01 RAP Working Papers_madidi

8/13/2019 01 RAP Working Papers_madidi

1/113

Rapid ssessment rogram

Biological ssedsment ofthe lto Madidi Regionand ad jacent areas ofNorthwest BoliviaMay 18 June 15 1990

CONSERVATION INTERNATIONAL

R P

Working

apers

DECEMBER 99


2/113

Rapid ssessment Program

Knowledg e of the pl;unet s biota is uneven. While som e regib e e n s t ~ ~ d i e dxtensively by biologists, others have scarcely beenby scientific exploration. In recent years, large-scale changes inworldwide have made it impel.ative that we analyze as Inany of k ~ i o w ~ i11? t lts ns possible while they co ntin ue to exist in relativtine states. Rapid assessm ents allow conserva tionists to evalcompare areas as a means of determining the biological base for ction priorities.

Conservation International s Rapid A ssessment Progr a~n Rsembles teams of world-renowned experts and host country sciegenerate first-cut assessments of the biological value of different eas, ancl regions around the worlcl. RAP aims to identify regions obiological im portance for conservation. An area s importance canacterized by its total biocliversity, its elegree of encle~nism, he ~of an ec osy ste~ n, r the elegree of risk of extinction on either a naa global scale. As a conservation tool, R AP precedes long-term inventory.

When satellite im ages of an area targeted for a RA P assessavailable, the team consults them prior to a trip to determine the forest cover ancl likely areas for exploration. O nce in-c oun t~y ,tists make o verflights in s~ na ll lanes or helicopters to identify foancl points for field transects. Gr oi~ nd ravel is more difficult:combination of vehicles, boats, pack animals, ancl foot travel is reget the team to re ~n ot e ites where few, if any, roads exist. Tripsfrom f o ~ ~ ro eight weeks.

On each trip, in-country scientists form a central part of tLocal experts arc especially critical to ~~nclcrstandingreas whereploration has been ~unclertaken. Furth erm ore, any subse quen t reseprotection of habitats following a RAP trip will invariably depeninitiatives of local scientists and conservationists.

We hope that Rcrl. Wol Itilig PCIIICI S ill contribute to knowledge of the earth s biodiversity. But even more i~nportanth ese s t ~ ~ d i e sill provide conservation groups around the world they need to help preserve the earth s biological heritage for fuerations.

Ted Park er ancl Brent BaileySeries Editors


3/113

Rapid ssessment Program R PWorking

Biological ssessment ofthe lto Madidi Regionand adjacen t areas ofNorthwest BoliviaM ay 18 June 15 1990

CONSERV TION INTERN TION L

DECEMBER 1991

apers


4/113

Rapid Assessnzerzt Working Pa11er s are occasionalreports published three to five times a year. Forsubscription information w rite to:Conservation InternationalPublications1015 18th Street NWSuite 1000Washington DC 20036Tel: 2021429-5660Fax: 2021887-5188Co nse~ vat io n nternat ional is a private , nonprofit organizat ionexempt from federal income tax under sect ion 501 c) 3) of the

Internal Revenue Cod e.

C o n ~ e r \ ~ a t i o nnternat ional , 199

Library of Congress Catalog Card

N umber 9 1-078 133

Printed on recycled paper


5/113

Table of Contents

AcknowledgmentsParticipants 3Organizational Profiles 4Overview 6Conclusions 6Opportunities 9Summary Of Field WorkFurther Research 12Technical Report 4Alto Madidi Region 14Pampas Region 25Apolo Region Mid-Eleva tion Wet Forest 26Apolo Region Mid-Elevation Dry Forest 30Appendices 341. Birds of the Alto Madidi 362. Birds of the Lower Rio Heath Bolivia/Peru 483. Birds of Ixiamas 614. Birds of Calabatea5 Mammal List 726. Observations on the Herpetofauna 747. Plant List: Alto Madidi Bajo Tuichi and the

Foothill Ridges8 Plant List: Apolo-M id-Elevation Wet Forest 939. P lant List: Apolo-M id-Elevation Dry Forest 10010. Plant List: Pampa-Ixiamas 104Bibliography 1 7


6/113

cknowledgments

Many people contributed to the success of this trip and production of thereport. Our thanks go to Eduardo Forno for coordinating extensive pre-trip arrangements and anticipating numerous details. Ana Martinet deMollinedo Marina de Montafio and Lourdes Larea of the CI Bolivia of-fice cheerfully accommodated the team and provided logistical support.Marcelo Sommerstein generously offered use of the Madidi camp andsupport from cam p personnel. Padre Nivardo graciously provided assis-tance with field transportation. Maria Marconi and Isabel Mercado of theBolivian Conservation Data Center provided valuable review of the ap-pendices. Dr.Sydney Anderson generously shared records from his fileson Bolivian mam mals. Stephen Nash prepared the map of the regionLeonor Greenidge helped prepare the manuscript and John Carr and TomSchulenberg provided valuable editorial assistance. Finally we gratefullyacknowledge the MacArthur Foundation for establishing the Rapid As-sessment Program and the Beneficia Foundation and CI members for ad-ditional financial support.

Editors NoteThis report is both a synthesis and compilation of the findings of the RA Pteam. The Overview section provides highlights of the trip and integratessuggestions and conclusions that were shared by team m embers after ana-lyzing their results. The Technical Report provides more detailed infor-mation on species and natural communities and is derived from separatereports submitted by the individual team mem bers. Authorship of sepa-rate sections is attributed to the contributors whenever possible thoughsome sections were combined in order to avoid overlap or repetition.

C O N S E R V A T I O N I N T E R N A T I O N A L Rapid ssessment Program


7/113

articipants

SCIENTIFIC PERSONNEL Flavio Hinojosa

Theodore A Parker IllMammalogistInstituto de Ecolo gia

OrnithologistConservation International

FIELD ASSISTANCE

Robin B FosterPlant EcologistConservation International

Hermes JustinianoPilot; Director Fundaci6n Am igos de laNaturaleza Santa CruzBolivia

Louise H. EmmonsMammalogistConservation International

Alwyn H. Gentry

Abel CastilloFundaci6n Amigos de la Naturaleza

Brent BaileyDirector of Biological ProgramsConservation InternationalBotanist

Conservation International Edward WolfStephan BeckBotanistHerbario Nacional de Bolivia

Silvia EstenssoroBotanistCentro de Datos para la Conservacibn deBolivia

EditorlwriterConservation International

EDITORS

Theodore A Parker Ill

Brent Bailey

RAP Working Papers One December 99


8/113

Organizational Profiles

Conservation International CI)Conservation International (CI) is an interna-tional, nonprofit organization based in Wash-ington, D.C., whose mission is to conservebiological diversity and the ecological proc-esses that support life on earth. CI employ s astrategy of ecosystem conservation whichseeks to integrate biological conservation witheconomic development for local populations.CI 's activities focus on developing scientif icunderstanding, practicing ecosystem manage-ment, stimulating conservation-based develop-ment, and assisting with policy design.Conservation International CI)1015 18th St. NWWashington, D.C. 20036 U.S.A.202-429-5660CI-BoliviaAvenida Villaz6n #1958, Of. 10-ACasilla 5633La Paz, Bolivia(5912) 341230

Centro de Datos para la Conservacibnde Bolivia CDC)The Centro de Datos para la Conservaci6n deBolivia (CDC) is a private, nonprofit Bolivianorganization created to contribute to the con-

servation of living resources of Bolivia. View-ing conservation as the appropriate use of theseresources, with the goal of improving the qual-ity of life of present and future generations, theCDC promotes sustainable development thatintegrates ecological, social, and econo mic fac-tors with basic principles of conserva tion. TheCD C's central mission is to provide the techni-cal base for the development of strategies, poli-cies, programs, and projects of protection andrational use of the country's biological heri-tage, and focuses on bridging the gap betweenthose who generate biological information andthose wh o use it.Centro de Datos para la Co nservacionde Bolivia CDC)Calle 26, Cota-Cota Casilla 11250La Paz, Bolivia(5912) 797399

Herbario Nacional de BoliviaHerbario Nacional de Bolivia is a botanical re-search center, established in 1983 by agree-ment between the Universidad Mayor de SanAndrCs and the Academia Nacional de Cien-cias de Bolivia under the sponsorship of theNeotropical Flora Organization in La Paz.

The Herbarium's principal goal is thestudy of the flora of Bolivia, through floristic

CONSERVATION INTERNATIONAL Rapid Assessment Program


9/113

inventories, establishment of botanical collec-tions, and development of basic and applied re-search projects, thereby contributing toscientific training at national and internationallevels in the Neotropical region.Herbario Nacional de BoliviaCasilla 10077La P az, Bolivia(59 12) 792582 an d 7924 16

lnstituto de Ecologia de BoliviaThe Instituto de Ecologia de Bolivia is a scien-tific training and research center of the Un iver-sidad Mayor de San AndrCs de La Paz. Itsprincipal objective is the enhancement of scien-tific capacity to resolve ecological problems inBolivia. To realize this goal, the Institutotrains professional biologists w ith an ecologicalorientation and undertakes basic and appliedresearch in programs of conservation, agro-ecology, and inventories of Bolivian flora andfauna.Instituto de Ecologia d e BoliviaCampus UniversitarioCota-Cota, Calle 27Casilla Correo Central 100 77La P az, Bolivia(5912) 792582 or 792416

Fundacion migos de la NaturalezaThe Fundaci6n Amigos de la NaturalezaFAN) is a private, nonprofit organization es-

tablished in 1988. FAN S mission is to protectBoliv ia s biological diversity. In collaborationwith other nongovernmental organizations, theBolivian government, and the internationalconservation community, FAN provides vitaltechnical and financial assistance to NoelKempff and Amboro National Parks and theRios Blanco y Negro Wildlife Reserve, pro-tected areas comprising a total of almost 7.5million acres. A long-term goal of FAN is toexpand the system of protected areas from thecurrent 2 percent to at least 10 percent of Bo-livia s national territory through the estab-lishment of new areas, training of conservationprofessionals, and a public environmental edu-cation program.Funda cion Amigos de la Naturaleza FAN)Av. Irala 42 1P.O. Box 2241Santa Cruz, Bolivia(591-33) 33806(591-33) 41327 (fax)

R P Working Papers One December 1991


10/113

verview

From 18 May to 15 June 1990, Conservation International s Rapid As-sessment Program team (Louise H. Em ~n ons , obin B. Foster, Alwyn H.Gentry, and Theo dore A . Parker, 111) and counterparts from Bolivian insti-tutions (Stephan Beck, Silvia Estenssoro, and Flavio Hinojosa) undertookrapid evaluations of fauna and flora of lowland and mo ntane forests in thedepartment of La Paz, on the eastern slope of the Andes in northwester11Bolivia.

