BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors nonprofit publishers academic institutionsresearch libraries and research funders in the common goal of maximizing access to critical research

Toward a Blueprint for Conservation in AfricaAuthor(s) THOMAS BROOKS ANDREW BALMFORD NEIL BURGESS JON FJELDSAringLOUIS A HANSEN JOSLIN MOORE CARSTEN RAHBEK and PAUL WILLIAMSSource BioScience 51(8)613-624 2001Published By American Institute of Biological SciencesDOI httpdxdoiorg1016410006-3568(2001)051[0613TABFCI]20CO2URL httpwwwbiooneorgdoifull1016410006-3568282001290515B06133ATABFCI5D20CO3B2

BioOne (wwwbiooneorg) is a nonprofit online aggregation of core research in the biological ecologicaland environmental sciences BioOne provides a sustainable online platform for over 170 journals and bookspublished by nonprofit societies associations museums institutions and presses

Your use of this PDF the BioOne Web site and all posted and associated content indicates your acceptanceof BioOnersquos Terms of Use available at wwwbiooneorgpageterms_of_use

Usage of BioOne content is strictly limited to personal educational and non-commercial use Commercialinquiries or rights and permissions requests should be directed to the individual publisher as copyright holder

August 2001 Vol 51 No 8 bull BioScience 613

Articles

In the last two decades various quantitative techniques for assessing conservation priorities have been

developed based on data about the distribution of species(Reid 1998 Williams 1998 Margules and Pressey 2000)These methods have been applied extensively in temperate re-gions such as North America at both the state (Csuti et al1997) and national (Dobson et al 1997) levels However bi-ological diversity is concentrated in the tropics and it is herethat conservation faces the most pressing threats (Raven1988) Furthermore fine resolution data are often so scarceand local land-use patterns so diverse as to limit our abilityto apply quantitative prioritization techniques at fine scales(Pimm and Lawton 1998) Hence such techniques may be par-ticularly appropriate for application in tropical areas and atcontinental scales Until recently this application has been re-stricted to single families (Kershaw et al 1994 1995) or or-ders (Hacker et al 1998) because continent-level species dis-tribution data from the tropics are rarely compiled

Recognizing this limitation the Zoological Museum ofthe University of Copenhagen embarked on a program tocompile continent-level data on the distributions of tropicalspecies Such data have allowed the use of quantitative con-servation prioritization techniques for birds for which dataare better than for other taxa in both South America andAfrica (Burgess et al 1997 de Klerk 1998 Fjeldsaring and Rah-bek 1997 1998 1999) Simultaneously the museum has com-piled data from Africa for three other major taxa mammalssnakes and amphibians (Burgess et al 1998) The Zoologi-cal Museum of the University of Copenhagen intends topublish these data as an atlas of African biodiversity

This article begins a series planned to extend these analy-ses across four major terrestrial taxa for an entire tropical con-tinent For each taxon we first present an overview of patternsof species richness and narrow endemism across the continentSecond we use these patterns to identify areas of high con-servation priority that can most efficiently represent eachgroup of species Finally we repeat the analysis for threatenedspecies (Baillie and Groombridge 1996) the immediate pri-orities for conservation in Africa

Numerous important issues remain to be addressed andwe conclude the article with a detailed discussion of them Inour ongoing research we are address four in particular Firstis the question of surrogacymdashhow well do conservation pri-orities for one taxon represent other taxa Second is the mat-ter of incorporating socioeconomic variables into the analy-ses both as pressure (eg human population) and as response(eg existing protected areas) variablesmdashthe need to do sois urgent The third issue involves scale We are using envi-ronmental models to increase the resolution of the databasesto scales relevant to local (rather than continental) conser-

Thomas Brooks (e-mail tbrooksconservationorg) is a postdoctoral research associate in the Zoological Museum University of Copenhagen Uni-

versitetsparken 15 DK-2100 Copenhagen Oslash Denmark and in the Conservation Biology Group Department of Zoology University of Cambridge

Downing Street Cambridge CB2 3EJ UK he is also the director of biodiversity analysis at the Center for Applied Biodiversity Science Conserva-

tion International 1919 M Street NW Suite 600 Washington DC 20036 Andrew Balmford is a lecturer in the Conservation Biology Group De-

partment of Zoology University of Cambridge Neil Burgess is the technical advisor to the Uluguru Mountains Biodiversity Conservation Project

Wildlife Conservation Society of Tanzania Pamba House PO Box 1668 Morogoro Tanzania Jon Fjeldsaring is a professor and the curator of birds

Louis A Hansen a database manager and Carsten Rahbek an associate professor in the Zoological Museum University of Copenhagen Joslin

Moore is a postdoctoral research associate in the Zoological Museum University of Copenhagen and in the Conservation Biology Group Depart-

ment of Zoology University of Cambridge Paul Williams is a researcher in the Biogeography and Conservation Laboratory Department of Entomology

Natural History Museum Cromwell Road London SW7 5BD UK copy 2001 American Institute of Biological Sciences

Toward a Blueprint forConservation in AfricaTHOMAS BROOKS ANDREW BALMFORD NEIL BURGESS JON FJELDSAring LOUIS A HANSEN JOSLIN MOORECARSTEN RAHBEK AND PAUL WILLIAMS

A NEW DATABASE ON THE DISTRIBUTION

OF VERTEBRATE SPECIES IN A TROPICAL

CONTINENT ALLOWS NEW INSIGHTS INTO

PRIORITIES FOR CONSERVATION ACROSS

AFRICA

614 BioScience bull August 2001 Vol 51 No 8

Articles

vation planning The last issue concerns consensus buildingThe databases and analyses should become available to on-going regional national and local conservation planningprocesses in Africa through workshops and collaborativework The consensus built by these processes is essential forensuring that any of these studies mentioned above results inthe effective conservation of biodiversity

The distributional databasesWe compiled data on the distribution of all currently recog-nized 1921 species of birds 940 species of mammals 406species of snakes and 615 species of amphibians found inmainland sub-Saharan Africa (south of 20 degrees North)onto a one-degree grid with each cell approximately 105km on each sideWe chose the one-degree resolution as a com-promise between the sampling inadequacy of finer resolutionsand the loss of biogeographic detail (especially in moun-tainous regions) at coarser resolutions (Rahbek and Graves2000) For the 1957 one-degree grid cells across the continentholding data we currently have a total of 828506 species-in-grid-cell data entries for the 3882 species overall This is a dy-namic database into which new grid-cell data entries are be-

ing added almost daily it is the most complete cross-taxonomic species distributional database for any tropical con-tinent Hereafter we refer to these species collectively asldquoAfrican terrestrial vertebratesrdquo for convenience althoughwe have not been able to include lizards in our database

For the larger and better-known species the data are esti-mates of recent distributions which we have taken fromstandard compilations and modified using the specialist lit-erature where necessary For smaller and less well-knownspecies we interpolated expected distributions by assuminga continuous distribution between confirmed records withinrelatively uniform suitable habitat using available informa-tion on speciesrsquo habitat associations and taking care to excludeknown gaps in distribution We checked interpolation byconsulting taxon specialists (all sources are available at the Website wwwzmucdkcommonwebresearchblueprint-africahtm)For the least well-known species records are plotted withoutinterpolation Figure 1 shows examples of each of these mapsOver half of the maps are primary sources that have never beenpreviously published or analyzed We recognize that suchbinary maps represent varying degrees of simplification ofactual distributions (Freitag et al 1996) but the urgency of

Figure 1 Examples of maps from our databases (a) Orycteropus afer a widespread large mammal We map recent rangebased on published maps modified with numerous other data (b) Laniarius funebris a bush-shrike restricted to EastAfrican drylands Our data follow generalized range maps (c) Leptotyphlops scutifrons a burrowing snake We interpolatebetween known records within (but not outside) suitable habitat and range (d) Geotrypetes seraphini a poorly knownamphibian Only known records are mapped Throughout the bar represents 10 degrees of latitude and the arrow indicatesnorth

614 BioScience bull August 2001 Vol 51 No 8

Articles

vation planning The last issue concerns consensus buildingThe databases and analyses should become available to on-going regional national and local conservation planningprocesses in Africa through workshops and collaborativework The consensus built by these processes is essential forensuring that any of these studies mentioned above results inthe effective conservation of biodiversity

