 ]uly-August 347

It is not clear how to predict where this satu-ration effect will manifest itself when one con-siders a variety of possible benefits. But if we ac-cept that benefit curves do, level off, we mustaccept tliat there are diininis^iing returns for pro-tecting ever more species in any particularecosystem. Given this pattern, a logical goalmight be to ensure that no major ecosystem suf-fers greater than a 10 percent loss, or a 20 percentloss, or a 50 percent loss of diversity.

Although scientists lack the knowledge toassign values to the services provided by theecosystems in all countries or regions, whatmatters is obviously not how many speciesthere are as much as what percentage of thenative diversity remains sheltered from de-struction. The hotspot approach would resultin high levels of protection for a few species-rich areas to the neglect of many others. Thus,setting conservation priorities using onlyhotspots as a guide could well bring on an un-fortunate side effect: more degradation of glob-al ecosystems than would take place if a morebroadly based strategy were used.

Thicker Branches on the Tree of LifeMost conservationists emphasize ecological orecosystem value when discussing the need topreserve biodiversity. But they often neglect anaspect of biodiversity that is just as important:its worth as a resource for future evolutionaryinnovation. In this regard, one has to considermore than just species. For example, a greateramount of evolutionary liistory and biologicaldistinctiveness is lost when the last species ofan entire genus or family becomes extinct thanwhen a species with many close relatives dis-

1/16-

10 15plant species richness before drought

20

Figure 5. Fcosystem functioning generally improves as the number of species rises, butthe relation is not linear. Typically, the benefits of increasing the species count accruequickly and then level out. Here, one sees that the resistance of an ecosystem to drought,measured as the ratio of biomass after tlie drought to the biomass before, depends onspecies numbei^-but only to a point. (Adapted from Tilnian and Downing 1994.)

appears. Some evolutionary biologists havechallenged the current focus on biodiversityhotspots as myopic, because it neglects theunique value of distinct evolutionary lineages,which represent very different life historiesand forms. Rather than simply directing effortstoward areas with rich sets of endemic species,conservation organizations might better con-centrate on saving higher taxonomic groupsunder threat.

One could, for example, try to rescue generathat are in danger of being entirely lost. To testthis idea, we looked at the distribution ofmammal and bird genera with more than onespecies for which the World ConservationUnion (IUCN) has listed all their constituentspecies as critically endangered, endangered,vulnerable or dependent on conservation. Foreach of the members of these highly tlirea tenedgenera, we then searched IUCN's latest"redlist" database of endangered species andrecorded all countries in which these rare crea-tures still occur naturally or have been reintro-duced. We then ranked the countries accordingto the number of highly threatened genera thatcan be found there.

This procedure provides a Hst that differsgreatly from one based solely on the richness ofendemic plants. In particular, this exercisepohits to a greater need for conservation in cer-tain African nations than does a hotspot-basedapproach. For example, Kenya, with only 265endemic plant species, has never been identi-fied as a biodiversity hotspot. Yet our rankingsuggests that the conservation of Kenyanwildlife should receive urgent attention. Afterall, this nation is home to species belonging toeach of the following eight highly threatenedmammalian genera: Alcelaphus (hartebeest),Comwcliaetes (gnu), Hippotragiis (a type of ante-lope), Orijx (another type of antelope), Otomops(a type of bat), Redunca (reedbuck), Rliynchocyon(elephant shrew) and Surdisorex (mole shrew).

Another shortcoming of the hotspot method-ology becomes apparent when one looks care-fully at how it is often applied. Although Myersand his colleagues initially proposed hotspotsas a means of setting conservation priorities at avery large scale and in a coarse manner, the no-tion of getting the most species per unit area ofland protected ("efficiency") has been translat-ed to much smaller spatial scales with poten-tially unfortunate consequences. For example,in the past few years two influential analyses ofbiodiversity within the U.S. have been used toshow how conservationists might efficientlyprotect species by focusing on just a few smallclusters of critical counties. This strategywould, however, fail to protect adequatelythose species that require large tracts of rela-tively undisturbed habitat.