The purpose of the expedition was to assess quickly the biologicalimportance of a vast, largely unexplored wilderness area in Provincia Itur-ralde, along the upper reaches of the Rios Heath, Madidi, and Tuichi. Theregion encompasses nearly 50,000 sq km of pristine forest and grassland,none of which currently receives protection und er Bolivian law.

ON LUSIONSResults for this area of Bolivia indicate a high diversity of flora and faunathat rivals the richest know n sites on the globe. Habitat heterogeneity, thegeneral species richness of Amazonian and Andean forests and their prox-imity to each other, relatively h igh precipitation, and nearly co mplete ab-sence of long-term human perturbation are among the related probablecause s for the high levels of species diversity enco untered by th e group.The region of northern La Paz, from the high Andes to the mouth ofthe Rio Heath in the lowlands, is likely to harbor more bird and mammalspecies than any other comparable area of Bolivia. It is possible that morethan 1,000 species of birds, or an amazing 11 percent of all bird species onearth, will eventually be recorded along a transect from the Andean grass-lands near Lake Titicaca to the lowland forests and savannas near themouth of the Rio Heath. The region hosts what are probably the most

CONSERV TION INTERN TION L Rapid ssessment Program


11/113


12/113

Results for thisarea of Boliviaindicate a highdiversity offlora andfauna thatrivals therichest knownsites on theglobe

species-rich forests in all of Bolivia on the an-cient river terraces and the adjacent slopes ofhills and ridges. Com bined with the adjacentTambopata-Candamo reserve in Peru, this areacould become a biodiversity reserve unsur-passed in all of South America.

We saw in our overflights that most of avast region of northern La Paz is uninhabitedor virtually so. It is therefore in an ideal statefor long-term land-use planning for conserva-tion and sustainable development.

We had already suspected that the forestsnear Bolivia's frontier with Peru at the base ofthe Andes (known as the 'lower yungas ')would be the riche st in plant species of any for-est in Bolivia. Our rapid assessment confirmsthis to our satisfaction. Th e region has goodrepresentation of the southwest Am azonian bi-ota. On e comp onent, the floodplain forests ,may be richer away from the Andes where therivers m eander mo re, and we suspect (but stilldo not know) that the middle or upper yungas(mountain slopes) are also richest in speciesnear the Peruvian border.

The Alto Madidi and portions of thelower Madidi have virtually intact faunal andfloral assembla ges. Only in the immediate vi-cinity of the airstrip and ca tnp have so me of theCedrel inga mars nzacho' in Spanish) treesbeen cut out. It is astonishing to see large Ce-drzla (cedro) near the lumber camp. The ab-sence of Swietenia (mahogany, known as nzaraor caob a), either as a result of muc h earlier log-ging or for lack of suitable habitat, is very for-tunate. It mean s there will be little loggingpressure on this area for several more years.The absence of hunting has left even the largegame animals near the rivers.

The Andes provide numerous habitattypes, including four elevational zones abovethe lowlands, all of which host different birdand small mamm al faunas. Our fieldwork atAlto Madidi (14 days) reveals how little weknow of bird distribution in northern Bolivia

CONSERV TION INTERN TION L

(and birds are the best-known vertebrategroup ).

The lowland savannas (pampas) are nota priori expected to be richer in plant speciesnear the northwestern frontier with Peru, butthey may contain more endemic species andcertainly are less disturbed here than anyw hereelse north of the Rio MamorC. Althou gh itsmammal list may not be large, the Pampas delHeath may be one of the only undisturbednatural habitats of its type, and serves as a ref-uge for species that are persecuted elsewhereand in need of protection. The pampas clearlyrepresent an important conservation priority inBolivia and South America.

At Ixiamas we found several grasslandbird species that are declining throughout m ostof their ranges in central South America (e.g.,Cock-tailed Tyrant, Alect~.uru s tricolor, andBlack-masked Finch, Coryphaspiza melano-tis). Tha t large populatio ns of such speciessurvive in northern and central Bolivia under-scores the conservation opportunities that stillexist in this region.

In addition to the northwestern borderevaluated here, one could mak e a similar argu-ment about the conservation importance of Bo-livia 's other border regions. Th e southernborder with Argentina and Paraguay has prob-ably the greatest richness within Bolivia ofsubtropical biota. The eastern border with Bra-zil (e.g. , the Serran ia de Huanchaca) certainlyshould have the greatest richness within Bo-livia of Brazilian Shield species. Th e cerradovegetation is disappearing in Brazil to a muchgreater degree than in the smaller area in Bo-livia.

Bolivia is not at the heart of any o ne ofthese large biotas, but it can claim to be themost important transitional country betweenthe major biotic regions of southern SouthAmerica. Our goal is not to single out Boliviaas the ideal parkland for southern So uth Ameri-can biota, but rather to draw attention to itsenorm ous biological wealth and relatively low



13/113

population pressure. The pressures of growing value of these forests, in addition to their bio-population and development on the neighbor- logical importance, is inestimable. The uppering countries may in many cases be too strongto protect these biotic system s. In Bolivia it isstill very possible.

CONSERVATION OPPORTUNITIESThe biological surveys discussed in this reportindicate that lowland and montane forests innorthern La Paz support the richest plant andanimal communities in the country. W e there-fore suggest the following opportunities fortheir protection:

Establishment of a large conservationunit in the department of La Paz that wouldinclude large areas of lowland and foothillforest along the upper Rios Heath andMadidi including the entire upper drainageof the Alto Mad idi), and the higher ridges tothe south e.g., the Serrania de Tutumo).The very high biodiversity of the Alto Madidiregion, particularly that of lowland forests atthe base of the mountains, but also of montaneforests on outlying and higher ridges to thewest, could be protected in a reserve that wouldborder the Tambopata-Candamo Reserve lyingalong the Peruvian side of the Rio Heath. Theresulting bi-national reserve would encompasssome of the richest forests in upper Amazonia,including the most diverse forests in all of Bo-livia. Up to 12 percent of all bird species onearth, for example, and the highest plant spe-cies diversity yet reported from Bolivia, havebeen found between 400 and 4,000 meters inthe headwater region between the Rio Tam-bopata, Peru, and the Rio Alto Madidi, Bolivia.

montane section of the proposed border reservecould be extended southeast to the Apolo-LaPaz highway currently under construction. Thebenefits of limiting colonization along suchroads e.g., to limit erosion that results from de-forestation for subsistence agriculture) areoften overlooked or ignored.

Establishment of a second conservationarea for watershed management, the protec-tion of diverse ecosystems, and developmentof ecotourism in the drainage of the RioTuichi. This could include the higher eleva-tions of mid-montane forest south of the Tuichiup to the burned plateaus northeast of Apolo.This would encom pass several long trails fromApolo to the lower Tuichi which could bemaintained for trekking, patrolling, and re-search, now that new access roads and air-planes make the trails of little use to thepopulation around Apolo.

The tall, lower montane forests of theSerrania de Esla bbn, along the western side ofthis valley, probably support the richest mon-tane plant and animal communities in Bolivia.The existing tourist lodge at Laguna SantaRosa would be a n excellent base for biologicalinventories of the surrounding lowland forestsas well as of montane forests to the west. Likethe Alto Madidi site described above, the lodgeis strategically located for monitoring eco-nomic development along the river.

Creation of extractive reserves aroundthe proposed M adidi conservation unit couldlead to the rational, long-term exploitation offorest-based products. Economically impor-tant forest resources, including large popula-

Extension of the boundaries of this Alto tions of valuable trees such as SwieterliaMadidi protected area southward along the macrophylla and Cedr.ela odorata occur inBolivia-Peru border to include the full gradi- this region. Unfortuna tely, these are being ex-ent of montane forests up to treeline includ- ploited rapidly, and perhaps not rationally, by aing the humid paramos). The watershed number of logging companies that have already

RAPWorking Papers One December 99

It is possiblet h a t . . 11percent of allbird species onearth willeventually berecorded alonga transect fromthe Andeangrasslandsnear LakeTiticaca to thelowland forestsand savannasnear the mouthof the RioHeath.


14/113

been granted concession s. Establishm ent ofextractive reserves where harvesting of suchspecies could be monitored and studied wouldbenefit the Bolivian economy far into the fu-ture. Forests along the lower Rio Heath, sup-port large numbers of castafia trees Ber thol ettiaexcelsa), which already yield economic re-wards for local people.

Construction of a dual-purpose bio-logical station and guard post at the s iteof the recently abandoned AserraderoMoira sawmill at Alto Madidi . T h e s t ra -t eg ic impor tance o f th i s ba s e fo r s c i en t i f i cs tud ie s a s we l l a s fo r s u rve i l l ance canno tbe ove rs t a t ed . Acces s by a i rp lane i s awonde r fu l conven ience fo r the ope ra t ion o fa s t a t ion , requ i r ing on ly ma in tenance o f thea i r s t rip . A l s o e ss en t i a l is ma in tenan ce o fthe road to Ix iamas . An improved ne tworkof t ra i l s a round the s t a t ion wou ld fac i l i t a t eb io log ica l exp lo ra t ions o f the s u r round ingfo re s t s . F lo ra l and fauna l inven to r i e s o fthe unexp lo red r idges , s ome a s h igh a s1 800 m, to the wes t o f A l to Madid i wou ldrevea l the p re s ence o f l a rge numbers o f en -demic s pec ie s . As a gua rd pos t , t he s ta t ioncou ld a l s o moni to r exp lo i t a t ive p rac t i ce ss uch a s t imbe r ex t rac t ion and go ld min ing .At p re s en t the re s eems to be l i t t l e con t ro love r s uch ac t iv i t i e s.