The distributional databasesWe compiled data on the distribution of all currently recog-nized 1921 species of birds 940 species of mammals 406species of snakes and 615 species of amphibians found inmainland sub-Saharan Africa (south of 20 degrees North)onto a one-degree grid with each cell approximately 105km on each sideWe chose the one-degree resolution as a com-promise between the sampling inadequacy of finer resolutionsand the loss of biogeographic detail (especially in moun-tainous regions) at coarser resolutions (Rahbek and Graves2000) For the 1957 one-degree grid cells across the continentholding data we currently have a total of 828506 species-in-grid-cell data entries for the 3882 species overall This is a dy-namic database into which new grid-cell data entries are be-

ing added almost daily it is the most complete cross-taxonomic species distributional database for any tropical con-tinent Hereafter we refer to these species collectively asldquoAfrican terrestrial vertebratesrdquo for convenience althoughwe have not been able to include lizards in our database

For the larger and better-known species the data are esti-mates of recent distributions which we have taken fromstandard compilations and modified using the specialist lit-erature where necessary For smaller and less well-knownspecies we interpolated expected distributions by assuminga continuous distribution between confirmed records withinrelatively uniform suitable habitat using available informa-tion on speciesrsquo habitat associations and taking care to excludeknown gaps in distribution We checked interpolation byconsulting taxon specialists (all sources are available at the Website wwwzmucdkcommonwebresearchblueprint-africahtm)For the least well-known species records are plotted withoutinterpolation Figure 1 shows examples of each of these mapsOver half of the maps are primary sources that have never beenpreviously published or analyzed We recognize that suchbinary maps represent varying degrees of simplification ofactual distributions (Freitag et al 1996) but the urgency of

Figure 1 Examples of maps from our databases (a) Orycteropus afer a widespread large mammal We map recent rangebased on published maps modified with numerous other data (b) Laniarius funebris a bush-shrike restricted to EastAfrican drylands Our data follow generalized range maps (c) Leptotyphlops scutifrons a burrowing snake We interpolatebetween known records within (but not outside) suitable habitat and range (d) Geotrypetes seraphini a poorly knownamphibian Only known records are mapped Throughout the bar represents 10 degrees of latitude and the arrow indicatesnorth

systematic conservation across Africa is too great to wait un-til better data are available (van Jaarsveld et al 1998a)

For mammals our taxonomy follows Wilson and Reeder(1993) modified according to recent species descriptionsand taxonomic opinions For the Galagonideae we followthe species names order and distributional data presented byKingdon (1997) For the larger mammal species we pro-duced range maps using standard references for the area(Dorst and Dandelot 1970 Haltendorf and Diller 1977 Skin-ner and Smithers 1990) We refined these using numeroussources including action plans and unpublished museum datafrom the International Union for the Conservation of Natureand Natural ResourcesmdashSpecies Survival Commission Forthe smaller species we compiled maps from the literature(hundreds of references) and discussed maps with dozens ofspecialists

For birds our taxonomy follows Sibley and Monroe (19901993) which in spite of some problems (Mayr and Bock1994) is the most recent and comprehensive treatment of theworldrsquos avifauna We include all terrestrial species and wa-terbirds that breed in the Afrotropics or that regularly visit thisregion as nonbreeding migrants but exclude pelagic and va-grant species (Dowsett and Dowsett-Lamaire 1993 Dowsettand Forbes-Watson 1993) We compiled base distributionmaps from the published volumes of the The Birds of Africa(Brown et al 1982 Urban et al 1986 1997 Fry et al 1988Keith et al 1992 Fry and Keith 2000) These base maps werethen checked refined and completed for unmapped speciesusing four major sources (Hall and Moreau 1962 1970 Snow1978 Stattersfield et al 1998) and dozens of other publishedreferences (including recent atlases) and unpublished re-search

For snakes we based our species list on Welch (1982)with the addition of newly described species and recent tax-onomic reinterpretations Data for species maps were com-piled from an extensive review of the literature and visits tomany museums We used range maps taken from publishedcompilations for poisonous snakes (Sprawls and Branch1996) and for southern Africa (Branch 1998) For amphib-ians our taxonomy follows Frost (1985) and Duellman(1983) We updated this list where necessary using recentlypublished papers and prevailing taxonomic opinions Data forthe maps come from a thorough literature review along withunpublished data provided by numerous specialists

Biogeographic patterns of Africanterrestrial vertebratesWe used WORLDMAP (Williams 1996) to assess richness en-demism and complementarity within our data Stacking thespecies maps on top of each other provides us with richnessmaps (Figure 2andashd) which show the variation in numbers ofspecies within each of the four groups across the continentBroadly areas with higher levels of energy-related variablessuch as primary productivity potential evaoptranspirationsolar radiation temperature and rainfall tend to have higherspecies richness as expected (Currie 1991 but see Rosenzweig

and Sandlin 1997) More specifically mammals are mostspeciose in the forest-dotted crescent around the north of LakeVictoria Outliers of particular richness run south through themountains of Kenya Tanzania and Malawi and west throughthe forestndashsavanna ecotone north of the Congo Basin into theGuinea forests Areas of lowest richness lie in the Sahel theHorn of Africa and the Kalahari Namib and Karoo drylandsBirds show a richness pattern remarkably similar to that ofmammals with the main differences lying in the relativelymore speciose miombo (Brachystegia Julbernardia andIsoberlinia) woodland of Zambia and in the relatively poorernorthern savannandashrain forest ecotone Snakes show a differ-ent pattern The Zambezian region contains the most specieswith large numbers concentrated also along the lower CongoRiver and in southern Cameroon the Albertine Rift thecoast of Kenya and Namibia The major drylands again havethe lowest richness but much of Ethiopia Tanzania andAngola are also depauperate Amphibians are most speciosein the Lower Guinea forests especially in Cameroon The Al-bertine Rift the Eastern Arc and the southern African forestsalso hold many species while northern Kenya and northernTanzania as well as the major drylands hold rather few Thecentral portion of the Congolian rain forest is also quite de-pauperate for all four taxa

One important limitation of the maps is sampling effortThis is particularly obvious for the snakes and amphibiansas indicated by the spotty appearance of their richness mapsOften the presence of spots of particularly high species rich-ness surrounded by apparently depauperate areas merely in-dicates that only a single locality has been surveyed This ef-fect is noticeably variable geographically an obvious regionthat has been only poorly sampled is northern Mozambique(Ryan et al 1999) However although this problem should notbe downplayedmdashand indeed we are taking steps to solve it bysupporting both deductive environmental modeling of speciesranges and the collection and compilation of new data (daFonseca et al 2000)mdashwe are confident that it is small relativeto the overall size of our databases

While patterns of species richness appear to be drivenlargely by current environmental conditions within geo-metric constraints (Currie 1991 Colwell and Lees 2000)patterns of species endemismmdashthe distributions of specieswith small rangesmdashmay be considerably more informative forinterpreting the historical evolutionary process (Kingdon1989 Fjeldsaring and Lovett 1997) Furthermore range restrict-edness may well be of key importance for conservation (Halland Moreau 1962) because species with small ranges tend alsoto have small population sizes (Brown 1984) and therefore in-creased risks of extinction on both counts (Pimm and Brooks2000) Following this logic Stattersfield et al (1998) mappedthe distributions of all bird species with ranges of less than50000 sq km to identify the ldquoendemic bird areasrdquo of theworld A common criticism of this approach is that it omitsspecies with ranges slightly largely than the arbitrary cutoff(Peterson and Watson 1998) To avoid this problem an al-ternative technique is to map ldquorange-size rarityrdquo as a contin-

August 2001 Vol 51 No 8 bull BioScience 615

Articles

uous variable for example by summing the reciprocals of therange sizes of all species in each cell so that all species in thedata make some contribution to the scores (Williams et al1996)

Figure 2endashh shows maps of range-size rarity In general aswith species richness the most endemic-rich areas lie in the

montane forests while the most en-demicpoor are in drylands and hy-drologically unstable plains For mam-malsmdashand similarly for frogsmdashspecieswith small ranges are concentrated inmontane forest around the equatorin the Guinea forests the AlbertineRift the Kenya highlands the EasternArc and Ethiopia The Sahara and theSahel are endemic poor but other dry-lands including the Horn of Africaand the Karoo are rather rich En-demic birds are similarly concentratedin the highlands of East Africa butwith relatively fewer species with smallranges in West Africa and more quitenoticeably along the Angola scarp Theextensive dry interior of southwesternAfrica is noticeably endemic poor Pat-terns of endemism in snakes are lesswell defined but areas holding nu-merous small-ranged species can bediscerned in Cameroon the AlbertineRift the Tanzanian coast and (per-haps surprisingly) the Horn of AfricaMuch of southern central Africa is en-demic poor for snakes but southernAfrica is quite rich