An added worry surfaces when you take along-term, evolutionary perspective on the

348 American Scientist, Volume 91

Figure 6. Conservation priorities can be set using any number of different strategies. Here the authors examined the distribution of bird andmammal genera that are in danger of being entirely wiped out and scored countries according to how many of these genera are still found there.The resulting map shows some similarities with tbe hotspots of Myers (dark shading) but also many differences.

problem. By focusing on conserving the mostspecies in the smallest possible area (for the sakeof cost-effectiveness), conservationists may in-advertently be altering the course of evolution.How? Protecting hotspots at small scales favorsspecies that can live in relatively restricted areas.Thus we might expect rodents to enjoy highspeciation rates relative to extinction rates. Atthe other extreme, species that require vast ter-ritories, and thus cannot be contained in cost-ef-fective hotspots, will suffer disproportionate ex-tinction relative to speciation. Hence primatesand large carnivores woi.üd be expected to wanerelative to their smaller mammalian counter-parts. Recent analyses of verttibrate extinctionand speciation rates suggest that this is exactlywhat is happening. Thus, as Donald A. Levin(of the University of Texas at Austin) and PhillipS. Levin (of the National Marine Fisheries Ser-vice) argued in these pages not too long ago(Macroscope, January-February 2002), the Earthmay well end up with a paucity of primates andrhinoceroses, and a surplus of rodents.

One hopes that eventuality will not actuallycome to pass. Indeed, one hopes that societywill ultimately prove able to protect habitatsand ecosystems throughout the world so thatevolution will be allowed to take a naturalcourse. But in the short term it's clear thatsome tough decisions need to be made in allo-cating scarce resources. Thus far, most conser-vationists have focused on biological value asthe major criterion for setting priorities. Butothers have explored different sides of theproblem, including the degree of immediatethreat to biodiversity, general feasibility andcost-effectiveness. Although there are manyways to go about doing this, it's hard to arguefrom scientific principles that any simple set of

variables can be cast into the ideal metric forscoring urgency. This leaves scientificallyminded conservationists with the unansweredquestion: What then should we do?

One thing is to make sure that the problem isframed properly. In recent times, biologists haveachieved a major conceptual advance by realiz-ing that conservation programs must span po-litical boundaries and that for planningpurposes biologically defined areas such as"eco-regions" represent natural units. Althoughthis mindset is useful for quantifying biologicalvalue, it is not terribly helpful for making deci-sions about tlie feasibility or cost-effectiveness ofa particular conservation project. When an or-ganization decides to work in a region, it willnecessarily have to deal with the regulatoryagencies, legal institutions and people of indi-vidual nations. Thus, although our conservationgoals may be biological, as long as feasibilityand cost-effectiveness are important, countrieswill remain essenfial units for considerafion.

One might thus be tempted to try to identifya comprehensive list of countries having greatbiological value and where an immediatethreat looms—but also where action is cost-ef-fective and feasible. Unfortiniately, conserva-tionists typically lack the information neededto assess all of the relevant variables. For in-stance, ecosystem services have not been wide-ly calculated, current rates of habitat loss areonly erratically available, and the uniquenessor irreplaceability of particular environmentsaro complicated to assess. One solution is toinvestigate many different approaches to seewhether certain areas are consistently identi-fied as being of high priority. In other words,are some conclusions robust, no matter whichcriteria are used?

www.americanscientist.org 2003 July-August 349

Clockwise (torn left: Nigel J. Dennis, Toni McHueh,Tim DavLs. Merlin D. ruttk'/i^oto Reseiirchers. Inc.

Figure 7. Although it does not appear on any hotspots list, Kenya deserves urgent attention because it harbors members of eight highly threat-ened mammalian genera, including the four shown here: Uippotragiis (a type of antelope, left), Rhytichocifon (elephant shrew, middle top), Oto-mops (a type of bat, middle bottom) and Ori/x (another type of antelope, right).

In this spirit, we have invested quite a bit oftime recently exploring all sorts of methods forsetting conservation priorities. And what we'venoticed is that some countries seem always tocome out on top. A few—such as India, Viet-nam, South Africa, the Philippines and Brazil—overlap with the top 25 hotspots that Myers

number of endangeredplanl speciesiv. 19 to 77• 1 8 to 18

.3or4

L_ 10 or no data

Figure 8. Using the threat to local species to identify certain areas as biodiversityhotspots, a strategy originally intended to establish conservation priorities on a broad,regional scale, has also been employed to evaluate small areas such as U.S. coun-ties—with great potential for misapplication, according to the authors. Although onecould economically concentrate conservation efforts using maps such as this one, the"hotspot counties" identified are commonly too small and too disjoint to provide ad-equate range for highly valued carnivores such as panthers, mountain lions, grizzlybears and wolves. (Adapted from Dobson et al. 1997).

and his colleagues identified. Interestingly,these five countries have relatively small per-centages of their land under conservation pro-tection, which suggests that they may be rela-tively sure bets, given that so much room forprogress exists there. Yet other nations appear-ing on a great many of our priority lists havenever been identified as biodiversity hotspots.