Establishment of a protected area en-compassing the Bolivian Pampas del He-ath and surrounding forests the onlyextensive areas of undisturbed pantanal-l ike grassland rem aining in northern Bo-livia. T his cou ld be accompl i s hed wi thou tmuch d i f f i cu l ty , a s the human popu la t iona long the lower R io Hea th i s ve ry s ma l land few i f any ca t t l e g raze the re . A b io -log ica l re s e rve wou ld p ro tec t hea l thy po pu-la t ions o f many l a rge m a m m a l s p e c i e s , a swe l l a s a h igh d ive rs ity o f b i rds s ee Ap-p e n d i x 2). P a m p a s o n t h e P e r u v i a n s i d e o f

the r iv e ra re a l read y p ro tec ted by l aw a s pa r t oft heTambopa ta -CandamoRese rve b u t a r e v e r ys ma l l compared to thos e on the Bo l iv ian s ide .Bo liv ian s and Peruvians living along the riverbelow the grasslands could find opportunitiesfor employment in managing and protecting thereserve.

As one o f the l a s t r ema in ing Bo l iv -i a n p a m p a s n o t o v e r r u n b y c a t t l e a n d s e e m -ing ly un ique in f lo r i s t i c compos i t ion , thePampas de l Hea th s hou ld be kep t f ree o froads . T he r ive r and s ma l l a i r s tr ips p ro -v ide adequa te acces s and con t ro l .

7 Protection of representative examplesof the unique plant communities that oc-cur in the semiarid valleys of the north-ern yungas such as the Rio Machariapodry forest. Numerous po ten t i a l ly th rea t -ened p lan t and an ima l s pec ie s may be re -s t r i c ted to s ma l l geograph ic and e leva t iona lzones wi th in the yungas reg ion . T he r e i san u rgen t need fo r rap id and in tens ive b io -log ica l inven to r i e s o f s uch a rea s .

Promotion of reforestation projects inthe densely settled inter-Andean valleyssuch as that surrounding Apolo. W ewere surprised to learn tha t soldiers a t theloca l mi l i t a ry ba s e a re s en t up to 5 kmaway to ga the r f i rewood and bu i ld ing ma te -r i a l s f rom ex i s t ing na t ive fo re s t s. E uca lyp-tus p lan ta t ions , wh ich commonly s upp lyf i rewood to the loca l popu la t ions in o the rd ry Andean va l l eys , have no t been cu l t i -va ted he re . W e would encourage the p lan t -ing o f o the r t ree s pec ie s , p re fe rab lyfa s t -g rowing na t ive s pec ie s .

Extension of a reserve or protectedarea along the Peruvian border to thenorthwest t ip of Pando. T his wou ld addm a n y p r i m a t e a n d s m a l l m a m m a l s p e c i e sno t found fu r the r s ou th , and would re s u l tin a re s e rve r iva l l ing Pa rque Nac iona l

CONSERVATION INTERNATIONAL Rapid Assessment Program


15/113

Man u of Peru as the biologica l ly r ichest pro-t ec ted a re a in t he wor ld .

SUMMARY O F FIELD WORKOver a two-week period, the group sur-

veyed lowland evergreen forests at AltoMadidi, a logging camp on the west bank of theRio Madidi about 20 km south of the Peruvianborder and 100 km n orthwest of the nearestsmall town, Ixiamas. Another ten days werespent in the area of Apolo, 125 km southwestof Ixiamas, in mid-elevation wet and dry for-ests. Brief field time was spent on the lowridges and savannas around Ixiamas. Over-flights crossed the broad Pampas del Heathnorthwest of Ixiamas along the Bolivia-Peruborder, and southeast of Ixiamas to Rurren-abaque on the Rio Beni.

Field methods varied according to eachspecialist. Gentry and Estenssoro collectedplants and data on woody species along a seriesof transects 2 meters and 5 0 m long in differenttypes of forests. Foster made qualitative as-sessments of the vegetation structure and plantcommunity composition of all habitats withinwalking distance, made lists of all the plantspecies observed, and contributed voucherspecimens of important plant species not foundin the transect. During overflights, he identifiedforest types in the region from the air. Beckdid general collecting of plant specimens, andobtained specimens of grass species for hisown research. Emmon s made daily and nightlyforest walks (totalling 85 hours at AltoMadidi), recording all mammal species seenand heard, in addition to those caught in smalllive traps. Hinojosa mist-netted bats and alsotrapped small mamm als. Parker and Castillosurveyed birds with the use of tape-recordersand mist nets.

At the Alto Madidi camp, where weather,logistics, and absence of human perturbationwere optimal, Parker identified 403 bird spe-

cies in 14 days along a transect of roughly sixkilometers by 200 meters. Of these, nine werefirst records for the country, 30 were recordedin the country for only the second time, and 52species were new for the department of La Paz.Among the 45 species of mammals Emmonsand Hinojosa identified in that period wereabundant populations of tapirs and spider mon-keys, testimony to the absence of local humanimpact. Of particular interest were two speciesof mammals previously unrecorded for Bolivia:a little big-eared bat (Mieonycter.is ~~icefor.i),and a spin y tree rat (Mesom ys Izispidus). In arare sighting of the Short-eared Dog (Atelo-cyrzus nzic~.otis), he individual carried a frog inits mouth; this is the first record of food habitsfor a wild individual.

Botanical results from the Madidi campare equally exciting, showing an unusuallyhigh diversity of plant species. Forests onfloodplains, high terraces and slopes, andridgetops, each host distinctive floras and con-tribute to the overall richness of the area. In atenth-hectare sample on forested low rollinghills, Gentry found 204 species greater than 2.5centimeters in diameter. According to hisanalysis, the average moist forest transectyields 15 2 species. Many of the species foundat Madidi are new to Bolivia; so me of them arelikely to be new to science.

While we were fortunate to have twoproductive weeks in the lowland forest,weather and logistical problems limited exten-sive coverage of the region s savannas. Ourbrief exposure to Pampas del Heath and thearea around Ixiamas, as well as our past experi-ence on the adjacent Peruvian pampas, how-ever, clearly point to the grasslands of theregion as a high priority for further explorationand conse rvatio n work . Botanically, initial in-dications are that the region s various savannasconstitute a far more complicated mosaic ofisolated habitat islands than a glance at the ma pwould suggest.

R P Working Papers One ecember 99


16/113

Future rapidassessmentsare urgentlyneeded inBoliviaconcentratingon specifichabitatsphysiographictypes andvegetationtypes. thepampas andry valleys

must be anurgent prioritybecause so fewremainuntouched byhuman activity.

Remarkable as these findings are, theyrepresent only t he first step in a recommendedresearch and conservation endeavor for the re-gion. Muc h remains to be discovered: A rapidassessment is a preliminary indicator of anarea s biological importance a nd appropriatelygenerates more questions than answers. Fromthis expedition, however, it is clear that longer-term, systematic biological inventories willconfirm the value of northern La Paz as a re-pository of impressive biological diversity thatis of global significance. Simultaneo usly, in-formation from this initial assessment is suffi-cient to call attention to the conservation valueof the region and to stimulate initiatives for itslong-teim management and protection.

DIRECTIONS FOR FURTHER RESE RCH

Rapid ssessmentsThe R AP team s overview of the mosaic ofvegetation types of northern La Paz serves itspurpose by focusing international attention onthe importance of this region. Future rapid as-sessments are urgently needed in Bolivia con-centrating on specific habitats, physiographictypes, and vegetation types.

The pampas offer a good starting point.It became obvious to us from our brief surveysof the westernmost patnpas of northern La Pazthat they differ radically from each other aswell as differing from pampas in the Beni andSanta Cruz region. To evaluate the conserva-tion im portance of the different Bolivian sa van-nas, an expert team of botanists, zoologists,and ecologists should be organized soon to in-ventory and compare them.

The team should go from one to anotherin succession, surveying all the different habi-tats within each pampas area. A smaller teamchec k should be done in high-water periods.

Floodplains should be evaluated for vari-ation along the length of rivers, and compari-sons made between floodplains from north to

CONSERVATION INTER NATIONAL

south and east to west. Alluvial fans should becompared along the whole length of the base ofthe Andean foothills. From this, scientistscould determine if the Beni river has reallybeen a major isolating mechanism, blockingthe migration of species from northern Boliviato central Bolivia as some evidence suggests.Similarly, the low foothill ridges should becompared along the base of the Andes. Thiswould be especially instructive if differentkinds of ridges could be categorized as theyhave been in this report.

Up to now, collecting in montane forestshas extended to little more than along the fewgood roads down from the altiplano. muchbroader comparison is needed to distinguishdifferent communities if they exist, to deter-mine the variation in altitudinal differences ofthese communities, and to discover importantdifferences between areas in diversity and en-dem ism. It is urgent that the inter-Andean dryvalleys be studied to find how many, if any, arestill reasonably intact and how much variationexists among them.

All of these suggested surveys areneeded soon. But the pampas and dry valleysmust be an urgent priority because so few re-main untouched by human activity. Many otherimportant habitats or vegetation types in Bo-livia-ranging from puna to chac o woodland-require an initial rapid assessm ent.

Remote ImageryMany regions of special interest to conserva-tion are subject to frequent cloud cove r or havenot been given priority by government officialsfor remote-sensing information. Frequently,modern radar and multispectral images are notimmediately available to the RA P team. Nev-ertheless, once the image information is avail-able, and especially once it can be processed bycomputer to reveal fine-scale patterns inground cover, the information can quickly beput to use and extrapolated to larger areas formapping. It remains to be seen how many of



17/113

the wet-forest plant communities recognizableon the ground can be distinguished with theseimages. It is already clear from this trip thatsatellite images are not sufficient at least byeye) to distinguish several of the importantplant communities.

