Amazon Soils

A/# o W. G. Sombroek

Amazon Soils

MN08201 39G Dit proefschrift met stellingen van WILHELMUS GERARDUS SOMBROEK, landbouwkundig ingenieur, geboren teHeiloo , 27augustu s 1934, isgoedgckeur d door de promotor, DR. IR. P. BURINGH, hoogleraar in de tropische bodemkunde.

De Rector Magnifkus vand e Landbouwhogeschool F. HELLINGA

Wageningen, 15 december 1965 W. G. Sombroek

Amazon Soils A reconnaissance of the soils of the Brazilian Amazon region

PROEFSCHRIFT ter verkrijging van de graad van doctor in de landbouwkunde, op gezagva n de Rector Magnificus, Ir. F. HELLINGA, hoogleraar in de cultuurtechniek, te verdedigen tegen de bedenkingen van een commissie uit de Senaat van de Landbouwhogeschool te Wageningen op vrijdag 25 februari 1966 te 16 uur.

7966 Centrum voor Landbouwpublikaties en Landbouwdocumentatie Wageningen Voor Willemijti STELLINGEN

1 Door de grote mate van overeenkomst in bodemkundige en andere factoren tussen het Amazonegebied en het centrale Congobekken zijn de resultaten van het Iandbouwkundig proefstation te Yangambi (Congo) van grote waarde voor de ont- wikkelingva n delandbou w in hetAmazonegebied . Dit proefschrift II De pleistocene terrasafzettingen van de Amazonevallei beslaan een grote oppervlakte enzij n inhe t algemeenva n interglaciale, interpluvialeouderdom . Dit proefschrift III Het succes van de op initiatief van Japanse immigranten in het Amazonegebied ontwikkelde pepercultuur illustreert dat met toewijdinge nvolhardin gi ndi ttropisch e probleemgebied veel kan worden bereikt.

IV De invoeringva n het bodemklimaat alsdeterminatiekenmer k bij een morfometrische bodemclassificatie iszowe lonlogisc hal sonpraktisch . SOIL SURVEY STAFF, 1960. Soil classification, a comprehensive system; Seventh Approxi mation. U.S. Dept. Agr., Soil Conservation Service, Washington. IDEM. Eighth Approximation 1964.

V Het is gewenst de bepaling van Si02:Al 203:Fe203 verhoudingen, door koken van de grond of van dekleifracti e metzwavelzuur ,gedeeltelij k gevolgddoo r koken metsoda , op te nemen bij het standaardlaboratoriumonderzoek van tropische gronden. Dit proefschrift VI Devvijz eva nverwerkin gva nbodemgegeven sva nhe t Mapanegebied inSurinam edoo r SCHULZ, is ongeschikt om het al of niet bestaan van een samenhangtusse n het voorkomen van de boomsoort Bruinhart(Vouacapoua amerkana) en de bodemgesteldheid vast te stellen. SCHULZ, J. P., 1960. Ecological studies on rain forest in Northern Suriname. Verh. Kon. Ned. Akad. van Wetenschappen, Afd. Natuurkunde. Tweede reeks, III (1), p. 130 en de tabellen VI en VII. VI[ Het onderbroken voorkomen vand e oerbank bij Jodensavanne in Suriname isnie t noodzakelijk het gevolg van vroegere erosie van deze laag; HEYLIGERS' conclusie dat deoerban k louterfossie l zou zijn isda n ook aanvechtbaar. HEYLIGERS, P.C , 1963. Vegetation and soil of a white-sand savanna in Surinam. Verh. Kon. Ned. Akad. van Wetenschappen, Afd. Natuurkunde. Tweede reeks, deel L1V (3).

VIM Opname vand e plantenecologie vantropisch e en subtropische gebieden als erkend onderdeel van het studieprogramma der Landbouwhogeschool isten zeerst egewenst .

IX Het voortdurende gevaar van mondiale besmetting met een overmaat aan radio- actieve nucleiden maakthe t urgent omteeltmateriaa l van dedoo r eeuwenlangeselecti e verkregen agrarische kwaliteitsrassen in onderaardse ruimten, zoals verlaten zout- mijnen, beschut te houden.