The richness and endemism mapscan be compared using a two-colortechnique (Williams and Gaston 1998)as shown in Figure 3 Mountainousequatorial regions are consistently highin both species richness and endemismwhereas the Sahara and KalahariDeserts have both low richness andlow endemism There tends to be dis-proportionately high endemism rela-tive to species richness in the UpperGuinea forests the Horn of Africa andthe Cape and Karoo In contrast theSahel and Zambezian lowlands havedisproportionately low endemismgiven their large numbers of species

Identifying priority areasfor conservationHow can we use these patterns of bio-geography to assess which areas arethe highest priorities for conservation

Because competing interests constrain the area available forconservation the issue can be viewed as a question of repre-sentation Given a fixed number of priority areas (say 50one-degree grid cells which is 25 of sub-Saharan Africarsquosland surface) we can seek a combination of areas to maximizeour representation of conservation unitsmdashin this case ter-

616 BioScience bull August 2001 Vol 51 No 8

Articles

Figure 2 Patterns of species richness (andashd) and endemism (endashh) in sub-SaharanAfrican vertebrates The most speciose and endemic-rich areasmdashgenerally thehighland forests of Lower Guinea the Albertine Rift upland Kenya the Eastern Arcand southern Africamdashare shown in red and the least speciose and most endemicpoormdashgenerally the drylands of the Sahel the Horn of Africa and southwesternAfricamdashare in blue All maps are on an equal frequency scale

restrial vertebrate species This is a ldquomax-imal coveringrdquo problem (Church et al1996)As a baseline against which to com-pare solutions to this problem we deter-mined how many species in each majortaxon would on average be represented in50 grid cells selected at random from themaps We calculated average representa-tions (plusmn25 tail) random selections ofcells will only achieve representations ofnumbers of species outside of this range5 of the time These random selectionsare given in Table 1

Perhaps the simplest method for se-lecting areas that might maximize speciesrepresentation would be to choose thosecells that hold the greatest numbers ofspecies These cells tend to be clustered ge-ographicallymdashfor mammals birds andsnakes in the Albertine Rift and for frogsin the Cameroon highlands (Figure 2andashd)As a result they tend to hold similar setsof species (Williams et al 2000) and soour total representation within sets of the50 richest cells for each group is there-fore poor (Table 1) For birds it is actuallyno better than choosing cells at randomA second alternative might be to selectareas that hold large concentrations ofspecies with small ranges The moststraightforward method for doing this isto select the grid cells that have the high-est range-size rarity scores (Figure 2endashh)These areas are less clustered than those forspecies richness As a result selecting the50 most endemic-rich cells achieves a bet-ter representation of all four major taxa

August 2001 Vol 51 No 8 bull BioScience 617

Articles

Table 1 The success of four methods for representing the terrestrial vertebrate species of sub-Saharan Africa in sets of 501-degree grid cells measured by absolute and percentage representation of species The methods are (1) choosing randomcells (plusmn25 tail) (2) choosing the richest cells (3) choosing the cells with the highest range-size rarity scores (the sum ofthe reciprocals of the range sizes of all species in the cell) and (4) greedy complementarity (choosing cells that progressive-ly add the greatest number of species to those already represented)

Method of representing cells Mammals Birds Snakes Amphibians

Species total 940 1921 406 615

50 cells chosen at random 525 plusmn 58 1515 plusmn 105 199 plusmn 30 225 plusmn 5556 plusmn 6 79 plusmn 5 49 plusmn 7 37 plusmn 10

50 richest cells 614 1414 249 39365 74 56 63

50 cells with highest range-size rarity 709 1554 320 49075 81 71 79

50 cells in greedy complementary set 841 1877 375 55289 98 84 89

Figure 3 A two-color technique for illustrating deviations from the overallrelationship between species richness and range-size rarity (Williams and Gaston1998) (a) Mammals (b) birds (c) snakes and (d) amphibians Increasing speciesrichness is mapped in blue and overlaid onto increasing range-size raritymapped in green A matched intensity of red is then added to this diagonal toproduce a neutral scale of whitendashgreyndashblack along which species richness andendemism positively covary Thus areas of high species richness and endemismare shown in white and areas of low richness and endemism in black In contrastdeviations from the overall positive linear relationship between richness andendemism are shown in increasingly saturated color with blue indicating anldquoexcessrdquo of richness and green an ldquoexcessrdquo of endemism Generally richness andendemism are both high on the equatorial mountains (white) and both low inthe deserts (black) Endemism tends to be disproportionately high in the UpperGuinea forests the Horn of Africa and the Cape and Karoo and low in the Saheland Zambezian lowlands

than does selecting the 50 richest cells (Table 1) Neverthelesswe can still do much better than this (Pressey and Nicholls1989) by considering the principle of complementarity

The idea of complementarity is to select a set of conser-vation areas that togethermdashrather than individuallymdashcon-tribute a maximum number of species towards the overall goalof representation (Pressey et al 1993) We have already usedthe first step in this process which is to select the grid cell thatholds the largest individual number of species In comple-mentary area selection however we then discount all speciesrepresented within this first cell and choose as our second cellthe one holding the largest number of species unrepresentedin the first cell Then we discount all species representedwithin this cell to direct us to where our third cell lies and soon until we reach our target We illustrate the first three stepsof this process for mammals in Figure 4

Repeating these steps 50 times for each major taxon givesus results that we can compare with our other selection meth-ods (Table 1) Clearly complementarity gives us an extremelyefficient answer to the question of where to site a represen-tative conservation network This technique is termed greedycomplementarity because it selects the area holding the largestnumber of unrepresented species at every step Numerousother complementary methods exist that achieve similar lev-els of efficiency but select areas different from those selectedby greedy complementarity (Csuti etal 1997)

An extension of the maximal cov-ering problem is to aim for a species-based rather than an area-based targetThus for example we could seek torepresent all of our species at leastonce in as few grid cells as possible anear-minimum set of areas To do thiswe simply repeat the steps illustratedin Figure 4 until we have representedevery species in each taxon We illus-trate this process in Figure 5 with therepresentation at each step by greedycomplementarity contrasted for thesame number of selections of cells atrandom (plusmn25 tail) The total num-ber of cells required to select everyspecies varies from 81 for snakes and87 for birds up to 109 for amphibiansand 127 for mammals

Interestingly birds despite beingthe most speciose of the four taxa canbe fully represented in fewer cells thanmost groups This can be attributed tothe fact that a smaller proportion ofbird species have very restricted knownranges than do the other three taxa(Figure 6) Thus birds in our databasehave a median range size of 144 cellsmammals 33 snakes 14 and amphib-

ians only 10 This effect is probably caused by three factorsthe generally larger range sizes of birds (a result of theirgreater dispersal abilities) our better knowledge of their dis-tributions and our greater use of continuous range maps torepresent their distributions in our databases The reasonfewer cells are required to represent all snakes than to repre-sent all birds despite the much greater representation ofbirds in small sets of areas is that there are so many fewer snakespecies overall than there are birds

Incorporating threatAlthough the representation approach to selecting conser-vation targets gives a set of priority areas it does not provideany information about the urgency of conservation in any par-ticular area (other than to suggest how many species each areacontributes toward the overall goal) For priority areas tomaintain their representation of species into the future theprocesses threatening these species must be taken into account(Nicholls 1998) This issue has not been extensively addressedto date but one initial way to tackle it is to prioritize consid-eration of red-listed species that are threatened with a ldquohighprobability of extinction in the wild in the medium-termfuturerdquo (Baillie and Groombridge 1996 annex p 9) This ap-proach is limited by a lack of data As with distributional in-formation the conservation assessment of most groups has

618 BioScience bull August 2001 Vol 51 No 8

Articles

Figure 4 Area selection using greedy complementarity for mammals (a) The firststep is to select the area holding the largest number of species the Ruwenzorimountains (b) All of the species occurring within this area are then subtractedfrom the database leaving a richness map of species not occurring in theRuwenzoris The new richest areamdashMount Kilimanjaromdashis then selected (c) Nextall the species occurring on Kilimanjaro are discounted The area with the greatestcomplementary richness and hence the next area for selection is Mount Nimba (d)Subtracting all the species occurring here indicates that the next area for selection isMount Cameroon We repeat this process until the conservation target is achieved

been very patchy Thus in Africa onlythree species of snakes and 11 speciesof amphibians have been listed asthreatened of which all but two (Nectophrynoides occidentalis andConraua goliath from the Guineaforests) are from South Africa (Bail-lie and Groombridge 1996)