Dispensing with ListsBiodiversity hotspots represent an initial andpioneering effort at establishing conservationpriorities. But by relying too much on counts ofplant species, this approach loses sight ofwhole ecosystems, habitats and the needs ofpeople. To work the metaphor a little harder,you might say that biodiversity hotspots leavetoo many places and people out in the cold.

Because many conservation threats are nowglobal in their origin and scope (for example,climate change and invasive species), place-based priorities risk disenfranchising toomany people from the challenge at hand. In-deed, on reflection, we worry that the initiallyappealing notion of getting the most species orgreatest biological value per unit area is, infact, a thoroughly misleading strategy. Howmuch of a victory would it actually be if peo-ple did manage to conserve only the 1.4 per-cent of the Earth's land surface that containsalmost half the world's vascular plants? Thereality is that people must make conservationprogress everywhere. Doing that requires nota ranking of theoretically deserving places buta prioritization that takes into account the ef-fectiveness of past conservation efforts. A per-formance-based system would not only hold

350 American Scientist, Volume 91

conservation organizations more accountable,it would also provide incentives to countriesattempting to implement conservation mea-sures, so that those nations demonstrating suc-cesses on the ground would be more likely toreceive funding in the future. Anyone who hasworked for a while in conservation knows thatcertain people and certain leaders can over-come enormous obstacles and do wonders inthe most unlikely places. Yet none of the es-tablished priority-setting schemes recognizessuch human factors.

We believe that the officers and directors ofall too many foundations, nongovernmentalorganizations and international agencies havebeen seduced by the simplicity of the hotspotidea. Perhaps that's why, for example, 10 per-cent of the World Bank's biodiversity projectsare located in a single country: Brazil. This factis particularly noteworthy because the WorldBank is the largest investor in biodiversity con-servation. And collectively, the three largestnongovernmental conservation organizations(The Nature Conservancy, World WildlifeFund and Conservation International) cluster adozen offices in Mexico and put many also inBrazil, Indonesia, Madagascar and the Philip-pines. Meanwhile, countries with vast biologi-cal resources such as Russia and Argentina to-gether host only three offices. Russia, in fact,rarely gets mentioned in conservation circles—perhaps because information on plant diversi-ty in that sprawling country is so lacking that itis ineligible for hotspot lists.

Unwavering support for the protection ofhotspots oversimplifies the difficult decisionsthat must be made in deciding which projectsto fund and where to invest money. Althoughbiodiversity hotspots are indeed an academi-cally appealing idea, blind adherence to this

mantra runs the risk of leaving the world witha sizable collecfion of species in a few areas butwith an environment that is otherwise largelydegraded. Rather than trying to identify denseconcentrations of species on a map, we andother conservationists should be more flexibleand should be prepared to reward effective ac-tions on the ground as they happen. If we doso, we will surely discover plenty of coldspotsdeserving of our attention.

BibliographyCostanza, R., R. d'Arge, R. de Groot, S. Ftirbcr, M. Gras-

so, B. Hannon, K. Limburg, 5. Noeem, R.V. O'Neill, J.Panielo, R. Raskin, T. Sutton and M. Van den Belt.1997. The value of the world's ecosystem services andnatural capital. Nature 387:253-260.

Dobson, A. P., j . P. Rodriguez, W. M. Roberts and D. S.Wilcove. 1997. Geographic distribution of endangeredspecies in the United States. Science 275:550-553.

Levin, R S., and D. A. Levin. 2002. The real biodiversitycrisis. American Scientist 90:6-8.

Myers, N. 1988. Tlireatened biotas: 'Hotspots' in tropicalfon;sts. The Fnvironmentaiist 8:187-208.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B.da Fonseca and J. Kent. 2000. Biodiversity hotspotsfor conservation priorities. Nrtfiirc 403:853-858.

Prendergast,J. R., R. M. Quinn,j. H. Lawton, B.C. Ever-sham and D. W. Gibbons. 1993. Rare species, the co-incidence of diversity hotspots and conservationstrategies. Nature 365:335-337.

Sisk, T. D., A. E. Launer, K. R. Switky and P. R. Ehrlich.1994. Identifying extinction threats. Bioscience44:592-604.

Tilman, D., and J. A. Downing. 1994. Biodiversity andstability in grasslands. Nature 371:363-365.

For relevant Web links, consult this issue of AmericanScientist Online:

http://v^Tvw.americanscientist.org/template / IssueTOC / issue / 394

bef spacPood Qnd uJno

www.americanscientist.org 2003 luly-August 351