X De internationale ontwikkelingshulp zal, evenals inhe tverlede n denational e sociale emancipatie, uitd esfee r vanliefdadighei d naar dieva ngemeenschapsplich t groeien en zal dan ook uiteindelijk inee n kader van bindende internationale overeenkomsten vastgelegd moeten worden.

XI Het bestaan vanee n communistisch gericht Cuba begunstigt desociaal-economisch e ontwikkeling van Latijns Amerika.

Proefschrift W. G. SOMBROEK Wageningen, februari 1966 Voorwoord/Acknowledgment s

De voltooiing van een proefschrift wordt beschouwd als de definitieve afsluiting van een academische studie. Daarom geeft deze plaats gelegenheid mijn dank te be- tuigen aan de Wageningse hogeschoolgemeenschap voor haar aandeel in mijn wetenschappelijke vorming. Hooggeleerde Buringh,d ewelwillendhei d waarmeeU i nee nz olaa t stadium voordi t proefschrift het promotorschap hebt willen aanvaarden, en Uw kritische nauwgezet- heid bij het doornemen van het manuscript stel ik bijzonder op prijs. Ik denk ook met voldoening terug aan de studie, onder Uw leiding, van de luchtfoto-interpretatie aan het ITC te Delft, welke in mijn huidige werkkring eengroo t en onmisbaar hulp- middelblijk t tezijn . Zeer geachte Mevrouw Edelman-Vlam, aan U wil ik mijn grote erkentelijkheid uit- drukken voor de bodemkundige opleiding en het wetenschappelijke enthousiasme die ik van wijlen Uw echtgenoot mocht ontvangen. Reeds tijdens mijn verblijf in Brazilie gaf hij zijn morele steun; maar vooral de bijna wekelijkse bezoeken die ik hem aan zijn ziekbed mocht brengen tijdens de uitwerking van dit proefschrift, hebi k hogelijk gewaardeerd. Ik acht het een voorrecht deze grote mens nog zo van nabij te hebben meegemaakt. Hooggeleerde Wellensiek,voora l Uw methodische benaderingva n wetenschappelijke vraagstukken blijft mij bij. Dat Uw bekende belangstelling voor het wel en wee van Uw vroegere studenten zich ook tot mij - die uiteindelijk een geheel andere richting dan de tuinbouwkundige is ingeslagen - blijft uitstrekken, stel ik bijzonder op prijs. Ook Uwpersoonlijk e belangstelling, hooggeleerde Schuflelen, waardeer ik bijzonder. Hooggeleerde Hellinga en hooggeleerde Van Lier, het zijn gedeeltelijk Uw beider colleges geweest die mij er toe gebracht hebben een werkkring te zoeken bij cultuur- technische projecten in ontwikkelingslanden.