Geology and ClimateMost of the plant communities are stronglysubject to the effects of geological substrateand climatic variation. Scientists could prob-ably map comm unity distribution without everlooking at an organism if enough informationwere available on the geographic distributionof geology and climate. Unfortunately, thegeological maps that exist are crude and inap-propriate, focusing on fault lines and the ag e ofthe substrate. Within any geological age therecan be a huge assay of different kinds of rock,often with radically different effects on thevegetation e.g., quartz sandstones and lime-stones). But we have no maps indicating thekind of rock exposed on the surface nor at thelevel where plants have their roots, and we lackan analysis of these substrates for the charac-teristics that are important to the soil and or-ganisms above them. Most soil maps based onsoil samples fail to take into consideration theeffect of the vegetation itself on the soil, andthe importance of deeper layers that arereached by plant taproots.

Similarly, climatic maps rarely provideinformation of importance to the organism. To-tal rainfall, for example, is not as important asits frequency du ring the year, its variation fromyear to yeas, its extremes, the cloud cover andwind over an area, the amount of fog precipita-tion, and draining or flooding once the precipi-tation reaches the ground.

More research and mapping of this na-ture, none of it very easy, will tell us much ofwhat we need to know about the distributionand maintenance of biotic communities impor-tant to conservation. Short of that, the plantcommunities themselves can be the best indica-tors.Follow up Inventory and Field GuideProductionThe Madidi-Tuichi region, with its consider-able importance to conservation in Bolivia, isclearly a priority for thorough inventory. Fieldguides developed to identify the differentgroups of organisms and describe the commu-nities in this area will almost certainly coverthe great majority of species in the rest of thelower yungas in Bolivia, as well as the south-ernmost populations of many species knownmai nly in coun tries to the north-but with abetter chance of survival in Bolivia. Any fol-low-up in providing such reference tools willhave an impact far beyond the Madidi area it-self.



18/113

Technical Report

ALTO MADlDl REGIONFrom 18-3 1 May 1990, the RA P team surveyed lowland evergreen forestsat Alto Madidi, a lumber camp on the south bank of the Rio Madidi about20 kilometers south of the Peruvian border and 100 kilometers northwestof the nearest small town, Ixiamas. This camp was a perfect base forstudy, being situated in lowland forest at the base of the Sessania del Ti-gre, the easternmost r idge of the Andes. The cam p was within walkingdistance of a variety of forest types, including young river-edge forest,more mature floodplain forest, and older forest on ancient river terracesand slop es of hills and higher ridges to the south an d west.

Physiography of Alto Madidi, Bajo Tuichi, and the FoothillRidges R. Foster)RidgesAccordin g to the available geological map s, the foothill ridges that reachabout 1,000 m altitude are compo sed of Ordov ician, Devonian, and Creta-ceous rock layers pushing up through ~~ nd ul at in g ills of Tertiary age.The old er rock is of the same age as that forming the hig her up to 2,500 Inaltitude) mid-elevation ridges of the middle yung as. Th ese are separatedfrom the foothill ridges by a broad Tertiary trough know n as the Madidi-Quiquibey Sincline, along which pass the Alto M adidi, lower Tuichi, andQuiquibey rivers. The Rio Beni bisects this trough between Rurrenabaqueand the Serrania Chepite.

The foothill r idges are mostly very steep on both flanks, formingknife-edg e crests. Most higher ridges exposed by landslid e hav e a palewhitish or yellowish color in contrast to the red of the Tertiary hills. Butthe rock is far from uniform in com po sit io ~~ . arge portions of the main

C O N S E R V A T IO N I N T E R N A T I O N A L Rapid ssess~ilent rogram


19/113

ridge and smaller outlying ridges are conlposedof a quartzite-like material often revealed inrectangular blocks on the tops of ridges, andwith considerable eroded sandy sedimentaround the base. Recent landslides may ac-count for somew here between a fifth to a tenthof the surface area.

HillsThe low Tertiary hills and terraces are mostlycomposed of a distinct dark pinkish-red clayeyrock. On exposed banks of the high, non-flooded Quaternary terraces, a few meters ofred Tertiary strata ar e always visible below theupper layer of gravel and boulder sediment.The rivers are apparently wearing downthrough the recent sediment and into the oldermaterial while simultaneously the Tertiarystrata are being lifted up by the same tectonicforces that are raising the foothill ridges.Anomalous within the red hills are a few ero-sion-resistant shields of what appear to besandstone or quartzite e.g., the Serrania delTutum o southwest of Ixiamas).

Where the Tertiary strata are compressedand subject to faulting, they are raised inblocks at a steep angle forming irregular hills.These blocks of soft red material are especiallyprone to landslide. Throug hout the area of Ter-tiary hills e.g., those along the east side of thelower T uichi), an impressive one-third of thesurface area consists of exposed recent land-slides or young successional vegetation onlandslides. Tho ugh this could have beencaused by severe earthquake, it is more likelythe consequence of the angle and softness ofthe strata and is probably a permanent dynamiccondition.

AlluviumThe hills are interrupted by landslide alluviumfrom higher ridges and reworked sedimentfrom the current river systems. The sedimentis of course derived from both the high ridgesas well as the hills, and tends to sort out or nix

R P Working Papers O n e

together differently depending on the condi-tions of deposition. This results in a mosaic ofsoils, at least in the recent alluvium, the mostobvious difference being between sandy andclayey soils . Th e abundance of sandy beacheson the Madidi is a testament to the importanceof quartzites or sandstones in the surroundingridges. Th e lack of meand er formation in therivers draining the Madidi-Quiquibey Sinclineis attributable to both the slope of the drainageand to its confinement by ridges and hills.

The river beaches in valleys with onlyweakly developed meanders are nearly static.They flood many times a year during the rainyseason, which eliminates the temporary vegeta-tion, but they only rarely or slowly form leveesfrom which a permanent vegetation can de-velop. More frequently, the river changescourse abruptly, forming an island, leaving alow abandoned channel, or leaving a broad pileof landslide rubble.

To the northeast of the foothills, the ero-sional sediment abruptly spreads out into a se-ries of overlapping alluvial fans formed bylandslides and by all of the streams drainingfrom and through the foothills . These fans ofwell-drained alluvium form a band of variablewidth along the base of the foothills, inter-rupted by inundated or poorly drained areaswith increasing frequency at greater distancefrom the foothill ridge. As is typical on alluvialfans, the streams draining the ridge constantlymake major jumps in their courses. They areappropriately referred to as arroyos since theyare frequently dly, having only small rainfallcatchment areas and considerable undergroundflow in the loose rocky soil.

Plant Communities of Alto Madidi,Bajo Tuichi, and the Foothill RidgesR. Foster

Beach and early riverine successionThe annual beach community of herbs andseedlings of woody plants only flourishes to-

December 99

Throughouttheh lkalong the eastside of the lowerTuichione-third of thesuij4me areaconsists ofexposed recentlandslides orYoungvegetcrtion onlandslides tsprobably a

permanentdyiucmiccondition.


20/113

ward the end of the dry season. We we re notable to determ ine if this was a particularly richor poor community based on the tiny plantsstarting to appear on the banks of the AltoMadidi in late May. Nearly all of them wouldbe weedy species of minimal conservation in-terest because of their pre-adaptation for colo-nizing human clearings.

The successional flora of river depositsalong the upper Rio M adidi is typical of mostof upper Amazonia, starting with the fast-growing treelets Tessal-ia ir~tegi.ifolia,Bac-charis sal ic i fol ia , Sal i s kun~boldt iai ia , andGynei.ium sugittatum. In the later stagesthere is a predominance of such species asbalsa Ochromcr pyramidale) that succeedbetter on the sandy alluvium predominatingon these rivers. Cecropia membi.anacea,more abundant on mud or silt beaches, is pre-sent but not common. Som e species absentf rom the Alto Madidi s tudy area may in par tbe limited by the paucity of finer silt depos-its. In any case, such forests are neither asabundant or as species-r ich as one would ex -pect to find on a more-meandering, silt-de-positing river system . From the overflight itappeared that the lower Madidi would besuch a system.

What d is t inguishes the area f rom theaverage young f loodplain forest on meander-ing rivers is the absence of strong dominanceby the canopy species Ficus insipida and C e -dizl a odoi-ata,and a greater representation ofthe smaller interstitial species such as Acacialoretensis, Nectandra i.eticulata, Tei.minaliaobloilga, and species of Inga, Erythrina, andSapium. This is to say that the communityhad much more evenness in relative abun-dance among the component species . Manyof the scattered successional forests on theAlto Madidi f loodplain may result f rom thesudden destruction and deposition followingmassive landslides in the headwaters or radi-cal shifts in river course. The relative inabil-ity of the fig and cedro to colonize in

CONSERVATION INTER NATIONAL

abundance may be somehow related to d isper-sal and es tablishment problems on such sub-strate.

On the upper pebbly and rocky beachesa more s tab le community develops, consis t-ing usually of 1nlpei.ata grass , the shrubsCa1liar1dr.a ar~ g~r stifo liand Adenaria f lori-bunda, and on s teeper banks, the tree Pithe-ce l lobi~rm ongi folium. Thickets of bambooGuadua sp .) occur locally o n the f loodplain ;

they are associated with areas of forest disin-tegration where floodwaters spill over a bankwith sufficient force to take down trees ordeposit a smothering layer of sediment. Th edense, spiny stands seem capable of persist-ing in one spot for many years, probably un-til the population finally flowers or getsswep t away by the river. Thes e thickets onlow floodplains could have been caused byhuman clear ing but we found no evidence forth is to be so . However , the bamboo th icketson h igher f loodplain ter races d id seem to b efrom human intervention, the only source offorest d is turbance large enough for bambooto become established with in th is commu-nity.

Older floodplain forestThe more mature f loodplain forest , h igherbut still subject to occasional flooding, iss imilarly not as abundant or as r ich in spec iesas on more extensive f loodplains to the nor thaway f rom the mountains . I t would , how-ever, be misleading to imply that it is impov-er ished , s ince there are hundreds of woodyplant species, and it is significantly richerthan floodplain forests seen in the Beni to thesouth (Foster, unpubl.) . Nevertheless, thisforest is mostly dominated by a single spe-cies, Poulser~ia ar.nlata, among the canopytrees, with occasional individuals of the ex-pected large Dipterys, Hura, Ceiba perltaiz-dra , Bi.osimum alicastr um, and Sloarzea cf.obtusifolia. In the understory Astr.occrryun~macr.ocaIy,~ redominates, but with conspicu-



21/113

ous numbers of Socratea esor.l.hiza 1riar.teadeltoi dea Otob a yarvi folia 11.ya11therajurense TrYchilia pleea na a nd Quarcir.ibeaw i f f i i .