The conservation status of mam-mals (Baillie and Groombridge 1996)and birds (Collar et al 1994) how-ever has been comprehensively as-sessed In Figure 7 we map the rich-ness of threatened African mammalsand birds The differences from Fig-ure 2andashdmdashthe richness maps for allspeciesmdashare quite striking Figure 7is instead more similar to Figure2endashhmdashthe endemism mapsmdashnec-essarily because small range size isone of the criteria for inclusion onthe red list (Baillie and Groombridge1996) For mammals the Horn ofAfrica coastal Namibia and CapeFynbos stand out as holding manythreatened species for birdsEthiopia the Angola Scarp and theSouth African highveld hold dispro-portionate numbers of threatenedspecies In addition the Guineaforests the Albertine Rift and theEastern Arc hold numerous threat-ened species

We can identify important areasfor the conservation of these threat-ened species using the same com-plementarity method that we usedfor all species To represent all 184threatened sub-Saharan Africanmammal species requires 68 one-degree grid cells whereas the 115threatened birds can be representedin only 41 cellsWe can then use theseareas as a top-priority baseline bydetermining the minimum numberof additional cells required to repre-sent all species of mammals and do-ing the same for birds It turns outthat we require only 56 additionalcells to represent all species of mam-mals and only 49 more for birdsThis means that selecting areas torepresent all species starting withthe threatened species is not sub-stantially less efficient a method thanselecting areas to represent species

August 2001 Vol 51 No 8 bull BioScience 619

Articles

Figure 5 Accumulation of greedy complementary sets of cells to represent all species in each major taxon (a) mammals (b) birds (c) snakes and (d) amphibians Wecompare the accumulation curves (red) with those achieved through the randomselection (green) of the same number of cells (plusmn 25 tail) The total number of cellsrequired to select every species varies from 81 for snakes to 127 for mammals

Figure 6 Proportionate frequency distributions of range sizes (numbers of 1-degreecells occupied) for sub-Saharan African vertebrates (a) mammals (b) birds (c)snakes and (d) amphibians The frequency distribution for birds is much less skewedthan for the other taxamdashmany more birds have larger ranges

without regard for their conservation status For mammals theformer method requires 124 cells and for the latter 127 forbirds the former method requires 87 cells and the latter 90

An alternative method for incorporating threat would beto map some measure (such as habitat destruction) of theprocesses threatening biodiversity across Africa The resultcould be integrated into a geographic map of conservation pri-orities (Figure 8) identifying the top priorities as those highlythreatened areas that make important contributions to a rep-resentative complementary set (Faith and Walker 1996a)Such data are becoming increasingly available for examplethrough the Global Land Cover Characterization (Web sitehttpedcdaacusgsgovglccglcchtml)

Which areas are the highest priorities for conservation inAfrica Table 2 gives the top 10 areas selected in comple-mentary sets to represent threatened mammals and birds andalso to represent all birds mammals snakes and amphibiansThe sets represent between half (threatened mammals) andfour-fifths (all birds) of the groups Most areas fall within themuch broader priority areas identified by various conserva-tion organizations (Mittermeier et al 1998 Olson and Din-erstein 1998 Stattersfield et al 1998) However even withinjust these six 10-cell sets no fewer than 28 of the 93 mainlandsub-Saharan biogeographic ecoregions (Underwood et al1999) are represented The most surprising of these are areasin central Angola (for mammals) Eritrea and Gambia (forbirdsmdasha result that is driven by the presence in these north-ern areas of overwintering and passage migrants) and the(Colophospermum) mopane woodland of northern Swazi-land the lowland forest of central Uganda and again Eritrea(for snakes) Nevertheless similar areas are selected againand againmdashthe Eastern Arc Guinea forests (especially MountCameroon) Cape FynbosAlbertine Rift and Ethiopian high-lands All of the areas listed in Table 2 are mapped in Figure8 They also all occur in a greedy near-minimum set for allspecies combined These are the highest continental-scalepriorities for conserving the diversity of African terrestrial ver-tebrates

Current and futuredirectionsThis study finds that five regionsmdashthe Eastern Arc Upper and LowerGuinea Cape Fynbos AlbertineRift and Ethiopian highlandsmdashemerge as the top conservation pri-orities across Africa Several fac-tors could bias this result Our datacould be misleading because of po-tentially inaccurate extrapolationsof the distributions of mammalsand birds or conversely because oflack of extrapolation for poorlyknown species The restriction ofour data set to terrestrial verte-brates could mean that we missimportant areas for other taxa A

third possibility is that threat to mammals and birds is apoor surrogate for threat to overall biodiversity We also donot incorporate flexibility into our area selection whichcould alter conservation priorities within homogenous regionsNevertheless that the same five regions are repeatedly se-lected as the highest priorities for the representation of dif-ferent taxamdashincluding threatened taxamdashstrongly suggeststhat these influences are small

What then are the next steps for conservation planning inAfrica We can discern five critical needs The first is to ad-dress the question of how well conservation priorities forthe groups with relatively good distributional data reflectthose for other groups and indeed for biodiversity as a wholeThis has been an issue of considerable debate At coarsescales there is high congruence between groups (Myers et al2000) At fine scales there appears to be little congruence be-tween either diversity patterns (Lawton et al 1998) or con-servation priorities (van Jaarsveld et al 1998b) However ev-idence from Uganda indicates that representative areas formost major taxa capture other groups very well indeed(Howard et al 1998) The verdict is still out however (Burgesset al 2000) If it emerges that few taxa are indicators for bio-diversity more generally an alternative strategy may be to relyon environmental surrogates (Faith et al 1996) Ideally suchapproaches to habitat representation approaches would becombined with species representation to avoid the loss ofparticular highly valued species

A second and related point is that much greater effort is re-quired in both data collection and compilation Clearly thecompleteness of both will affect the selection of conservationpriorities to a great degree (Freitag and van Jaarsveld 1998)Even for well-known taxa much work remains Witness theabsence of comprehensive assessments of the conservation sta-tus of African snakes and amphibians as well as the geo-graphic biases in data collection Further testing the idea ofcongruence properly requires data on taxa not coveredhere A particularly acute need is for the compilation of com-parable distribution data for plants because conservation

620 BioScience bull August 2001 Vol 51 No 8

Articles

Figure 7 Patterns of richness of threatened sub-Saharan African vertebrates (a)mammals and (b) birds The most speciose areasmdashgenerally the Guinea forests theAlbertine Rift the Eastern Arc Ethiopia and South Africamdashare shown in red and theleast speciosemdashgenerally the drylands of the Sahel and southwestern Africa and thecentral African lowland rain forestsmdashare in blue

without regard for their conservation status For mammals theformer method requires 124 cells and for the latter 127 forbirds the former method requires 87 cells and the latter 90

An alternative method for incorporating threat would beto map some measure (such as habitat destruction) of theprocesses threatening biodiversity across Africa The resultcould be integrated into a geographic map of conservation pri-orities (Figure 8) identifying the top priorities as those highlythreatened areas that make important contributions to a rep-resentative complementary set (Faith and Walker 1996a)Such data are becoming increasingly available for examplethrough the Global Land Cover Characterization (Web sitehttpedcdaacusgsgovglccglcchtml)

Which areas are the highest priorities for conservation inAfrica Table 2 gives the top 10 areas selected in comple-mentary sets to represent threatened mammals and birds andalso to represent all birds mammals snakes and amphibiansThe sets represent between half (threatened mammals) andfour-fifths (all birds) of the groups Most areas fall within themuch broader priority areas identified by various conserva-tion organizations (Mittermeier et al 1998 Olson and Din-erstein 1998 Stattersfield et al 1998) However even withinjust these six 10-cell sets no fewer than 28 of the 93 mainlandsub-Saharan biogeographic ecoregions (Underwood et al1999) are represented The most surprising of these are areasin central Angola (for mammals) Eritrea and Gambia (forbirdsmdasha result that is driven by the presence in these north-ern areas of overwintering and passage migrants) and the(Colophospermum) mopane woodland of northern Swazi-land the lowland forest of central Uganda and again Eritrea(for snakes) Nevertheless similar areas are selected againand againmdashthe Eastern Arc Guinea forests (especially MountCameroon) Cape FynbosAlbertine Rift and Ethiopian high-lands All of the areas listed in Table 2 are mapped in Figure8 They also all occur in a greedy near-minimum set for allspecies combined These are the highest continental-scalepriorities for conserving the diversity of African terrestrial ver-tebrates