There are many, persons and institutions, which have enabled me to make the present study a reality. My thanks to Dr.Ir .Jaa p Bennema,unti l recently a member of the Brazilian Soils Commission delegated by FAO, can hardly be overstated. He was my mentor and continuous stimulator, both duringth e latter part of the stay in Beleman d duringth e elaboration of the manuscript in Holland. His dedication to the study of the tropical soils in general and those of Brazil in particular, which may permeate through this thesis, has served as an example. Secondly, I wish to thank the co-members of the FAO/SPVEA team in Belem. Particularly Mr. Tom Day, soil scientist, who guided me during the first year and trained me in the American systematics of soil description; Dr. Paul Sutmoller, veterinarian, who tugged me along to the most remote grazing areas; Mr. Hans Huffnagel, agricultural economist; and the forest inventorists Mr. Bas Glerum and Mr. Gerard Smit,m y discussion-eager companions during the long stays in the forest. Right with them, there are the Brazilian counterparts attached to the Mission. Par ticularly dear is the remembrance of Eng. Civil Heber Rodrigues Compasso, chief of the draughting section and photo interpretation expert; Antonio Vahia de Abreu, and Eng. Agron. Jose Benito Sampaio, which latter joined me in the soil survey field work. It was my good fortune to have contact in the same sphere of cooperation and friendship with the members of the Soils Section of the Instituto Agronomico do Norte, namely Professor Lucio Salgado, and Engs. Agron. Walmir Hugo dos Santos and ftalo Claudio Falesi. If the present study is of use for their part in the agricultural development of Amazonia, then a promise is kept and a main aim of my effort fulfilled. Agradeco muito todo o pessoal da Missao FAO/SPVEA, bem no escritorio como no campo. Muitas vezes Ihes deve ter sido dificil perceber qualquer utilidade nas tare/as monotonas deabrir caminhos, de tradear, colecionar, conduzir, tracar e re-tracar, dactilografar e re-dactilografar, estencilar, traduzir e calcular. O que,apesar de tudo isto, cada um delesfez seu trabalho com diligencia e abnegacao, sempre sera para mim uma das numerosas gratas lembrancas do Brasil. Tambem e motivo deprofundo agradecimento a ajuda que me prestaram institutos em Belem e em outras partes do Brasil. Da SPVEA sera lembrada mormente o Dr. Nady Bastos Genu, do IAN o Professor Jose Maria Condurii. Outras assistencias valiosas foram prestadas numa e outra ocasidopor Rodobrds, Petrobrds, Rio Impex, Belterra Estate, Icomi, Prospec. With the members of the Federal Brazilian Soils Commission of the CNEPA in Rio de Janeiro a much cherished contact developed, ensuring the integration of the Amazon soil survey work in the great national effort. I gratefully record the help given by the IQA, where, notwithstanding its very busy programme, so many Amazon soil samples were analysed. The cordiality of Dr. Leandro Vettori, head of the Soils Laboratory, even extended to checking of the main part of the Portuguese texts of this study. During the elaboration of the manuscript in Holland, I received help and suggestions from many sides. The Dutch SoilSurve y Institute (STIBOKA),i nth e person of itsdirecto r Dr. Ir.F . W. G. Pijls, isgreatl y thanked. Although I had no previous connections with this Institute, its help was large and readily. Dr. Ir. H. W. van der Marel, head of the Soils Labo ratory, provided for very valuable additional analysis data, and Messrs. J. J. Jantzen, J. P. Heerema and Th. C. Vos of the Cartography and Map Editing Sections devoted much time to the preparation of the appending maps and some text figures. Acknowledgment is due to the Agricultural State University of Wageningen, for my appointment as temporary research worker. The Department of Soil Science of this University rendered hospitality and facilities during this time. Special thanks is due to the draughtsmen of the Department, Messrs. G. Buurman and W.F . Andriessen for their excellent preparation of the many textfigures, an d to Mr. Z. van Druuten for hisdevote d attention toth e photographic material. The soils laboratory of the Royal Tropical Institute of Amsterdam provided for someextr a analysisdata ; Ir.W .P .Stakma no fth eWageninge n Institutefo r Land and Water Management Research (ICW) kindly determined physical data on selected samples, and HVA Company of Amsterdam made available their collection of Amazonsoi ldata . Several persons helped with critical remarks on parts of the manuscript in its preliminar form. Gratefully acknowledged are the suggestions given by Prof. Dr. D. J. Doeglas, Dr. G. C. Maarleveld and Dr. D. J. G. Nota (Chapter 1); Dr. K.Sy s (Chapter 2); Ir. T. W.G . Dames (Chapter 3);Ir . D. Heinsdijk, Ir. B.B .Gleru m and Dr. Ir. I. S.Zonnevel d (Chapter 4); Ir. H. Ph. HufTnagel and Ir. J. M.va n Staveren (Chapter 5). Manythank sar edu et o Mr.R . Lyallan d Mr.T .Brook ,a swel la st oMrs .B .Dooren - bos, Mrs. E.Groenendij k and Mrs. K. Johnston. They all withstood the sweltering weather of Sokoto,Nigeria , for the arduousjob s ofcorrectin g the English and typing the final text respectively. Miss C. van Twisk of Amsterdam helpfully corrected the final Portuguesetexts . The Centre for Agricultural Publications and Documentation (PUDOC), in par ticularly Mr. M. J. M. Osse, deserves special thanks for reducing the manuscript to the book form in which it isno w presented. It must have been atime-consumin g task for him inm yabsenc et oaccommodat ewithi n a type area of 28 x 43cicer o a manu script that in its arrangement and form was still reminiscent of the vast, untamed Amazon jungle, and to build it up into a comprehensible and uniform whole, with scarcely more aid than italics and bold face indications.