High terraces and low ridge slopesProbably the most species-rich forests in all ofBolivia are on the ancient river terraces and es-pecially the slopes of the hills and ridges. Inmost areas these would be separate physi-ographic entities with sufficiently distinct flo-ras to be treated as distinct communities. Heretheir close proximity and overlap in flora jus-tify lumping them together.

The terraces are more uniform in topog-raphy than the slopes and as a consequ ence arealso more uniform in community compositionthroughout, and apparently less species-rich.Virtually all the species of the terraces werealso found on the slopes, but the reverse wasfar frorn true. The terraces are probably morediverse than they would otherwise be becauseof the seed rain coming in from the slopes.

The slopes, in addition to frequently hav-ing a year-round source of groundwater fromhigher up, have m ore heterogeneity in m oistureconditions, soil conditions, light conditions,and a patchwork disturbance regime in theform of landslides and lateral slumps alongstreams. The latter not only permits the long-term survival of numerous rare species fromchance colonization of the clearings, but it alsoprovides the conditions long-term minimalcompetition) for many of the well-dispersedfloodplain species to get established. An exam-ple is Poulsenia which is found down on thefloodplain and up on the slopes, but not on theintervening high terraces.

These forests have unusually large adultpopulations of Ficus spherzophylla. Theseenormous-crowned, large-fruited trees seem tooccur at a density of one per two or three hec-tares througho ut the area. While not register-ing as abundant along transects by counting thenumber of tree stems greater than 10 cm in di-

ameter, they make up a significant fraction ofthe biomass and crown area, and probably ofedible fruit production. Many of the tapir scatsencountered were made up principally of fruitmaterial of this species, and Ficus spheno-phylla fruits are likely an important food re-source for many other vertebrates and insects,especially as their fruiting is not confined toone season.

Other conspicuously abundant large treesinclude Apuleiu Cedr.elinga CopaifernHlrber.ode/~d r.on Hyerolzinza alchor~neo idesManilkar.a Parirzar.i Pterygota SterculiaTachiguli and Terr.agastris. Abundant smallertrees are Apeiha nzenzb~.anncea Batocarpu san~uzonicus rzgu various Lauraceae, Pour-ounza minor- Tetr~or~chicliunznd Virola culo-phylla. Abundant treelets and shrubs areCayl.7ar.i~ sol a Coussareci sp., Hirtella 1.0-cernosa Hyospathe elegans Pa1icour.e~purzicea Pausandra tr.iarzue Perebeu hlmzilisPiper augusturn Piper obliqun~ leu~.othyriunzkrxkovii Sipar.~i~za ecipiens and Stylogynecauliflora.

On the alluvial fans of the larger drain-age canyons at the base of the main foothillridges e.g., near Ixiamas), the distinction blursbetween floodplain, terrace, and slope comm u-nities. Th e species distinctive of each are oftenfound mixed together, solnetimes even withridgetop species. Species distributions on theseareas tend to be extremely patchy, often withlocal dominance by just a few species evenwhen the species-richness of the whole area ishigh.

A. Gentry): Structurally, the Madidilowland forest is fairly typical of Amazonianforests . Th e most unusual s tructural feature isthe relatively high density of lianas 93 lianas2.5 cni diam. in 0 1 ha vs. a Neotropical aver-age of 69). Also noteworthy is that many ofthe climbers are hemiepiphytic, with the fernPolybotrya the second most common speciesin the entire sample) especially prevalent. Thisis a feature usually more closely associated



22/113

with cloud forests than with Amazonian for-ests. The implication is that the Madidi areaenjoy s unusually high rainfall. Th e density oftrees 10 cm dbh is also greater than normal inAmazonia . Man y of the larger trees are palms,with seven individuals of Il.ial.tea and six ofEutelpe making these the two most commontaxa of trees 10 cm dbh. This structural preva-lence of palms contrasts with their relativelylow diversity.Wet ridgetopsW e were unable to visit any of the highest foot-hill ridgetops which approach 1,000 m in alti-tude. On a somewha t lower (600-700 m)ridgetop at the northernmost tip of the outlyingrange near the Alto Madidi (a ridge summittouched by clouds intermittently, though withfew large trees), the woody plant compositionwas almost indistinguishable from that of thelowe r slopes. Th e most striking difference isthe abundance of epiphytes and the moss coveron trunks and branches on the ridgetop. Veryfew species of trees from mid-elevation wetforest were found (Clusia, Calyl~tr.anthes,andanother Myrtaceae).

The higher ridges probably have thesame species composition, but with an increas-ing frequency of mid-elevation wet forest spe-cies and an increasing density and diversity ofepiphytes.

Dry ridgetopsOn the steep slopes below the wet ridges andon lower ridgetops that do not reach the cloudlevel, there is a distinctive community ofplants. Som e of the species (e.g., Rino rea vi-ridifolio, Mou1.ir.i nzyl.tilloides) occur in thelower areas but have their peak abundance onthe dry ridges. Man y others seem to be foundonly in this habitat. Th e most characteristicplants are a new tree species in the Sterculi-aceae that was recently described by Gentry asReevesiu smithii and previously collected onthe foothills to the south in Beni, and several


shrub species mainly in three families: Ruta-ceae (Eryth~.ochito~z,Galipea, Esenbeckia,etc.), Euphorbiaceae (e.g., Acirloton), and Viol-acea e (Ri1zor.e~ spp.).

My suggestion as to why this communityis distinct is that it supports species with un-usual drought tolerance. Th e drainage, windexposure, and the annual lack of precipitationduring several months of dry season probablyimpose severe water stress on seedlings as wellas adult plants, and only the most tolerantplants survive. Many of these plants are appar-ently unable to compete in the more mesic con-ditions elsewhere. Th e only other feasibleexplanation is that these ridges have some un-usual geochemica l composition. However,judging from other places where I have seenthese same species growing, the first hypothe-sis seems the more probable.

(A. Gentry): Th e forest on the low ridg-etops that form the first row of Andean foot-hills is distinctly differe nt. A 0.1 ha transectsample between 360 and 380 m on the firstridge was made and compared with the transecton the lower slopes. Incredibly, there is anoverlap of only about 23 spp. out of the ap-proximately 200 spp. in each of these samples(Identifications in ~ Und Lauraceae couldmodify this numb er slightly). In well-knownand speciose families like Bignoniaceae andPalmae, none of the species occur in both habi-tats.

(A. Gentry): Th e ridgetop forest differsfrom the lowland forest not only at the specieslevel, but also in the relative importanc e of dif-ferent families. Although Leguminosa e (18and 20 spp.) is the most diverse family in bothsamples (as nearly everywhere in Amazonia),the next most speciose fam ily on the ridgetop isRubiaceae (ca. 18 spp. vs. 3 in the lowlandsample). This remarkable diversity is morecharacteristic of premontane than of lowlandforests. Other speciose families noticeably bet-ter represented on the ridgetop are Euphor-biaceae, Bombacaceae, and Sapindaceae.



23/113

Several interesting montane elements foundonly on the ridge include Aiphn~zes Con-h z i n e u and Styr.us.Quartzite ridgetopsOn the ridges or portions of ridges with quartz-ite outcrops, there is a high proportion of gen-era and species not seen anywhere else in thefoothill region. The trees are short with small,tightly packed crowns of uniform height, giv-ing the ridge a bald or slick look from a dis-tance. Many of the taxa, such asGI-ujjferlr.iedia paristhmiiml Frezier.cr Stjlr.~ ~Aspletzizrnl rutcrceunz and Mupro~rnen areknown to be associated with m ore acidic, nutri-ent-poor soils. The upper slopes of theseridges are dominated by two species of Aspi-clos11er.nlcr and Hurrzir iastr-iml trees, and theshrub layer is dominated by Geonoma clever-saand an unusual, small-leaved Psychotr-ia.Among the other distinctive plants throughoutthe knife-like ridges are the lianas Distictelluand Br-edenzej1er.a.

From their alti tude (500-700 m) and po-sition with respect to the main foothill ridge,the quartzite ridges visited on this trip wouldotherwise have qualified as dry ridgetops,though with very few species in common withthe latter. Perhaps the substrate has differentwater retention characteristics than otherridges. In any case, i t remains to be seen whatwould be the community composition of awet quartzite ridge higher up in the clouds.

Phytogeography of Alto Madidi, RioBajo Tuichi, and the Foothill RidgesA. Gentry)

Floristically, there seem to be s ome m inor pe-culiarities associated with the Madidi forests.Although the forest is composed of typicalAmazonian elements, with Leguminosae andMoraceae being the most speciose families rep-resented in the transect sample (followed byBignoniaceae, Lauraceae, Sapotaceae, Melas-

tomataceae, Meliaceae, Myristicaceae, Myrta-ceae, and Chrysobalanaceae); some intrinsi-cally tropical families such as palms,Annonaceae, Co nnaraceae, Lecythidaceae, andPiperaceae are relatively underrepresentedcompared to adjacent Peru. Others like Sapo-taceae, Meliaceae, Euphorbiaceae, and suchsouthern families as Myrtaceae an d Proteaceae,are unusually well represented. To what extentthese may be sampling artifacts, perhaps asso-ciated with edaphic peculiarities of the limitedarea surveyed, or to what extent they representbroader biogeographical patterns remains un-clear. For several families like Rutaceae (3spp.), Moraceae (18 spp.), and Bignoniaceae16 spp.), the Madidi lowland forest 0.1 ha

sample is amo ng the most species-rich y et sam-pled in the world. Th e unusually well-repre-sented families are generally those associatedwith relatively rich soils, and it is likely that theMadidi soils are relatively fertile for Ama-zonia, which may also account for the highmammal biomass of the area.

