8/11/2019 Tree Species in Amazonia

1/4

Proc. Nati.Acad. Sci. USAVol.85, pp . 156-159,January 1988Ecology

Tree species richness of upper Amazonian forests(Amazonia/raiforest/diversity

ALWYNH. GENTRYMissouriBotanicalGarden, P.O.Box 299,St. Louis,MO63166-0299Communicated by Peter H. Raven, August 31, 1987

ABSTRACT Upper Amazonian datafor tree speciesrich-ness in 1-hectareplots ar e repor ted. Allplants 21 0 cmdiameter were censused and identifiedin sixplotsin Amazo-nianPeru and one on the Venezuela-Brazilborder.The twoplotsfrom theeverwet forestsnear Iquitos,Peru,arethemostspecies-richin theworld,with 300species21 0 cm diameterin single hectares;al lof th ePeruvian plotsareamongthemostspecies-richever reported.Contraryto acceptedopinion,upper Amazonian forest, and perhaps Central Africanones,have as manyor more treespeciesas comparable Asian forests.Very hightree speciesrichnessseems to be a generalpropertyof mature lowland evergreen forestson fertile to moderatelyinfertile soils on al l threecontinents.

That Southeast Asian rain forestsare the world srichestintree specieshas been widelynoted in both scientificandpopularliterature(1-5).Thisimpressionstems from thefactthathectareparcelsof Southeast Asian foreststypicallyhaveincluded120-200 speciesof trees .10 cm diameter at breastheight(dbh),whereas similarsamplesof Africanand neo-tropicalforestsgenerallyhave 60-120 tree species(figure1.5in ref. 1 . Trees are one of the few k inds o forganismsthatshow this pattern,which contrastsmarkedly witha-diversityfiguresfor birds (6-9), reptiles and amphibians (10, 11),butterflies ref. 12;G. Lamas, personalcommunication),andbats (G .Creightonand L. Emmons, personalcommunica-tions), where neotropicalcommunities have more speciesornonflyingmammals ref. 13; L. Emmons, personalcommu-nication),where community speciesrichnessis similarinsimilar forestsfrom continentto continent.Even for plants,other kindsof comparisons at a much largerscaleshow manymore speciesin theNeotropicsas a whole thanin SoutheastAsia or Africa(14, 15).

Two seriousproblems in the extant database confoundintercontinentalcomparisons of tree speciesrichness.Thefirstis themajor intersitevariationin treespeciesrichnessona givencontinent.Although theAsian tree plotsspan thefullgamut of habitattypes from poor soil,relativelylow diver-sity, heath or kerangasforeststo highdiversity,continuallymoistrainforestson relativelyfertile soils,al ltheequivalentneotropicaland Africandata sets have been until recentlyfrom sites anticipatedto be at the low end of the tropicaldiversitygradient(16). In the Neotropics,this has beenfurtherconfounded by relianceon local tree identifiersor materos who always lump togetherunder the same ver-nacularname many speciesof largetaxonomicallycomplexgenera such as Inga,Eschweilera,Protium, or Iryanthera.

Thus, theneotropicalspeciesrichness valuesfortreeswithwhich Whitmore (1)compared theAsian datawere mostly forintrinsicallyrelativelylow diversityforestsin specialhabitats[seasonallyinundated igap6, white sand caatingaorcampinarana (17, 18)]or from CentralAmerican forests(e.g.,see ref. 19), which ar e comparativelydepauperate in tree

species for biogeographicalreasons c f. ref. 15). Manyrelativelylow-diversityneotropicaldata sets (17, 20-22reliedheavilyon matero identifications.Apparently theonly Amazonian terra firme site for which a completebotanicalinventoryof treeswithspeciesnumbers calculateby herbarium comparison of vouchers has been published isa CentralAmazonian plotnear Manaus (23).This plothad asmany tree speciesin a hectare(ha) as many comparableSoutheast Asian plotsdespiteusing15cm ratherthan 10cmdbh as thelower diameter cutoff.Moreover, theManaus sitehas a pronounced dry season and relatively low annuaprecipitation(==2000mm), factorsexpected to correlatewithrelativelylow species richness(16, 24, 25).