Current and futuredirectionsThis study finds that five regionsmdashthe Eastern Arc Upper and LowerGuinea Cape Fynbos AlbertineRift and Ethiopian highlandsmdashemerge as the top conservation pri-orities across Africa Several fac-tors could bias this result Our datacould be misleading because of po-tentially inaccurate extrapolationsof the distributions of mammalsand birds or conversely because oflack of extrapolation for poorlyknown species The restriction ofour data set to terrestrial verte-brates could mean that we missimportant areas for other taxa A

third possibility is that threat to mammals and birds is apoor surrogate for threat to overall biodiversity We also donot incorporate flexibility into our area selection whichcould alter conservation priorities within homogenous regionsNevertheless that the same five regions are repeatedly se-lected as the highest priorities for the representation of dif-ferent taxamdashincluding threatened taxamdashstrongly suggeststhat these influences are small

What then are the next steps for conservation planning inAfrica We can discern five critical needs The first is to ad-dress the question of how well conservation priorities forthe groups with relatively good distributional data reflectthose for other groups and indeed for biodiversity as a wholeThis has been an issue of considerable debate At coarsescales there is high congruence between groups (Myers et al2000) At fine scales there appears to be little congruence be-tween either diversity patterns (Lawton et al 1998) or con-servation priorities (van Jaarsveld et al 1998b) However ev-idence from Uganda indicates that representative areas formost major taxa capture other groups very well indeed(Howard et al 1998) The verdict is still out however (Burgesset al 2000) If it emerges that few taxa are indicators for bio-diversity more generally an alternative strategy may be to relyon environmental surrogates (Faith et al 1996) Ideally suchapproaches to habitat representation approaches would becombined with species representation to avoid the loss ofparticular highly valued species

A second and related point is that much greater effort is re-quired in both data collection and compilation Clearly thecompleteness of both will affect the selection of conservationpriorities to a great degree (Freitag and van Jaarsveld 1998)Even for well-known taxa much work remains Witness theabsence of comprehensive assessments of the conservation sta-tus of African snakes and amphibians as well as the geo-graphic biases in data collection Further testing the idea ofcongruence properly requires data on taxa not coveredhere A particularly acute need is for the compilation of com-parable distribution data for plants because conservation

620 BioScience bull August 2001 Vol 51 No 8

Articles

Figure 7 Patterns of richness of threatened sub-Saharan African vertebrates (a)mammals and (b) birds The most speciose areasmdashgenerally the Guinea forests theAlbertine Rift the Eastern Arc Ethiopia and South Africamdashare shown in red and theleast speciosemdashgenerally the drylands of the Sahel and southwestern Africa and thecentral African lowland rain forestsmdashare in blue

priorities for plantsmdashfor example in the Succulent Karoomdashmay not be well reflected by vertebrates and vice versa (forexample in the Albertine Rift) (Myers et al 2000) This workis now under way (Lovett et al 2000) Clearly data for fresh-water and marine conservation planning are also urgently re-quired and ultimately should be meshed with terrestrial datato inform an integrated conservation strategy

Third an important research front involves the use of en-vironmental data often derived from remote sensing to buildmodels of expected species distributions These techniquesprovide an indirect method to fill sampling gaps in species dis-tributions In addition they have the potential to greatly in-crease the resolution of species distributions This research isdeveloping along two parallel lines of inquiry one involvingthe use of expert opinions regarding speciesrsquo habitat prefer-ences to construct deductive models the other necessitatingthe compilation of point locality data to build inductivemodels (Corsi et al 2000) Environmental models can also pro-vide information on probability of occurrence rather than justpresence which is important to address threats to speciesrsquo vi-ability and persistence (Lawes and Piper 1998) These ap-proaches have already been applied with great promise to largeAfrican mammals (Boitani et al 1999)

A fourth area for urgent attention is engaging the social sci-ences to provide useful socioeconomic data of the same res-olution and quality as biodiversity data There are some en-couraging signs that such collaborations are improving at leastin temperate regions Natural resource managers have al-ready begun to compile protected area data in formats thatcan be combined with biodiversity data to enable ldquogap analy-sisrdquo to identify those species that fall through holes in the cur-rent protected-areas network (Scott et al 1993) In Africanational-level gap analyses have been conducted (Lombard1995 Howard et al 2000) but no analysis has been carriedout at the continental level at which species representationin protected areas is probably quite high Meanwhile envi-ronmental economists have developed methods for buildingland values into priority-setting analyses (Faith and Walker1996b Ando et al 1998) Nevertheless numerous other so-cioeconomic factors critical to the outcome of conservationprograms have yet to be addressed especially in the tropics(Singh 2000) Balmford and Long (1994) and Fjeldsaring and Rahbek (1998) for example pointed out that areas of highavian endemism also hold dense human populations andrapid rates of habitat loss If this is the case human popula-tion density and growth rates must be explicitly factored into

August 2001 Vol 51 No 8 bull BioScience 621

Articles

Figure 8 Geographic map showing the 42 grid cell areas and 28 ecoregions highlighted in Table 2 asconservation priorities for sub-Saharan African vertebrates

conservation priority setting Other examples of importantmdashand often sensitive and contentiousmdashparameters include thedistributions of military conflict refugee movements timberand mining concessions commodity production bushmeathunting and the narcotics trade Broadly some methods arealready in place to deal with all of these factors once they havebeen adequately quantified

Finally and most crucial of all the results of conservationpriority-setting exercises need to be put into action Data-driven analysis such as that presented here is increasinglyproviding answers on continental scales The challenge remainsto move this analysis to the national and local scales of con-servation implementation (da Fonseca et al 2000) Onepromising strategy for doing so is through workshops that pro-vide local participants with rigorous quantitative biodiversitydata to forge a consensus on specific conservation priorities

and actions (Mittermeier et al 1995) Successful workshopsof this nature were held in December 1999 in Ghana to ad-dress the Upper Guinea region and in March 2000 in Gabonto address the Congo Basin Only through such local owner-ship of the fine-scale planning process can effective and effi-cient conservation in Africa be implemented

AcknowledgmentsMany thanks to P Agnelli J L Amiet the late W F H AnsellE Baker N Baker P Bates S Bearder W Bergmans L Boi-tani B Branch D Broadley G Bronner C Cabral M D Car-leton A Channing C Claessen M Colyn W Cotterill GCowlishaw C Dardia G Davies H M de Klerk J-P drsquoHuartF Dieterlen N Dippenaar B Dowsett F Dowsett-LemaireJ M Duplantier J Fahr M B Fenton C Fitzgibbon CGans L Granjon P Grubb T Halliday D C D Happold R

622 BioScience bull August 2001 Vol 51 No 8

Articles

Table 2 Conservation priorities for sub-Saharan African terrestrial vertebrate species Each list gives the first 10 areasselected in greedy complementary sets for the group in question We also give the cumulative percentage of species withinthe group represented by each arearsquos inclusion in the overall set The most frequently selected areas lie in five geographicregions the Eastern Arc Forest the forests of Upper and Lower Guinea the Cape Fynbos the Albertine Rift and theEthiopian Highlands Biogeographic regional names (in italics) follow Underwood et al (1999)

Prioritythreatenedmammals Threatened birds All mammals All birds All snakes All amphibians

1 Usambaras 10 N Udzungwas 11 Ruwenzori 23 Ruwenzori Foothills 36 SE Congo 18 Mt Cameroon 19Eastern Arc Forest Eastern Arc Forest Albertine Rift Montane Forest Albertine Rift Montane Forest West Congolian Forest-Savanna Cameroonian Highlands Forest

2 Mt Cameroon 19 Mt Cameroon 20 Kilimanjaro 32 Usambaras 50 N Swaziland 29 S Udzungwas 29Cameroonian Cameroonian Highlands East African Montane Forest Eastern Arc Forest Zambezian and Mopane Woodland Eastern Arc Forest Highlands Forest

3 S S Afr 24 Mt Nimba 27 Mt Nimba 40 Eritrea 57 Mt Cameroon 36 Mt Nimba 35Montane Fynbos Guinean Montane Guinean Montane Ethiopian Xeric Grassland Cameroonian Highlands Forest Guinean Montane Forestand Renosterveld Forest Forest and Shrubland

4 NE DR Congo 29 Ruwenzori 34 Mt Cameroon 45 Soutpansberg 63 Sokoke 41 Upemba Plain 40Northeastern Congolian Albertine Rift Cameroonian S Africa Bushveld N Zanzibar-Inhambane Central Zambezian Lowland Forest Montane Forest Highlands Forest Coastal Forest Miombo Woodland