Some of the plant species are rare ones,many are new to Bolivia, and a few are new toscience. A new Arr-uhiduea (Gentry 70382)will be described as A . uffiirzis a new Distictis(Gentry 70258) as D. occiclentalis. Severalother Bignoniaceae a re new to the country. Atthe generic level, the palms Wettirzia and Wend-lurzeliella are also new to Bolivia, as are thegenera Ar~thocliscirs f the Caryocaraceae, andPter-ygota of the Sterculiaceae, the latter beingquite abundant here. Even many of the com-mon species such as Aspidosl~e~.nzuanzboput-ense (the latter only recently described fromPeru) have apparently never been collected inthe country before.

Human Impact on VegetationA. Gentry and R Foster)

Except within a kilometer or (at most) twofrom the airstrip and sawmill, the forest ap-pears to be virtually undisturbed, unless ma-

RAP Working Papers One ecember 99


24/113

Theconservationimplicationsare obvious:Protection of alarge section ofrainforest inthis region w illpreserve asgood a sampleof biologicalreal estate aswouldconserving amoreequatorial one

hogany (or m ara, Swietenia macrophylla inthe Meliaceae) had been removed sometimeearlier when it was a military prison camp.This area was remote to begin with and themere presence of the prison may have consid-erably discouraged lumbering. Ironically, thelocation of the sawmill was based on the misi-dentification from the air of the numerousmara macho (Ce dre lhg a carenaefor.nzis, Le-

guminosae) which greatly resemble mahogany.Once constructed, the mill was mainly used tosaw mahogany brought in from the Amazonplain to the east. Th e abundant remains of for-mer prisoners and associated rumors of ghostshave also had a considerable effect on keepingthe lumbermen from wandering or hunting farfrom the camp.Ced/.elinga is a very valuable andsought-after wood in the rest of the upper Ama-zon, and it is now greatly reduced in most ofLatin America. But in this area, it has onlybeen cut dow n right near the sawmill. The rewere several large intact trees right along themain entry road. Amo ng the large trees in thetransect sa mple is a Cedr-ela 52 cm in diameter.Apparently in this region of Bolivia, one of thelast places where mature mahoga ny trees stilloccur in appreciable numbers, harvesting pres-sure on even the second most valuable hard-wood species has been negligible to date,another argument for the conservation impor-tance of th e region.

Finally, the high density of large lianas(5 individu als greater than 10 cm in diame ter in0.1 ha) provides an independent indicator thatthe forest is old, since an abund ance of large li-anas is probably the single best physiognomicindicator of very old o r primary fo rest.

Plant D iversity R. Foster a n dA. Gentry(R. Foster): The unusually high plant speciesdiversity of the Alto Madidi area results fromthe close juxtaposition of different floras, in-


cluding those of the current floodplain on theflatlands and those of the adjacent foothillr idges. We made note of 113 families, 528genera, and 988 species in a little over a week.(See Appendix 7 .

The floodplain flora alone at AltoMadidi is probably significantly less rich inspecies than floodplain forests further out onthe Amazon plain, even along the same river.The ridges and slopes constrain any large de-velopment of floodplain forest; meanders arelimited and oxbow lakes are rare. This con-finement probably restricts dispersal into andmaintenance of species on the active flood-plain. The recent floodplain of the upper RioMadidi, while not poor in species, does nothave the richness of other larger meanderingrivers I have seen to the north in Peru.

However, the adjacent hills and higherridges provide numerous small refuges and dis-turbance opportunities maintaining a largeplant species pool that is available to colonizethe nearby lower slopes and the non-inundatedhigh floodplain terraces. They also provide aheterogeneous soil derived from the alluvialmixing of two o r more distinct geological sub-strates. Consequ ently these areas have an es-pecially high diversity of plant species incomparison to areas of extensive and uniformold floodplain terraces without adjacent hills,and are comparable in r ichness to almost anyother upper Amazonian forest.

While neither the r idge flora nor the re-cent floodplain flora itself may be unusuallyrich in species in comparison to similar habitatsfarther north in Peru, the mixing of these flo-ras, in the intermediate habitats which domi-nate this area probably accounts for part of thehigh regio nal diversity.

Except for on the ridgetops, the precipi-tation in this area does not appear to exceedto 3 m per year. Even if the total rainfall ishigher, the land is probably still subject to aprolonged annual dry season of 3 to months.The epiphyte load on the trees appears little



25/113

different from that along the Rio Manu, Peru, lated to the unusually high precipitation at thewhich has an annual rainfall of less than 2 m. base of the adjacent corner of the AndeanThe higher r idges (above 900 m) clearly re-ceive more precipitation, at least in the form offog drip from the frequent clouds; mosses,leafy liverworts, and filmy ferns are abun dant.Lower slopes-benefit from this higher precipi-tation on the ridgetops through a year-roundabunda nce of groundwater percolating downfrom above. The local habitats most subject todrought are probably the lower r idges andhigher hills that d o not reach cloud line, andsecondly, the high, flat terraces on sandy-grav-elly alluvium. The se areas probably suffer asevere drop in water table when rainfall de-clines during thehabitat may getmorning low fog.

dry seaso n, though the lattersome moisture from early

(A. Gentry): The most str iking aspect ofthe Rio Madidi forest is its high diversity. The204 species greater than 2.5 cm dbh in a tenthhectare transect-sample of the forest on the lowrolling hills just behind the floodplain is ashigh as in the Iquitos, Peru area. It is signifi-cantly higher than the values for equivalentsamples in adjacent Madre de Dios, Peru,which average about 140 species. In general,forests from areas with stronger dry seasons areless diverse, and in Amazonian Peru the moresoutherly forests are the least diverse. On ecould therefore assume that the B olivian forestswould be (relatively) floristically depauperate.Thus the extremely high (plant community) di-versity of the Alto Madidi area is a distinct sur-prise. Although actual data are scarce, the highplant diversity of the Madidi area may be re-

foothills. Th e conservation implications areobv ious : Protection of a large section of rain-forest in this region will preserve as good asample of biological real estate as would con-serving a more equatorial one. Data from planttransects corroborate the high diversity notedin the area (See Table 1

Birds of lto Madidi T. ParkerThe avifauna of the lowland forests in the AltoMadidi area was found to be unusually r ich andis similar to those of tw o well-studied localitiesin nearby southern Peru, the Tambopata Re-serve (Parker 1982; Donahue and Parker, un-publ. data) and the Cocha Cashu BiologicalStation in Manu National Park (Terborgh et al. ,1984; Terborgh et al. , 1990). The latter lists,both of ca. 540 spp., are based on inventoriesin areas of roughly 5,00 0 and 1,000 ha, respec-tively. Th e area surveyed at Alto Madidi overa period of only 14 days, consisted mainly of atransect ca. 6 km-long by 200 m, throughyoung river-edge forest (ca. 100 m), maturefloodplain forest (ca. 400 m), an d older foreston somewhat hilly alluvial terraces (ca. 5.5km).

In this small area we recorded 403 spe-cies of birds (Appendix 1). This num ber prob-ably represents about 9 5 percent of the residentbird community. Based on more prolongedfieldwork at Tambopata and Cocha Cashu, wepredict that an additional 75 species will even-tually be found at Alto Madidi, including a

TABLE 1. VEGETATION TRANSECTS LOWLAND MOIST FORESTS.At ea ch site, sum of 10 transects, each 2x5 m, includes plants with stems diameter 22.5cm, at breast height.f Total Total Liana Liana Tree Tree Trees Tree z

fam. SPP. ind. SPP. ind. SPP. ind. 10 cm d h lOcm dbhSPP. ind.

Avg ois~ orest 46 152 373 35 68 116 3 4 42 64adidi 61 2 4 434 53 93 151 341 56 8

RAPWorking Papers One ecember 99


26/113

small number of uncommon or rare residents, most of which occurred in river-edge forests.and a larger number of austral and Nearctic mi- Four aerial species could not be assigned to agrants. The absence of oxbow lakes and theirmarshes accounts for the few waterbirds on thelist; the Tambopata and Co cha Cashu study ar-eas include large oxbows with numerous resi-dent and transient water and m arshbirds.

The high avian diversity at Alto Madidiwas not unexpected . It reflects both the habitatheterogeneity of the region, especially the com-bination of riverine and hill forest habitats in asmall geographic area, as well as the generalspecies r ichness of upper Amazonian forestsnear the base of the Andes (see Terborgh 1985for a discussion of additional causes). A break-down of Alto Madidi bird species diversity byhabitat reve als that upland (ter.r.cr ,firme ) forestsupports the richest com munity (1 82 spp .), fol-lowed by floodplain/river-edge forest (143spp.), marsh/water birds (33 spp.), low second-growth (21 spp.) , and migrants (20+ spp.) ,

habitat. A compa rison of bird diversity withinter.r.rrfir.n~e nd riverine forests in other parts ofAmazonia (Table 2) shows the Alto Madidi for-ests to be equally diverse, and the total list ofresident forest species to be almost as high asthose available for the richest known sites.

The total bird list would be m uch higherif we had included lower montane forests (at900- 1200 m) on the ridges within 15 k m to thewest of the st ~t dy ite. At these elevations onthe Serrania Pilbn, a southerly extension ofridge just a few kilolneters to the south of theAlto Madidi camp, Parker recorded (duringfieldwork in June 1989 43 bird species notlisted in Appen dix 1. Man y additional specieso c c ~ t r t even higher elevations to the west (seeAppendix 4), ant1 in natural grasslands alongthe Rios Heath and Madidi to the east (see Ap-pendices 2 and 3 .