Thispaper reportsvaluesfor tree speciesdiversities instandardizedsample plotsin upper Amazonia, an area withgenerallybetter soils, higherannual precipitation,and aweaker dry season than the few Amazonian sites fo r whichtreeplotdatawere previouslyavailable. Allof thesesitesarein or near areasproposed as Pleistocenerefugia,unlikemanyof the previouslysampled areas (26). In six 1-haplots inAmazonian Peru and one in theupper Rio Negro drainageonthe Brazil-Venezuela border,all plants -10 cm dbh wereidentified.Complete specieslists,standcharacteristics,andintersite floristic comparisons will be reportedelsewhere(27).Each tree in theseplotsis permanently marked with analuminum tag and voucher specimens ar e depositedin theherbariaof the MissouriBotanicalGarden, UniversidadNacional de San Marcos, and Universidad Nacional deAmazonia Peruana. Each plotwa s laid out in closed-canopymature foreston flator slightlyundulatingterrain. The plotswere selectedto be homogeneous except fora smallstreamlepassingnear or through fiveofthe plots. Such rivuletsar e apotentialcomplicatingfactor,buttheyare so omnipresent inAmazonia that it is difficult to locatea square hectareplotthat does not intersectone Therefore,I have assumed that,alongwithtree fallsand poorlydrainedpatches,theycon-stitute an unavoidable partof themicrogeographic within-sitvariationof a hectareof Amazonian forest. Slightintrasitedifferencesin soil fertilityare associatedwith thesemicrotopographicdifferenceson th elocalsoilcatena at some siteshowever, thesedifferencesare trivial when compared to themajor between-sitedifferencesin soil nutrients(25, 27).

If a valueof

8/11/2019 Tree Species in Amazonia

2/4

Proc. Natl.Acad. Sci. USA 85 (1988)

sandy upland terra firme,and th e other is on rich alluvium.Th e site on the Venezuela-Brazil border,near the base ofCerro Neblina, is on highlyleached ultrapoorwhite sand.Th esites and theirsoilsare more completely described elsewhere(25, 27 , 30).

The four full-tropical Peruvian plots all include morespecies .10 cm dbh/ha than previouslyreportedfo r anyneotropicalsite. Th e two plots in th e continuallymoistIquitosarea both includealmost 300 species-10 cm dbh/ha,far exceeding any other known site in th e world. Even theCerro Neblina foresthas more species .10 cm dbh/ha thanmost previouslyreportedneotropicalvalues. Allof th eupperAmazonian plotsare as richas or richerthan th eManaus site(23)in species(Table1 .Thus, th esinglehigh-diversityfigurepublishedfor the Neotropics, previouslyconsidered some-thingof an anomaly in th econtext of the data summarized inFig. 1, is more typical of neotropicalforests than therelativelylow-diversityplotscollatedby Whitmore 1 .

Although comparable data for 1-haplots in th e diverseforestsof CentralAfrica are not available,data forplants .10cm dbh extractedfrom 0.1-hasample plots at Makokou,Gabon, include as many speciesas do similarsizesubplotsofmost of th eneotropicalhectares(Fig.2). Moreover, thereareat least 389 tree speciespotentially.10 cm dbh at Makokou(35), suggestingthat the number of species-10 cm dbh/hamay well be as great as in many Southeast Asian andneotropicalsites. Such high valuesare probably typicalofCentral African forests, since0.1-haplots near Korup Na-tional Park Cameroun, and on th e lower slopesof Mt.Cameroun have, respectively,38 and 43 species-10 cm dbhand 0.64-ha plotsat Korup have up to 138 species-10 cm dbh(Fig. 2 .

I conclude that th eeverwet forestsofupper Amazonia maybe the world s richestin tree species.Th e dipterocarpforestsof Southeast Asia,although remarkably diverse,are not asexceptionalas has been supposed; rather, theirhighdiversityof trees in small sample plots is shared with some SouthAmerican forestsand possiblysome CentralAfrican forestsas well. Indeed, it is hard to imagine a more diverseforestthan at Yanamono where there are only twiceas manyindividuals.10 cm dbh as speciesin a 1-ha patch of forest(Table1), with 63 of th e speciesrepresentedby singleindividualsand only 15 of thespeciesrepresentedby morethan two individuals.In the first 50 trees sampled atYanamono, only two specieswere repeated;therewere 58speciesin the 65 individualsin thefirst 0.1-hasubplotand 90speciesin the 115 individualsin th e first 0.2-hasampled.

300

250

200

._

0

E

z

lOC

SC

0

8

0

80

0

1

r ea

0

ha

FIG.1. Number of species .10 cm diameter on smalhomogenous mixed-species tropical lowland rain forestSoutheast Asia;circles, Neotropics; squares, Africa.Soliare from Table 1; left open square, combination of two noncous 0.1-hatransect samples at Makokou, Gabon; rightopenaverage number of specieson 0.64-haplotsat Korup, Camerother data are from Whitmore (figure1.5 in ref. 1 .