The author is indebted to the Food and Agriculture Organisation of the United Nations of Rome, Italy, and to the Superintendencia doPiano de Valorizagao Econo- mica daAmazonia of Belem do Para, Brasil, for their permission to publish the data assembled duringhi sassociatio nwit hth e FAO/SPVEA Missioni nBelem . Thanks are due to the Dutch Government's Bureau for International Technical Assistance (NEBUTA), for financing the author's association with FAO during the period 1959-1962. The Landbouwhogeschoolfonds (Agricultural University Fund Foundation), the Nederlandse Organisatie voor Zuiver Wetenschappelijk Onderzoek (the Netherlands Organisation for theAdvancemen t of Pure Research) and the Centre for Agricultural Publications and Documentation (PUDOC) are rendered many thanks for their willingnesst o givefinancial suppor t for thepublicatio n ofthi swork . This thesis willals ob epublishe d asAgricultura l Research Reports 672 © Centre for Agricultural Publications and Documentation (PUDOC), Wageningen, 1966 No part of this book may be reproduced and/or published in any form, photoprint, microfilm or anyothe rmean swithou t written permission from thepublisher s Printed in the Netherlands Contents

Introduction 7

Some Termsan dAbbreviation s 9

Chapter I TheEnvironmenta lfactor s ?. . 11 1.1 The Geographic Location 11 1.2 The Climate 12 1.3 The Geology 14 1.4 The Geomorphology 18 1.4.1 Sketcho fproto-Amazoni a 18 1.4.2 The Planicie.Th e Plio-PIeistocene Plateau or Amazon Planalto ... 20 1.4.2.1 Description of Occurrence of Belterra Clay 20 1.4.2.2 Origin ofBelterra Clay 23 1.4.3 The Planicie.Th e Pleistocene Terraces 26 1.4.3.1 Origin of Terraces 26 1.4.3.2 Terrace Levels in theGuamd-lmperatriz Area 30 1.4.3.3 Terrace Levelsthroughout theAmazon Valleyproper 33 1.4.3.4 Terrace Levelsin Relatively HighSituated Areas 37 1.4.4 The Holocene Terrains 38 1.4.5 Levels of Occurrence of Fossil Plinthite (Laterite). Its Grouping and Its Dating 39 1.4.5.1 Fossil Plinthite in the Guamd-lmperatriz Area (northern half- southern half) 39 1.4.5.2 FossilPlinthite throughout Amazonia 47 1.5 The Vegetation 51 1.5.1 The Forest Vegetation 53 1.5.1.1 LowlandForests 53 1.5.1.2 UplandForests 54 1.5.1.3 SpecialUpland Forest Types 56 1.5.2 The Savannah and Savannah-Forest Vegetation 56 1.5.2.1 LowlandSavannahs andSavannah-Forests 58 1.5.2.2 UplandSavannahs andSavannah-Forests 58 1.6 Paleo Climate and Paleo Vegetation 60 Chapter II The MainAmazo n Latosolsan dPlinthiti c Soils 63 II.1 Evolution of the Concepts of 'LatosoFan d 'Laterite' or 'Plinthite' ... 63 11.2 Latosols 65 11.2.1 The Definition and Subdivision of Latosols as Applied in Brazil . . 65 11.2.1.1 TheModal Latosol 66 11.2.1.2 The Distinction of the Latosolsfrom other Soils;Latosolic-B versus Textural-B 67 11.2.1.3 The Subdivision of the Latosols (introduction; description of the subdivision of Table6) 72 11.2.2 The Main Amazon Latosol ". . 76 11.2.3 Comparison with other Classification Systems 92 11.3 Plinthitic Soils 95 11.3.1 Origin of Plinthite 95 11.3.2 Plinthite inAmazoni a 98 H.3.2.1 Plinthite-in-formation (formation ofplinthite below orin the lowest part of the solum;formation of plinthitewithin thesolum: Ground Water . Lateritesoils) 99 II.3.2.2 FossilPlinthite (introduction;fossil plinthite below thesolum; fossil plinthitewithin thesolum) Ill