5 N Bale 34 Drakensberg 40 Cape Town 50 Mt Cameroon 68 Namib Scarp 46 Ruwenzori Foothills 45Ethiopian Montane Drakensberg Montane Lowland Fynbos and Cameroonian Highlands Namibian Savanna Albertine Rift Montane Grassland and Grassland Woodland Renosterveld Forest Woodland ForestWoodland and Forest

6 Mau 38 Yabello 46 Vumba 54 Mt Moco 71 Upemba Plain 50 E S Afr 49East African Montane Somali Acacia-Commiphora E Zimbabwe Montane Angola Montane Central Zambezian Miombo Drakensberg MontaneForest Bushland and Thicket Forest-Grassland Forest-Grassland Woodland Grassland Woodland

and Forest

7 NW Somalia 42 S Itombwe 51 Awash 57 Mt Elgon 74 Usambaras 53 Cape Town 53Somali Montane Albertine Rift Montane Forest Ethiopian Montane Forest East African Montane Forest Eastern Arc Forest Lowland Fynbos and Xeric Woodland Renosterveld

8 N Tai 45 Gabela 56 Mid-Angola 59 Cape Town 76 Accra 56 S Bale 56Western Guinean Angola Montane Forest- Angolan Miombo Woodland Lowland Fynbos and Eastern Guinean Lowland Ethiopian Montane Lowland Forest Grassland Renosterveld Forest Grassland and Woodland

9 Mt Oku 48 Mombasa 59 N Itombwe 62 Ankober 78 Eritrea 59 Usambaras 58Cameroonian Highlands North Zanzibar- Albertine Rift Montane Ethiopian Montane Ethiopian Xeric Grassland and Eastern Arc Forest ForestForest Inhambane Coastal Forest Forest Grassland and Woodland Shrubland

0 N Itombwe 50 Nyungwe 62 NE DR Congo 64 Gambia 80 Kampala 61 Mulanje 60Albertine Rift Montane Albertine Rift Montane Northeastern Congolian West Sudanian Savanna Victoria Basin Forest-Savanna South Malawi MontaneForest Forest Lowland Forest Mosaic Forest-Grassland

1

Hoffman M E Holden P Howard B Hughes R HuttererN H G Jacobsen P Jenkins W Jetz J Kerbis D Koch A JL Lamberis M Languy M Largen H Leirs A Linzey JMutke J Oates B Patterson J-L PerretA Plumtre J C Poyn-ton P A Racey J B Rasmussen G B Rathburn L RobbinsV SalewskiA Schioslashtz D SchlitterA M Simonetta J D Skin-ner S Spawls W Stanley M E Taylor P Taylor V van Cak-enberghe E van der Straeten E van Dijk H van RompaeyW Verheyen V Wallach N Winser R Wirth D Yalden andBirdLife International for providing data to S Galster AJakobsen J B Larsen T Lehmberg P V Nielsen T S Rom-dal L L Soslashrensen and M M Westergaard for entering datato A Andaryati A Channing R Gutberlet D Moyer and HS Thompson for providing photographs and to H M deKlerk D Faith D C D Happold W Jetz J Lovett S LPimm A Plumtre W Sechrest E Underwood M Wisz andthree anonymous reviewers for help with the manuscriptFunding was provided by the Danish Research Council theDanish Council for Development Research the Isaac NewtonTrust of the University of Cambridge and Conservation In-ternationalrsquos Center for Applied Biodiversity Science

References citedAndo A Camm J Polasky S Solow A 1998 Species distributions land val-

ues and efficient conservation Science 279 2126ndash2128Baillie J Groombridge B 1996 1996 IUCN Red List of Threatened Animals

Gland (Switzerland) International Union for the Conservation of Na-ture and Natural Resources

Balmford A Long A 1994 Avian endemism and forest loss Nature 372623ndash624

Boitani L Corsi F De Biase A DrsquoInzillo Carranza I Ravagli M Reggiani GSinibaldi I Trapanese P 1999 AMD African Mammals Databank ADatabank for the Conservation and Management of the African Mam-mals Rome Istituto di Ecologia Applicata

Branch B 1998 Field Guide to the Snakes and other Reptiles of SouthernAfrica 3rd ed Cape Town (South Africa) Struik

Brown JH 1984 On the relationship between abundance and distributionof species American Naturalist 124 255ndash279

Brown LH Urban EK Newman K 1982 The Birds of AfricaVol 1 LondonAcademic Press

Burgess N de Klerk H Fjeldsaring J Crowe T Rahbek C 1997 Mapping Afrotrop-ical birds Links between atlas studies and conservation priority analy-ses Bulletin of the African Bird Club 4 93ndash98

Burgess N Fjeldsaring J Rahbek C 1998 Mapping the distributions of Afrotrop-ical vertebrate groups Species 30 16ndash17

Burgess N de Klerk H Fjeldsaring J Crowe T Rahbek C 2000 A preliminary as-sessment of congruence between biodiversity patterns in Afrotropical for-est birds and forest mammals Ostrich 71 286ndash290

Church RL Stoms DM Davis FW 1996 Reserve selection as a maximal cov-ering location problem Biological Conservation 76 105ndash112

Collar NJ Crosby MJ Stattersfield AJ 1994 Birds to watch 2 The World Listof Threatened Birds Cambridge (UK) BirdLife International BirdLifeConservation Series no 4

Colwell RK Lees DC 2000 The mid-domain effect Geometric constraintson the geography of species richness Trends in Ecology and Evolution15 70ndash76

Corsi F deLeeuw J Skidmore A 2000 Modeling Species Distributions withGIS Pages 389ndash434 in Boitani L Fuller TK eds Research Techniques inAnimal Ecology New York Columbia University Press

Csuti B Polasky S Williams PH Pressey RL Camm JD Kershaw M KiesterAR Downs B Hamilton R Huso M Sahr K 1997 A comparison of

reserve selection algorithms using data on terrestrial vertebrates in Ore-gon Biological Conservation 80 83ndash97

Currie DJ 1991 Energy and large-scale patterns of animal- and plant-species richness American Naturalist 137 27ndash49

da Fonseca GAB et al 2000 Following Africarsquos lead in setting priorities Na-ture 405 393ndash394

de Klerk HM 1998 Biogeography and Conservation of Terrestrial Afrotrop-ical Birds PhD dissertation Rondebosch (South Africa) University ofCape Town South Africa

Dobson AP Rodriguez JP Roberts WM Wilcove DS 1997 Geographic dis-tribution of Endangered Species in the United States Science 275550ndash553

Dorst J Dandelot P 1970 A Field Guide to the Larger Mammals of AfricaLondon Collins

Dowsett RJ Dowsett-Lemaire F 1993A Contribution to the Distribution andTaxonomy of Afrotropical and Malagasy birds Liegravege (Belgium) TauracoPress

Dowsett RJ Forbes-Watson AD 1993 Checklist of Birds of the Afrotropi-cal and Malagasy RegionsVol 1 Species Limits and Distribution Liegravege(Belgium) Tauraco Press

Duellman WE 1983 Amphibians in Africa and South America Evolution-ary history and ecological comparisons Pages 200ndash243 in Goldblatt Ped Biological Relationships between Africa and South America NewHaven (CT) Yale University Press

Faith DP Walker PA 1996a Integrating conservation and developmentIncorporating vulnerability into biodiversity-assessment of areas Bio-diversity and Conservation 5 431ndash446

mdashmdashmdash 1996b Integrating conservation and development Effective trade-offs between biodiversity and cost in the selection of protected areas Bio-diversity and Conservation 5 417ndash429

Faith DP Walker PA Ive J Belbin L 1996 Integrating conservation andforestry production Exploring trade-offs between biodiversity and pro-duction in regional land-use assessment Forest Ecology and Manage-ment 85 251ndash260

Fjeldsaring J Lovett JC 1997 Geographical patterns of old and young species inAfrican forest biota The significance of specific montane areas as evo-lutionary centres Biodiversity and Conservation 6 325ndash346

Fjeldsaring J Rahbek C 1997 Species richness and endemism in South Ameri-can birds Implications for the design of networks of nature reserves Pages466ndash482 in Laurence WL Bierregaard RO Jr eds Tropical Forest Rem-nants Chicago Chicago University Press

mdashmdashmdash 1998 Continent-wide conservation priorities and diversificationprocesses Pages 139ndash160 in Mace GM Balmford A Ginsberg J edsConservation in a Changing World Cambridge (UK) Cambridge Uni-versity Press

mdashmdashmdash 1999 Naringr skoven er vigtig biodiversitet i tropeskovene og priori-tering (When the forest is important Biodiversity of tropical forestsand priorities) Den Ny Verden 4 44ndash57