Moreover, many of th e site s common speciesweincluded in th e sample, suggestingthat the localtree

Table 1. Numberof speciesand stems in contiguous1-haplotsin upper AmazoniaNo. of individuals No. of species

Trees .10 Trees .30 Total Trees .10 Trees .30Site Coordinates Total cm dbh Lianas cm dbh species cm dbh Lianas cm dbh Families L

Yanamono 316S, 72054W 606 580 26 110 300 283 17 81 58Mishana 3047S, 73030W 858 842 16 83 289 275 14 54 50Cocha

Cashu 11045S, 71030W 673 650 23 110 204 189 15 68 48Cabeza de

Mono 1020S. 75018W 544 520 24 81 185 169 16 49-50 40Tambopata

alluvial 12049S, 69043W 540 526 14 100 165 155 10 54 41Tambopata

terra firme 12049S, 69043W 602 585 17 80 181 168 13 51-52 42Neblina base

camp 0050N, 6610W 513 493 20 84 10 2 89 13 24 32Data ar e fortrees .10 cm dbh and lianas .10 cm greatestdiameter.Sites are all in Amazonian Peru,except Neblina base c

Venezuela-Brazil border;al l sites arebetween 140 and 400 m altitude.Speciesnumbers in some familiesarebased in parton morphosp

in a few casesmorphospecies limits areunclear ora voucher specimen has been lost. The limitingvaluesfor total number of specisite are 294-303 (Yanamono), 285-291 (Mishana),201-204 (Cocha Cashu),185-190 (Cabeza de Mono), 165-170 (Tambopata alluvial)(Tambopata terra firme),101-103 (Neblinabase camp). In each case, the most likely valuefor number of speciesin the plot is give

Ecology:Gentry

I

8/11/2019 Tree Species in Amazonia

3/4

Proc.Natl.Acad. Sci. USA 85 (1988)

300

25 0

C_

a0L-Co

Ez

200

150

100 teblinaBadas

0.5 1Plot Size (ha)

FIG.2. Speciesarea curves for plants.10 cm dbh in subplotsof some 1-hatree plots.Solidlines, neotropicalsites reportedhere;dashedlines, Asian sites; the Gunung Mulu dataare from ref. 32, and the Andulau and Badas dataare from ref. 33. Stars,CentralAfricansites; the0.1-havaluesare, from to pto bottom,Mt. Cameroun, Korup, Makokou; the 0.2-havalueis Makokou; the lower 0.64-havalueis theaveragefor 13 5plotsat Korup (34),and the higherone indicatesthe highestindividualplotvalue(D. Thomas, personalcommunication).

richnesswould continueto increaserapidlywithincreasing

plot size.Upper Amazonian forests may generallyhave the mostdiversefloral and faunalassemblages in the world.TheCocha Cashu site in Manu NationalPark has the world slargestpublishedinventoryof birds 9 . The Tambopata sitehas even more bird species(36), and the highestknownbutterflydiversityin theworld (12).The world slargestlocalinventoryof mammals is from Balta, Peru (A. Gardner,personalcommunication),themost diversereptilianfauna isfrom theIquitosarea (37,38),and themost diverseamphibianfauna is from Santa Cecilia, Ecuador (31). The new datapresentedhere suggestthat, instead of beinga strikinganomaly, patternsof tree speciesrichnessparallelthose o fbirds, butterflies, reptiles and amphibians,and mammals,withthe world sgreatestlocalconcentrationsof speciesinthe relativelymoistand fertileforestsnear the base of theAndes. Whether these patternsrelate more to the effectsofhypothesized Pleistocenerefugiaor to modern ecology,aswellas thedegreeto which theyma y prove to be replicatedin othergeographicalregions,remain unclear.From a plan-etaryperspective,th eemerging generalitythatdiversityma ybe uniquelyconcentratedin upper Amazonia suggeststhatspecialfocuson preservingremnants of theserapidlydisap-pearingecosystems is of the utmost conservationalimpor-tance.

I thankV. Remsen, P. Raven, P. Ashton,and L. Emmons forreview comments and the NationalGeographicSocietyand U. S.Agency for InternationalDevelopment (DAN-5542-G-SS-1086-00)for support of th e research.G. K. Creighton,W. Duellman,L.Emmons, A. Gardner,G. Lamas, and T. Parker made availableunpublisheddata.