ChapterII I TheSoil sClassified ,an dthei rGeographi cOccurrenc e 120 111.1 The Methods of Study 120 III.l.l Field Studies , ... 120 III.1. 2 Laboratory Studies (physical analysis; chemical analysis; mineralogical analysis) 121 111.2 The Soils Classified 125 111.3 The Occurrence of the Soils in Relation to the Various Geomorphologic Units 161 111.3.1 The Soils of the Undulating Terrains with Outcropping Crystalline Basement 162 111.3.2 The Soils of the Undulating Terrains with Outcropping Paleozoic, Mesozoic or Early Tertiary Deposits 162 111.3.3 The Soils of the Cretaceous and/or Early Tertiary Peneplanation Surfaces 164 111.3.4 The Soils of the Planicie 165 111.3.4.1 TheSoils of theAmazon Planalto 165 111.3.4.2 TheSoils of the Pleistocene Terraces (mainsoils; profile develop mentin dependence of texture andterrace level; other well-drained soils; im perfectlyor excessively drained soils; Terra Preta soil; the Pleistocene terraces of western Amazonia) 166 111.3.5 The Soils of the Holocene Terrains 178 111.4 Soil Associationsan d Land-Units 178 Chapter IV The Soils and the Vegetative Cover . • 182 IV.l The Uplands with Forest Cover 183 IV.1. 1 Evaluation of Non-Edaphic Factors 183 IV.1.1. 1 The Influence of the Climate (timber volume; occurrence of individual tree species) 183 IV.1.1. 2 TheInfluence of Man (secondaryforests; selective cutting; Indians; tabocal) 186 IV.1. 2 Soils and Forest Characteristics 189 IV.1.2. 1 Land-Units and Forest Inventory Units 189 IV.1.2. 2 Soils and Timber Volume (soil texture; subsoil compactness and cipoal) 191 IV.l.2. 3 Soils and Occurrenceof Individual Tree Species (soil texture; plinthite concretions; the eco-site of mahogany; palms) 198 IV.2 The Uplands with Savannah or Savannah-Forest Cover 209 IV.2.1 Primarily Non-Edaphic Upland Savannahs 212 IV.2.1.1 The UplandSavannahs ofEastern Amapd Territory 212 IV.2.1.2 The Upland Savannahs of South-Eastern Marajo Island .... 213 IV.2.1.3 The Upland Savannahs at the Northbank of the Lower Amazon River 214 IV.2.2 Upland Savannahs and Savannah-Forests of Edaphic Origin ... 215 IV.2.2.1 Savannahs and Savannah-Forests of the Planicie (introduction; field observations; conclusions) 215 IV'.2.2.2 Savannahs and Savannah-Forests outside the Planicie 222 IV.3 The Lowlands with Forest Cover 223 IV.4 The Lowlands with Savannah or Savannah-Forest Cover 225 IV.4.1 The Lowland Savannahs of the Lower Amazon Region 225 IV.4.2 The Lowland Savannahs of Eastern Maraj6 Island 225

Chapter V Chemical and Physical Qualities of the Main Amazon Soils, and their Agricultural Occupation 227 V.l The Soils of the Lowlands 229 V.2 The Soils of the Uplands Outside the Planicie 229 V.3 The Freely Draining Kaolinitic Soils of the Planicie 230 V.3.1. Chemical and Physical Qualities of the Freely Draining Kaolinitic Planicie Soils 231 V.3.1.1 The Soils underPrimeval Forest Cover(chemical qualities; physical qualities) 231 V.3.1.2 The Soils under Influence of Man (the soils under the shifting culti vation system; the soils under anthropogenic savannah; the soils influenced byPre-Columbian Indianoccupation; thesoils under manuring and fertilizing) 248 V.3.1.3 The Soils with Horizons of FossilPlinthite 259 V.3.2 Agricultural Land Capability Evaluation for the Freely Draining Kao linitic Planicie Soils, and their Adequate Management 260 V.3.2.1 Agricultural Land Capability Evaluation 260 V.3.2.2 Soil Management (maintenance of soil organic matter; exploiting of soil moisture reserve; tillage and erosion control;fertilizing) 260 V.3.2.3 Systems of Crop Production 264