TABLE 2 BIRD SPECIES RICHNESS AT 6AMAZONIAN FOREST LOCALITIES OF COMPARABLE SIZE. SPECIES TOTALSSites Upland forest Floodplain forest Total Latitude ReferenceLimoncoch a, EC 19O+ 513 2O 55 's Pearson et al. 1977Sucu sari, PE 207 50 3' 16 's Parker unpubl. listYanamono, PE 220 510 3'23's Parker utipubl. listVainilla, PE 20 328 3'46's Robbins et al. in press, Parker unpubl. dataRio Shesha, PE 216 360 8 09 's O'Neill et al. unpubl. listMucden, BO 207 288 lo 00 's Remsen and Parker unpubl. listCocha Cashu, PE 217 526 ' 11 51'S Terborgli et al. :984Tambop ata, PE 196 142 554 12' 3 6 's Parker unpubl. listAlto Madidi, BO 182 143 403 13' 10 's Parker unpubl. listHuanchaca, BO 192 14' 00 's Bates and Parker unpubl. listCachoeira NazarC, BR 236 82 447 lo 0 20 's Stotz and Schulenberg unpubl. listFazendas Esteio 23P. Alegre, Diamona, BR

i 59 's Stotz and Bierregaard 1989

Reserva Ducke, BR 209 35 2O 55 '~ Willis 1977, Stotz in litt.Raleigli Falls, SU 225 362 4O SO'N Davis 1982, rnirneographed listItaituba, Rio Tapajos, BR 190+ 35O+ 4O SO'S Parker and Schulenberg unpubl. reportAltamira, Rio Xingu, BR 15O+ 260 3O 20 's Grav es and Zusi 1990*Additional species restrictetl to floodplain forestI Upland species poorly knownCountry abbreviat ions : BO-Bollv~a, R - B r a z ~ l , C-Ecuatlol..PE-Peru, SU-Sul.iname.

CONSERVATION INTERNATIONA L Rapid ssessment Program


27/113

We predict that mo re than 1,00 0 birdspecies, or an amazing 11 percent of all birdspecies on earth, will eventually be recordedalong a transect from the Andean grasslandsnear Lake Titicaca to the lowland forests andsavannas near the mouth of the Rio Heath.

Our fieldwork at Alto Madidi revealshow little we know of bird distribution innorthern Bolivia (and birds are the best-knownvertebrate group ) . Of 403 species found, ninewere new to the Bolivia list, 30 w ere recordedin the country for only the second time (thefirst records of these were reported by Parkerand Remsen in 1987), and 5 2 species were newfor the department of La Paz.

Species new to Bolivia include an in-con spic uou s pa rrot (Nar~no l)sittcrcu lcrchillecre)recently discovered in southern Peru (O'Neillet al. , 1991), the little-known and uncommonScarlet-hooded Barbet (Ellbucco tricinkrre),Crested Foliage-gleaner (Autonlolus clor.scrlis),Rufous-tailed X eno ps (Xer~ ops lillei.i), Undu-lated Antshrike (Fr-eclerickena ~lnd~tligeru),Ash-breasted Gnateater (Cor~opopl~agcrer.uvi-arlcr), Olive-striped Flycatcher (Mionectes oliva-ceus), Black-and-white Tanager (Conoth i.a~rpissl~ec~iliger.cr) , and Casqued Oropendola(Cly/>icter.us osei yi). By far the most unex-pected species observed during our days inlowland La Paz was an Arctic Tern (Sternerparailisaea) found and photographed atLaguna Santa Ro sa along the lower Rio Tuichiduring our reconnaissance of that area on 24May. Not only was this species previou sly un-known in Bolivia, it is also the first record ofthis pelagic tern for the interior of South Amer-ica.

Interesting features of the resident avi-fauna at Alto Madidi included the presence ofnine species of forest tinamous, an unusuallylarge number for su ch a small area, and 16 spe-cies of parrots. Orange-ch eeked Parrots(Pionopsittu bar.i.abancli), although here re-ported for only the second time in Bolivia,were unusually common; we repeatedly ob-

RAPWorking Papers One

served small groups in the canopy and flyingoverhead, and occasionally noted as many as12 in one flock. The abundance of this andseveral other parrot species may h ave been re-lated to the presence of large numbers of fruit-ing fig trees (Ficus spherlopl~ylla describedabove). Red-and-green Macaws (Al-ir cl~loi-011-te1.u) were also comm on. Large cracids, in-cluding heavily hunted species such asRazor-billed Curassow (Mitu tuher.osa) andSpix ' s Guan ( Per ~e lope acqucrc~c), were alsorelatively numerous.

The large number of antbirds (42 spe-cies) found in the study area reflects the mixingof foothill (e.g., Mjlrn~ecizrrortis) and riverineforest sp ecies (M)rr.rneciza goelclii). So me spe-cies, such as the Rufous-capped Antthrush(Foi.rnicai.ius coln~ir),were inexplicably scarce,whereas others know n from nearby areas to thenorth (e.g., M yr.n ~o the iul a oi.rzcrtu) were notfound at all.

The Alto Madidi avifauna, especiallywhen combined with that of the montane for-ests close by to the west and that of pristine sa-vannas not far to the northeast along the RioHeath (see Appendix 2), is unquestionably therichest known from any region in Bolivia, ifnot all of South America. Furth ern~ ore, bout10 percent of the bird species found in these ar-eas are endemic to a relatively small (


28/113


29/113

high numbers at Alto Madidi are of consider-able importance fo r conservation.

There are large populations of KinkajousPotos flavus) and O lingos Bassar. icyon gab -

bii) at Alto Madidi. Both these and spidermonkeys may benefit from the unusually highnumbers of enormous, free-standing figs Ficussphenophylla) in this forest.

As is typical of most of lowland Ama-zonia, the terrestrial, nocturnal fauna is domi-nated by spiny rats Proechinzys spp.). Twospecies were collected, but the habitats werenot adequately sampled for rodents, and otherspecies of Proechinzys may occur. Many othersmall mammals are to be expected.

Of special interest for conservation weretwo sightings, by Parker and Castillo, of aShort-ear ed Dog Atelocynus microtis). This isone of the rarest mammals in Amazonia, withfew reliable recorded sightings by scientists.In one sighting, the dog had a large frog in itsmouth , which is the first record of food habitsfor a wild individual. Although short-eareddogs could occur in all lowland Amazonianparks or reserves from Colombia to Peru, andsouth of the Amazon in Brazil, there is appar-ently no information to confirm that they areactually present in any of them. It would there-fore be of importance for the conservation ofthis species to preserve an area where they a redefinitely known to be present.

The forest 3 km west of Ixiamas seemsto be an extension of lowland evergreen forestsimilar to that at Alto Madidi. The eight mam-mals recorded at this site are all common andwidespread throughout western Amazonia.The area seen h ad clearly been under consider-able subsistence hunting pressure for sometime.

P MP S R GIONParker and Castillo found 35 bird species inthe savanna habitats at Ixiamas, including 40

species 29 percent) not previously recordedfrom the department of La Paz see Appendix3). All of these were known in Bolivia fromonly a few localities to the east and south in thedepartm ent of Beni. Thes e discoveries reflectthe paucity of distributional data available forBolivian grassland vertebrates. Of mo re thanthree large areas of grassland in lowland LaPaz, not one had been previously inventoried;in fact, only one reliable bird list is availablefor the ca. 100,000+ sq km of natural, lowlandgrasslands in the country see Remsen 1986).The ecological distributions of grassland birdspecies within the complicated mosaic of sa-vanna plant communities see Haase and Beck1989) are largely unknown.

At Ixiamas we found several grasslandbird species that are declining throughout mostof their ranges in central South Ame rica e.g.,Cock-tailed Tyrant, Alectl-uvus t~ .ico lo~ ., ndBlack-masked Finch, Corypkaspiza rnelanotis).That large populations of such species survivein northern and central Bolivia underscores theimportanc e of establishing reserves of one kindor another in Bolivia.

Residents of Ixiamas reported that largeareas of fairly pristine grassland and galleryforest, with f ew o r no cattle, are situated in un-populated areas between Ixiamas and the RioBeni to the east. Even large mam mals e.g. ,marsh deer) are said to survive in numbersthere. Identifying and preserving at least por-tions) of such areas should be given high prior-ity.

One pristine grassland area of significantconservation importance is situated along theBoliviaperu border on the lower Rio Heath.Our overflights of the Pampas del Heath re-vealed large expanses, possibly of as much as10,0 00 hectares, of pantanal-lik e grasslands,and cerrado-like woodlands surrounded byvery extensive riverine and upland evergreenforests. Vertebrate and plant surveys in thesehabitats on the Peruvian side of the river re-vealed high bird diversity including the only

R PWorking PapersOne ecember 99


30/113

The initialindication isthat thesevarioussavannasconstitute a farmorecomplicatedmosaic ofisolated habitatislands thanglance at themap wouldsuggest

Peruvian populations of at least 4 grasslandspecies), large populations of large mammals,and undisturbed grasslands and cerrado-likevegetation which differ floristically from thoseto the south see Appendix 2; Parker, unpubl.data; Gentry, unpubl. data).

Gentry saw only one species of plant onthe Ixiamas pampas that was held in commonwith the Peruvian pamp as. We noted 57 fami-lies, 120 genera, and 169 species in the area.See Appe ndix 10.) Beck observes that the

Ixiamas pampa has little in common floristi-cally with the savannas he has studied fartherto the east. Th e initial indication is that thesevarious savannas constitute a far more compli-cated mosaic of isolated habitat islands than aglance at the map would suggest.

The Pampas del Heath is as yet almostunexplored for mammals, but as one of theonly pristine grassland systems not yet de-graded by cattle or severely hunted and too-fre-quently burned by man, it represents one of thechief conservation priorities and opportunitiesin Amazonian South America. Marsh DeerB1nstocer.u~ clichotomus), extrem ely endan -

gered or threatened over much of their geo-graphic range, are found in the Pampas delHeath. There are also persistent rumors of Ma-ned Wolves Ckrysocyon hrachyu rus). On ouroverflight we saw that the Bolivian pampas arecriss-crossed with innumerable, deeply worntracks of large mammals, presumably made bytapirs or deer. Parts of the savann as aredensely studded with termite mounds, whichshould provide favorable habitat for anteatersand armadillos. Although its mammal list maynot be large, the Pampas del Heath may be oneof the only undisturbed natural habitats of itstype and a refuge for species that are perse-cuted elsewhere and in need of protection.