Freitag S van Jaarsveld AS 1998 Sensitivity of selection procedures forpriority conservation areas to survey extent survey intensity and taxo-nomic knowledge Proceedings of the Royal Society of London B 2671475ndash1482

Freitag S Nicholls AO van Jaarsveld AS 1996 Nature reserve selection in theTransvaal South Africa What data should we be using Biodiversityand Conservation 5 685ndash698

Frost DR 1985 Amphibian Species of the World A Taxonomic and Geo-graphical Reference Lawrence (KS) Association of Systematics Collec-tions

Fry CH Keith S 2000 The Birds of Africa Vol 6 London Academic PressFry CH Keith S Urban EK 1988 The Birds of Africa Vol 3 London Aca-

demic PressHacker JE Cowlishaw G Williams PH 1998 Patterns of African primate di-

versity and their evaluation for the selection of conservation areas Bio-logical Conservation 84 251ndash262

Hall BP Moreau RE 1962 A study of the rare birds of Africa Zoology Bul-letin 8 313ndash378

August 2001 Vol 51 No 8 bull BioScience 623

Articles

Hoffman M E Holden P Howard B Hughes R HuttererN H G Jacobsen P Jenkins W Jetz J Kerbis D Koch A JL Lamberis M Languy M Largen H Leirs A Linzey JMutke J Oates B Patterson J-L PerretA Plumtre J C Poyn-ton P A Racey J B Rasmussen G B Rathburn L RobbinsV SalewskiA Schioslashtz D SchlitterA M Simonetta J D Skin-ner S Spawls W Stanley M E Taylor P Taylor V van Cak-enberghe E van der Straeten E van Dijk H van RompaeyW Verheyen V Wallach N Winser R Wirth D Yalden andBirdLife International for providing data to S Galster AJakobsen J B Larsen T Lehmberg P V Nielsen T S Rom-dal L L Soslashrensen and M M Westergaard for entering datato A Andaryati A Channing R Gutberlet D Moyer and HS Thompson for providing photographs and to H M deKlerk D Faith D C D Happold W Jetz J Lovett S LPimm A Plumtre W Sechrest E Underwood M Wisz andthree anonymous reviewers for help with the manuscriptFunding was provided by the Danish Research Council theDanish Council for Development Research the Isaac NewtonTrust of the University of Cambridge and Conservation In-ternationalrsquos Center for Applied Biodiversity Science

References citedAndo A Camm J Polasky S Solow A 1998 Species distributions land val-

ues and efficient conservation Science 279 2126ndash2128Baillie J Groombridge B 1996 1996 IUCN Red List of Threatened Animals

Gland (Switzerland) International Union for the Conservation of Na-ture and Natural Resources

Balmford A Long A 1994 Avian endemism and forest loss Nature 372623ndash624

Boitani L Corsi F De Biase A DrsquoInzillo Carranza I Ravagli M Reggiani GSinibaldi I Trapanese P 1999 AMD African Mammals Databank ADatabank for the Conservation and Management of the African Mam-mals Rome Istituto di Ecologia Applicata

Branch B 1998 Field Guide to the Snakes and other Reptiles of SouthernAfrica 3rd ed Cape Town (South Africa) Struik

Brown JH 1984 On the relationship between abundance and distributionof species American Naturalist 124 255ndash279

Brown LH Urban EK Newman K 1982 The Birds of AfricaVol 1 LondonAcademic Press

Burgess N de Klerk H Fjeldsaring J Crowe T Rahbek C 1997 Mapping Afrotrop-ical birds Links between atlas studies and conservation priority analy-ses Bulletin of the African Bird Club 4 93ndash98

Burgess N Fjeldsaring J Rahbek C 1998 Mapping the distributions of Afrotrop-ical vertebrate groups Species 30 16ndash17

Burgess N de Klerk H Fjeldsaring J Crowe T Rahbek C 2000 A preliminary as-sessment of congruence between biodiversity patterns in Afrotropical for-est birds and forest mammals Ostrich 71 286ndash290

Church RL Stoms DM Davis FW 1996 Reserve selection as a maximal cov-ering location problem Biological Conservation 76 105ndash112

Collar NJ Crosby MJ Stattersfield AJ 1994 Birds to watch 2 The World Listof Threatened Birds Cambridge (UK) BirdLife International BirdLifeConservation Series no 4

Colwell RK Lees DC 2000 The mid-domain effect Geometric constraintson the geography of species richness Trends in Ecology and Evolution15 70ndash76

Corsi F deLeeuw J Skidmore A 2000 Modeling Species Distributions withGIS Pages 389ndash434 in Boitani L Fuller TK eds Research Techniques inAnimal Ecology New York Columbia University Press

Csuti B Polasky S Williams PH Pressey RL Camm JD Kershaw M KiesterAR Downs B Hamilton R Huso M Sahr K 1997 A comparison of

reserve selection algorithms using data on terrestrial vertebrates in Ore-gon Biological Conservation 80 83ndash97

Currie DJ 1991 Energy and large-scale patterns of animal- and plant-species richness American Naturalist 137 27ndash49

da Fonseca GAB et al 2000 Following Africarsquos lead in setting priorities Na-ture 405 393ndash394

de Klerk HM 1998 Biogeography and Conservation of Terrestrial Afrotrop-ical Birds PhD dissertation Rondebosch (South Africa) University ofCape Town South Africa

Dobson AP Rodriguez JP Roberts WM Wilcove DS 1997 Geographic dis-tribution of Endangered Species in the United States Science 275550ndash553

Dorst J Dandelot P 1970 A Field Guide to the Larger Mammals of AfricaLondon Collins

Dowsett RJ Dowsett-Lemaire F 1993A Contribution to the Distribution andTaxonomy of Afrotropical and Malagasy birds Liegravege (Belgium) TauracoPress

Dowsett RJ Forbes-Watson AD 1993 Checklist of Birds of the Afrotropi-cal and Malagasy RegionsVol 1 Species Limits and Distribution Liegravege(Belgium) Tauraco Press

Duellman WE 1983 Amphibians in Africa and South America Evolution-ary history and ecological comparisons Pages 200ndash243 in Goldblatt Ped Biological Relationships between Africa and South America NewHaven (CT) Yale University Press

Faith DP Walker PA 1996a Integrating conservation and developmentIncorporating vulnerability into biodiversity-assessment of areas Bio-diversity and Conservation 5 431ndash446

mdashmdashmdash 1996b Integrating conservation and development Effective trade-offs between biodiversity and cost in the selection of protected areas Bio-diversity and Conservation 5 417ndash429

Faith DP Walker PA Ive J Belbin L 1996 Integrating conservation andforestry production Exploring trade-offs between biodiversity and pro-duction in regional land-use assessment Forest Ecology and Manage-ment 85 251ndash260

Fjeldsaring J Lovett JC 1997 Geographical patterns of old and young species inAfrican forest biota The significance of specific montane areas as evo-lutionary centres Biodiversity and Conservation 6 325ndash346

Fjeldsaring J Rahbek C 1997 Species richness and endemism in South Ameri-can birds Implications for the design of networks of nature reserves Pages466ndash482 in Laurence WL Bierregaard RO Jr eds Tropical Forest Rem-nants Chicago Chicago University Press

mdashmdashmdash 1998 Continent-wide conservation priorities and diversificationprocesses Pages 139ndash160 in Mace GM Balmford A Ginsberg J edsConservation in a Changing World Cambridge (UK) Cambridge Uni-versity Press

mdashmdashmdash 1999 Naringr skoven er vigtig biodiversitet i tropeskovene og priori-tering (When the forest is important Biodiversity of tropical forestsand priorities) Den Ny Verden 4 44ndash57

Freitag S van Jaarsveld AS 1998 Sensitivity of selection procedures forpriority conservation areas to survey extent survey intensity and taxo-nomic knowledge Proceedings of the Royal Society of London B 2671475ndash1482

Freitag S Nicholls AO van Jaarsveld AS 1996 Nature reserve selection in theTransvaal South Africa What data should we be using Biodiversityand Conservation 5 685ndash698

Frost DR 1985 Amphibian Species of the World A Taxonomic and Geo-graphical Reference Lawrence (KS) Association of Systematics Collec-tions