References 267

Summary 274

Sumario 283

Samenvatting 293

Appendices 1 Mapa dos solos — Area Guamd-Imperatriz, regiao ao longoda parte superior da rodovia BR-14 Soil map — Guama-Imperatriz area, region along the upper part of the BR-14 highway 2 Mapa dos solos — Area Araguaiana de Mogno Soil map — Araguaia Mahogany area 3 Mapa semi-detalhado dos solos e ocorrencia dePau Amarelo— Unidadede terra 'Candiru1 Semi detailed map of soils and occurrence of Euxylophora paraensis — Land Unit 'Candirii' 4 Perfil da estrada — Trecho Sao Miguel do Guamd-Imperatriz da rodovia BR-14 Cross-section of the road — Stretch Sao Miguel do Guama-Imperatriz of the BR-14 highway 5 Esboco da geologia e geomorfologia — Trecho Sao Miguel do Guamd-Impera triz darodovia BR-14 Sketch of the geology and geomorphology — Stretch Sao Miguel do Guama- Imperatriz of the BR-14 highway 6 Esboco dageologia, pedologia e vegetacdo — Area Araguaiana de Mogno Sketch of the geology, the soils and the vegetation — Araguaia Mahogany area 7 Campos naturais ni regiao do Rio Para e baixo Tocantins Natural savannahs in the region of the Rio Para and the lower Tocantins 8 Stratigraphic sections of (a) the Amazon basin, (b) the Acre basin, (c) the Marajo basin and (d) the Maranhao basin Secedes estratigrdficas da (a) bacia amazonica, (b) bacia de Acre, (c) bacia de Marajo e (d) bacia de Maranhao 9 Analytical data ofth eProfile s2-18 :Groun d WaterLaterit e and related soils, and the Profiles 19-23: soilswit h fossil plinthite Dadosanaliticos de Perfis2-18: solosLaterita Hidromorfica e aparentados; ede Perfis 19-23:solos com 'plinthite'fossil 10 Analytical data of the Profiles 24-53: representative profiles of the applied soil classification Dadosanaliticos dePerfis 24-53: perfis representantes da classificacaodesolos aplicada

The manuscript was submitted in October 1964. The Appendices 1-7 were already printed in 1963. but unfortunately with the provisional title: 'A study of soils of the Amazon valley'. In view of the Brazilian application of the maps, the Portuguese version of theexplanator y texts is putfirst. O manuscrilodesta publicacaofoi apresentado em outubro 1964. Os Apendices 1-7 foram impressos id em 196J. infelizmente com titulo de livropro visor io: 'A study of soilsof the Amazon valley'. Forcausa da aplica- cao brasileirodesles mapas.o lexlo explicativo emporlugues foi colocadoem primeiro lugar. "Theluxurious vegetativecover led to theearly assumption that thesupporting soil was extremelyrich. However, once it was realised thatsuch growth isbased largely ona closednutrient cycle ontop of the soil,and in viewof thefailure ofagricul turalsettlements in the region,this opinion wascompletely reversed. Indeed, the greatmajority of the Amazon soils are 'poor' inthe chemicalsense." Introduction