APOLO REGION, MID ELEVATIONWET FOREST CALABATEA)

Physiographyo Mid-elevationYungas: Apolo to CalabateaR. Foster)Th e middle yungas is an extensive area of highrelief mostly between 1,000 and 2,000 m, andApolo is in the center of this region. Th e areawest to the main Andes is geologically mappedas completely of Ordovician age. I t is dis-sected with an irregular reticulate drainagefrom a labyrinth of twisting and turning ridges.T o the east by contrast, the r idges are long,straight, and sharply defined, separated m ostlyby long gradual slopes and occasional vertical-walled mesas. These are mapped as narrowparallel strata of Ordovician, Devonian, Car-boniferous, Cretaceous, and Tertiary age notoccurring in any predictable order) . Apolo isin a broad, gently sloping valley bordered by asteep escarpment on its east f lank and rollinghills to the west.

The significance of this contrast betweenthe western and eastern parts of the northernyungas is not yet clear since we were only ableto fly ove r the eastern part. From the air, how-ever, the eastern part seemed to hav e consider-ably more diversity of vegetation types,presumably from the exposure of very differentgeological strata.

The Ordovician substrate, as seen in theroad cuts in the Yuyu drainage to the south ofApolo and in exposure throughout the denudedApolo Valley, is a wild variety of thin, softstrata tilted at a steep angle or even vertical.Mos t of these m ulti-colored layers are like softclay- you can stick your fingers all the wayin-and mo st of the rest are at best very brittleshale. The hardest strata seem to be the narrowbands of sandstone. These are also most likelyto be on ridgetops.



31/113

This substrate seems remarkably perme-able to water. For all the heavy rain we wit-nessed at the Calabatea Road Camp, thestreams down the slopes hardly increased. Ero-sion seemed almost entirely caused by land-slides except for along the large stream at thebottom of the valley. Th e Apolo valley floor issimilarly surprising in the lack of stream ero-sion in spite of the heavy rains, severe denuda-tion, and near abse nce of soil. Presuma bly thewater percolates through the rock. It wouldbe instructive to f ind out how far the water ta-ble drops during periods of drought, and the ef-fect this has on the plants.

The center of the middle yungas is prob-ably the driest part. Layers of moist air comingfrom the east off the Amazon plain over thelower Tuichi valley and from the southwest upthe deep river valleys of the Mapiri drainageare intercepted b y the series of high ridg es botheast and west of the Apolo area.

Plant Communities R. FosterThe elevation range observed (1,000 to 1,650m) is where the montane flora (e.g., Clethr-a,Lap lacea , Pod ocarp us, Schefflera) m eets thelowland flora (e.g. , Sim arou ba, Sloanea , Synl-phorzia, Tachig ali) . Th e overlap is consider-able, with even the palms Dictyocar-yumlamar-ckianum and Iriar-tea deltoidea growingside by side in places. Even the montane gen-era Hedyosnlunz and Br.unellia were found tobe more common at the valley bottom than onthe ridges above. Gentry considers the distinc-tion between m ontane an d lowland forest typeshere to be more abrupt than have presented.His sample trarlsect of 0 .05 ha at 1,500- 1,550m was comprised almost entirely of lowlandtaxa. Although Gentry could not sample themontane forest at Calabatea, he states that asample from Incahuara, farther south in the de-partment of La Paz in the same ridge system, isavailable and appears to represent the samevegetation type.

High ForestThe best-developed forest (up to 30 m tall) is atthe bottom of the valleys, and comes up theslopes especially on old landslide debris. Sim i-larly, species richness is greatest in this forestbut clearly decreases up slope to the ridge.While tree ferns are common and epi-phytes certainly more abundant than in lowlandforest, there was not such a superabundance ofepiphytes nor heavy moss cover to suggestyear-round high humidity except on the verytopmost r idges.

The most abundant tree appeared to beHyer-or~ima p. (Euphorbiaceae), but the forestis most easily characterized by the numeroustree species of Rubiaceae (Cirlchona and closerelatives), Melastomataceae (Topobea, Mico-nia, etc.) and L auraceae.

Stunted ForestForest on ridges is clearly shorter ( less than 20m), as is expected. But on large parts of theridges the forest is stunted with a canopy oftenno more than 5 to 10 m tall. The easiest expla-nation is that the stunting (without much differ-ence in f lora) has to d o with the water-holdingcapacity of the soil. However, nothing wasseen to substantiate or refute this idea. It ispossible that are as deliberately burned in ex-ceptionally dry years regenerate in a stuntedform. Several areas of recent burning to pro-vide forage for pack animals were found alongan old trail down the ridge. These are nowgrowing back to bracken feln (P ter id i~ m~quil-in un ~) xcept in the more settled areas wherethe burning is continued to keep an area openfor cattle pasture.

Palm ForestWhile palm species in general are few in thisarea (six seen), some ridges and slopes aredominated by the conspicuous emergent palm,Dictyocar.yum 1amur.ckianum. The se areas areso conspicuous that from a good vantage pointwith binoculars on e can easily pick out all the



32/113


33/113

cidated, but its discovery, especially if borneout as a general pattern, is of more than trivialtheoretical significance.

Prior data from a second, higher eleva-tion site at Sacramento Alto 2,500 m) in theCoroico valley, department of La Paz, are alsoavailable and are of interest in comparison withdata from Calabatea and Incahuara. At this al-ti tude, the switch from lowland to montanetaxa has been completed with a concomitantloss in species diversity to 91 spp. 2.5 cm dbhin 0.1 ha. Melastomataceae 15 spp.) and Ru-biaceae 10 spp.) are the most speciose fami-lies, just as they are at Calabatea. With areduced num ber of species, Lauraceae 7 spp.)share canopy dominance with such montanetrees as species of Alchornea Clethra Gordo-nia Hedyosmum Melio sn~ a Myrsine Rhanl-nus Syn~plocos and Weinn~annia. At thisaltitude, Compositae completely dominate theliana compo nent of the flora 7 of 16 free-climbing species). Forests such as this un-doubtedly occur above 2,000 m on the slopeswest of Calabatea.

Gentry): Th e 110 spp. in 500 m2 atCalabatea im plies that at least 14 0-15 0 spp. oc-cur in 0.1 ha, about the same as in the Inca-huara sample 147 spp.). Because of theincomplete sample from Calabatea, only a verybroad outline can be suggested with respect tospecies diversity and floristic composition.The two most diverse families are melastomsand rubiacs ca. 12 spp. each), with legumes 9spp.), Lauraceae and Moraceae 6 spp. each),and Bignoniaceae 4 spp.) following as themost speciose. The forest at Incahuara is char-acterized by a predom inance of melastom s 18spp.) and Rubiaceae 12 spp.) in the understoryand lower canopy, but differs in the over-whelm ing diversity of Lauraceae in the canopy24 spp. in 0.1 ha the greatest Lauracea e di-

versity yet found in my 0.1 ha samplesthroughout the world, and the majority of thespecies are likely undescribed ). How the ex-treme prevalence of this preeminently bird-dis-persed family might be related to the localavifauna or other elements of the biota remainsan unexplored question.

Plant diversity R. Foster andA Gentry)Foster): The floristic richness observed here is

nothing exceptional we noted 1 13 families,251 genera, and 390 species ee Appendix8). It is certainly far less than forest at similarelevations in southern and central Peru. The rewas no clearly dominant large tree, but thesame one or two dozen common tree speciescould usually be found throughout the areafrom ridgetop to ravine bottom, from high for-est to stunted forest, and from palm forest tobamboo forest. Eve n the remnan t pockets offorest on the hills adjacent to Apolo many kilo-meters away had almost the same composition.The unimpressive diversity and uniformity issurprising given that this is the transition zonebetween montane and lowland floras with con-sequent mixing of species from each.

Birds of Mid-elevation Wet ForestCalabatea) T. Parker)

The Calabatea avifauna is comprised of a mix-ture of highland and lowland species Appen-dix 4). Of 169 species found between ca.1,300 and 1,600 m, 80 species are montane,most occurring in low- and mid-elevation for-ests 900-1,500 m) from Colombia south tocentral Bolivia. The rest are lowland specieswhich were at or near the upper limits of theirelevational ranges. Ten of the mon tane speciesare endemic to the yungas region of extremesouthern Peru/northern Bolivia, as are numer-ous well-marked subspecies. Most of the mon-tane bird species found at Calabatea occurredprimarily abo ve 1,50 0 n~ in forests charac-terized by numero us highland plant genera seeabove).

RAPWorking Papers ne ecember 99


34/113


35/113

from their refuges in the ravines and riverbanks. Th e forest in these latter areas is nearlyevergreen.

The deciduous forest continues down thevalley as far as the eye can see. It is probablethat the Alto Tuichi to which this valley drainshas similar vegetation, as would the valleysparallel to and north of the Machariapo on intoPeru. It is important to explore this possibiltyon the chance there are some valleys with less-disturbed vegetation.

The overwhelmingly dominant tree (inthe transect sample, half of all trees 20 cm dbhand two-thirds of all trees 3 0 cm) is Anadenan-thera colubrina a mimosoid legume with adistinctive spreading crown noted for its hal-lucinogenic indole alkaloids. Also abundantin the canopy is an Acacia Astrorziurn Schiiz-opsis and a short Ceiba. Understory trees in-clude a conspicuous large Echinopsis? cactus,Triplaris and Capparis. Distinctive and com-mon shrubs include the startling combinationof a strange Opuntia (60 plants in 0.1 ha, by farthe most common species in the sample) withsmall, flattened joints and straight, erect stemoccasionally reaching 12 cm dbh and 12 m ormore tall, mixed with Clavija and a smallTrichilia. Epiphytic orchids and bromeliadsare abundant. The evergreen patches includetypical lowland moist forest trees such as Gall-esia Platyrniscium M yroxylo~ z Clarisia bi-flora Cecropia polystachya Ficus juruerzsisetc.

Structurally this forest is not very differ-ent from moister forests, with 465 stems 2.5 cmdbh including 77 trees, 2 lianas, and a strangler10 cm dbh. The most striking structural anom-aly is the presence of 134 individual lianas(compared to an average of 7 for all dry for-ests sampled) probably relating to a history ofpast disturbance.

Scattered individuals of commerciallyimportant timber taxa such as Amburana cear-ensis Tabebuia impetiginosa and Cedrela sug-gest that the forest may formerly have been

more diverse. My

01 RAP Working Papers_madidi

Documents

Transcript of 01 RAP Working Papers_madidi