Fry CH Keith S 2000 The Birds of Africa Vol 6 London Academic PressFry CH Keith S Urban EK 1988 The Birds of Africa Vol 3 London Aca-

demic PressHacker JE Cowlishaw G Williams PH 1998 Patterns of African primate di-

versity and their evaluation for the selection of conservation areas Bio-logical Conservation 84 251ndash262

Hall BP Moreau RE 1962 A study of the rare birds of Africa Zoology Bul-letin 8 313ndash378

August 2001 Vol 51 No 8 bull BioScience 623

Articles

mdashmdashmdash 1970 An Atlas of Speciation in African Passerine Birds LondonBritish Museum (Natural History)

Haltenorth T Diller H 1977 Field Guide to the Mammals of Africa includingMadagascar London Collins

Howard PC Viskanic P Davenport TRB Kigenyi FW Baltzer M DickinsonCJ Lwanga JS Matthews RA Balmford A 1998 Complementarity andthe use of indicator groups for reserve selection in Uganda Nature 394472ndash475

Howard PC Davenport TRB Kigenyi FW Viskanic P Baltzer MC Dickin-son CJ Lwanga J Matthews RA Mupada E 2000 Protected area plan-ning in the tropics Ugandarsquos national system of forest nature reservesConservation Biology 14 858ndash875

Keith S Urban EK Fry CH 1992 The Birds of Africa Vol 4 London Aca-demic Press

Kershaw M Williams PH Mace GM 1994 Conservation of Afrotropical an-telopes Consequences and efficiency of using different site selectionmethods and diversity criteria Biodiversity and Conservation 3 354ndash372

Kershaw M Mace GM Williams PH 1995 Threatened status rarity and di-versity as alternative selection measures for protected areas A test usingAfrotropical antelopes Conservation Biology 9 324ndash334

Kingdon J 1989 Island Africa Princeton (NJ) Princeton University Pressmdashmdashmdash 1997 The Kingdon Field Guide to African Mammals London

CollinsLawes MJ Piper SE 1998 There is less to binary maps than meets the eye

The use of species distribution data in the southern African sub-regionSouth African Journal of Science 94 207ndash210

Lawton JH et al 1998 Biodiversity indicators indicator taxa and effects ofhabitat modification in tropical forest Nature 391 72ndash76

Lombard AT 1995 The problems with multi-species conservation Dohotspots ideal reserves and existing reserves coincide South African Jour-nal of Zoology 30 145ndash163

Lovett JC Rudd S Taplin J Frimodt-Moslashller C 2000 Patterns of plant diversityin Africa south of the Sahara and their implications for conservation man-agement Biodiversity and Conservation 9 33ndash42

Margules CR Pressey RL 2000 Systematic conservation planning Nature405 243ndash253

Mayr E Bock WJ 1994 Provisional classifications v standard avian se-quences Heuristics and communication in ornithology Ibis 136 12ndash18

Mittermeier RA Bowles IA Cavalcanti RB Olivieri S da Fonseca GAB1995 A Participatory Approach to Biodiversity Conservation The Re-gional Priority Setting Workshop Washington (DC) Conservation In-ternational

Mittermeier RA Myers N Thompsen JB da Fonseca GAB Olivieri S 1998Biodiversity hotspots and major tropical wilderness areas Approachesto setting conservation priorities Conservation Biology 12 516ndash520

Myers N Mittermeier RA Mittermeier CG da Fonseca GAB Kent J 2000Biodiversity hotspots for conservation priorities Nature 403 853ndash858

Nicholls AO 1998 Integrating population abundance dynamics and dis-tribution into broad-scale priority setting Pages 251ndash271 in Mace GMBalmford A Ginsberg J eds Conservation in a Changing World Cam-bridge (UK) Cambridge University Press

Olson DM Dinerstein E 1998 The Global 200 A representation approachto conserving the earthrsquos most biologically valuable ecoregions Conser-vation Biology 12 502ndash515

Peterson AT Watson DM 1998 Problems with areal definitions of en-demism The effects of spatial scaling Diversity and Distributions 4189ndash194

Pimm SL Brooks TM 2000 The sixth extinction How large where andwhen Pages 46ndash62 in Raven PHWilliams T eds Nature and Human So-ciety The Quest for a Sustainable World Washington (DC) NationalAcademy Press

Pimm SL Lawton JH 1998 Planning for biodiversity Science 279 2068ndash2069Pressey RL Nicholls AO 1989 Efficiency in conservation evaluation Scor-

ing versus iterative approaches Biological Conservation 50 199ndash218Pressey RL Humphries CJ Margules CRVane-Wright RIWilliams PH 1993

Beyond opportunism Key principles for systematic reserve selectionTrends in Ecology and Evolution 8 124ndash128

Rahbek C Graves GR 2000 Detection of macroecological patterns in South

American hummingbirds is affected by spatial scale Proceedings of the

Royal Society of London B 267 2259ndash2265

Raven PH 1988 Our diminishing tropical forests Pages 119ndash122 in Wilson

EO ed Biodiversity Washington (DC) National Academy Press

Reid WV 1998 Biodiversity hotspots Trends in Ecology and Evolution 13

275ndash280

Rosenzweig ML Sandlin EA 1997 Species diversity and latitudes Listening

to arearsquos signal Oikos 80 172ndash176

Ryan P Bento C Cohen C Graham J Parker V Spottiswoode C 1999 The

avifauna and conservation status of the Namuli Massif northern Mozam-

bique Bird Conservation International 9 315ndash331

Scott JM et al 1993 GAP analysis A geographic approach to protection of

biological diversity Wildlife Monographs 123 1ndash41

Sibley CG Monroe BL Jr 1990 Distribution and Taxonomy of Birds of the

World New Haven (CT) Yale University Press

mdashmdashmdash 1993 A Supplement to Distribution and Taxonomy of Birds of the

World New Haven (CT) Yale University Press

Singh A 2000 Application of geospatial information for identifying prior-

ity areas for biodiversity conservation Pages 276ndash280 in Raven PH

Williams T eds Nature and Human Society The Quest for a Sustainable

World Washington (DC) National Academy Press

Skinner JD Smithers RHN 1990 The Mammals of the Southern African Sub-

region 2nd ed Pretoria (South Africa) University of Pretoria

Snow DW 1978 An Atlas of Speciation in African Non-Passerine Birds Lon-

don British Museum (Natural History)

Sprawls S Branch B 1996 A Guide to the Poisonous Snakes of Africa Cape

Town (South Africa) Struik

Stattersfield AJ Crosby MJ Long AJ Wege DC 1998 Endemic Bird Areas of

the World Priorities for Biodiversity Conservation Cambridge (UK)

BirdLife International BirdLife Conservation Series no 7

Underwood E DrsquoAmico J Olson D Dinerstein E Burgess N Itoua I 1999

Terrestrial Ecoregions of Africa (Map)Washington (DC) World Wildlife

Fund

Urban EK Fry CH Keith S eds 1986 The Birds of Africa Vol 2 London

Academic Press

mdashmdashmdash 1997 The Birds of Africa Vol 5 London Academic Press

van Jaarsveld AS Gaston KJ Chown SL Freitag S 1998a Throwing out

biodiversity with the binary data South African Journal of Science 94

210ndash214

van Jaarsveld AS et al 1998b Biodiversity assessment and conservation

strategies Science 279 2106ndash2108

Welch KRG 1982 Herpetology of Africa A Checklist and Bibliography of

the Orders Amphisbaenia Sauria and Serpentes Malabar (FL) Krieger

Publishing

Williams PH 1996 WORLDMAP 41 Priority Areas for Biodiversity Lon-

don British Museum (Natural History)

mdashmdashmdash 1998 Key sites for conservation Area-selection methods for bio-

diversity Pages 211ndash249 in Mace GM Balmford A Ginsberg J eds Con-

servation in a Changing World Cambridge (UK) Cambridge University

Press

Williams PH Gaston KJ 1998 Biodiversity indicators Graphical techniques

smoothing and searching for what makes relationships work Ecography

21 559ndash568

Williams PH Gibbons D Margules C Rebelo A Humphries C Pressey R

1996 A comparison of richness hotspots rarity hotspots and comple-

mentary areas for conserving diversity using British birds Conservation

Biology 10 155ndash174

Williams PH Burgess ND Rahbek C 2000 Flagship species ecological

complementarity and conserving the diversity of mammals and birds in

sub-Saharan Africa Animal Conservation 3 249ndash260

Wilson DE Reeder DM 1993 Mammal Species of the World A Taxonomic

and Geographic Reference Washington (DC) Smithsonian Institution

624 BioScience bull August 2001 Vol 51 No 8

Articles