1. Whitmore, T. C. (1984)TropicalRain Forests of theFar East

(Clarendon,Oxford).2. Ashton, P. S. (1977)Ann. Mo . Bot.Gard. 64, 694-705.3. Caufield,C. (1985)In the Rainforest(Knopf,New York).4. Gan, Y. Y., Robertson,F. W., Ashton,P. S., Soepadmo, E.

Lee,D. W. (1977)Nature (London) 269,323-325.5. Yeom, F. B. (1984)Unasylva 36,2-16.6. Amadon, D. (1973)in TropicalForest Ecosystems in Africa

and South America: A Comparative Review, eds. Meggers, B.,Ayensu, E. Duckworth, W. (SmithsonianInst., Washington,DC), pp. 267-277.

7. Pearson,D. L. (1977)Condor 79,232-244.8. Karr,J. (1976)Ecol.Monogr. 46,457-481.9. Terborgh, J. W., Fitzpatrick,J. W. Emmons, L. (1984

Fieldiana Zool.21, 1-29.10. Inger,R. F. (1980)Am. Nat. 115,761-770.11. Scott,N. J. , Jr. (1976)Biotropica8, 41-58.12. Lamas, G. (1985)Rev. Peru Entomol. 27,59-73.

13. Emmons, L. H., Gautier-Hion,A. Dubost G. (1983)J.Zool.199,209-222.14. Raven, P. H. (1976)in Conservation of Threatened Plants,

eds. Simmons, J. , Beyer,R., Brandham, P., Lucas, G. Parry,V. (Plenum,New York),pp. 155-179.

15. Gentry,A. H. (1982)Ann. Mo . Bot. Gard. 69, 557-593.16. Gentry,A. H. (1982)Evol.Biol.15, 1-84.17. Black,G. A., Dobzhansky, T. Pavan, C. (1950)Bot. Gaz.

111,413-425.18. Anderson, A. B. Benson, W. W. (1980)Biotropica12,

235-237.19. Hartshorn,G. S. (1983)in Costa Rican Natural History,ed .

Janzen,D. (Univ.Chicago Press,Chicago),pp. 118-157.20. Davis,T. A. Richards,P. W. (1934)J. Ecol.22, 106-15521. Schulz,J. P. (1960)Verh.K. Ned. Akad. Wet. Afd.Natuurkd.

53, 1-267.22. Uhl,C. Murphy, P. G. (1981)Trop.Ecol.22, 219-237.23. Prance,G. T., Rodrigues,W. A. da Silva,M. F. (1976)Acta

Amazonica 6, 9-35.

15 8 Ecology:Gentry

8/11/2019 Tree Species in Amazonia

4/4

Proc. Natl.Acad. Sci. US A 85 (1988)

24. Gentry, A. H. (1987)Ann. 1st Simp6sio do Tr6pico (Omedo,Feb.19-24,1987 (Dept.Difusfiode Technologia,Brasilia,D. F.).

25. Stark,N. Gentry, A. H. (1987)Bioscience, in press.26. Prance, G. T. (1982)BiologicalDiversificationin the Tropics

(Columbia, New York).27. Gentry, A. H. (1988)Ann. Mo . Bot. Gard., in press.28. Stark, N. (1971)Int. J. Trop.Ecol.12(1),24-50.29. Oficina Nacional de Evaluaci6n de Recursos Naturales (1976)

Mapa Ecologico de l Peru (ONERN, Lima), pp. 1-145 and 4map pages.

30. Gentry, A. H. Emmons, L. H. (1987)Biotropica19, 216-227.31. Duellman, W. E. (1978)Univ.Kans. Mus. Nat. Hist. Misc.

Publ. 65, 1-352.

32. Proctor,J. , Anderson, J. M., Chai,P. Vallack,H. W.J. Ecol.71, 237-260.

33. Ashton, P. (1964)Oxf.For. Mem. 25, 1-75.34. Gartlan,J. S., Newbery, D. M., Thomas, D. W. Wate

P. G. (1986)Vegetatio 65 , 131-148.35. Gentry,A. H. Dodson, C. H. (1987)Ann. Mo . Bot.

74, 205-233.36. Donahue,

P., Parker,T., Sorrie,B.

Scott,D.

(1988)B

the Tambopata Nature Reserve Madre de Dios,Peru, in37. Dixon, J. R. Soini,P. (1975)Contrib.Biol. Geol.

Public Mus. 4, 1-58.38. Dixon, J. R. Soini, P. (1976) Contrib .Biol. Geol.

Public Mus. 12, 1-91.

Ecology: Gentry