The name 'Amazonia' immediately conjures up visions of damp, vine-entangled greenery,infeste dwit h preying tarantulas and snarlingjaguars . Such notions of 'the green hell' or'the steamingjungle 'ar e widespread among the average European and American, but alsoamon gman y aninhabitan t ofsouther n Brazil.The yhav ebee n fed on vast numbers of fictional impressions on the region, through vivid pictures on school walls,lus hdescription s innovels ,popula r magazines,an d impressive illustrated books.Whe n their authors did not entirely draw on their ownimagination , or simply copy from previous narratives, they often only peeped through riverside vegetation or a patch of secondary forest and took their pictures at the Belem zoo. They did a disservicet othos eintereste d inthi spar to fth eworld .Al lth emor eadmiratio ni sdu et o such early explorers of Amazonia as COUDREAU, BATES, KATZER and LECOINTE for theirperseveranc ean d theveracit yo fthei rpublications . One'sfirst sensatio n on actuallypenetratin gth eprimeva lfores t ofcentra lAmazoni a is one of deception. There are no entangling vines, hardly any wildlife, no steaming atmosphere. One soon discoverstha t the dangers of snakes,jaguar s and wild Indians (the cobras, oncase indios of the indigenous caboclo) are verylimited . The everlasting drone of insectsi sth e only realnuisance . It takessom etim e before onebecome s fully aware ofth ecathedral-lik e majesty of the highfores t withit schoi r of rasping cigarras and howlingguaribas, and beginst o marvela tth edelicat e igapo growthalon gth ecoo l rivulets. The present soilstud y isth e result of three yearso f plodding along straight transects through primeval forest, accompanying an experienced forest inventory team. Very divergent evaluations have been made of the Amazon soils.Th e luxurious vegetative coverle dt oth eearl yassumptio ntha tth esupportin gsoi lwa sextremel yrich .However , oncei twa srealize dtha t suchgrowt hi sbase d largely on a closed nutrient cycle on top of the soil, and in view of the failure of agricultural settlements in the region, this opinion was completely reversed. Indeed, the great majority of the Amazon soils are 'poor' in the chemical sense. But until recently the only factual information on whicht obas esuc ha conclusio nwa sth e 1926repor t of MARBUTan d MANIFOLD,whos e investigations were an aftermath of the famous rubber boom. After its collapse, no new data on the Amazon soils were gathered until well after the second world war, whenth eSPVEA ,th eFedera l BrazilianDevelopmen t Board forAmazonia ,wa sse tup . Drawing on the pioneering work of DAY, this book gives a new description and evaluation ofth e main Amazon soilsbase d on present-day standards of soil study. It is endeavoured to combine the advantages of the morphometric method, as now applied in the U.S.A., with those of the more physiographical approach of certain European soil scientists. Thus, on the one hand, detailed morphometric descriptions aregive n ofrepresentativ e profiles and classification ofth emai n soilsi sfull y checked. On the other hand, much attention is given to the pedogenetic factors. Several of these factors as exhibited in Amazonia, have been little described, unlike those in such comparable regions as the central Congo basin. It was therefore necessary to relate extensively our ownan d others'dat a on such aspects asgeomorphology , which is found to be of great importance for the understanding of the local pattern ofsoils . The purpose of the study is twofold. On the one hand it isa release of scientific data of more than regional interest - on the Amazon soils themselves, the genesis of tropical soils in general and on soil- plant relationships ina tropical region with limited human influence. But it is hoped that this publication will also serve as a kind of handbook for those Brazilians, and their associates, who are directly concerned with the Amazon soils and their use.A n attempt istherefor e madet o facilitate future soil surveys in the region by providinga legendt o allth esoil shithert o encountered andb ydescribin gthei rgeographi c pattern. All existing data on the qualities of the main soils and their management is also provided. Theavailabl efield an d laboratory data onlycove ra par t ofth eregio n and they often giveles sinformatio n than mightb ewished .Moreover ,conclusion s onloca lpedogene sis and soil-plant relationships are often only tentative. The resulting picture of the Amazon soilsi stherefor e far from complete, but it may wellope n the way to further soil studies on the immense region known as Amazonia. Finallyi tshoul db estate dtha tth eopinion san dconclusion sexpresse d inthi spublica tion are the author's responsibility, and may not be taken to represent the official opinions or policies ofeithe r the Superintendencia do Piano de Valorizagdo Economica daAmazonia or the Food and Agriculture Organisation of the United Nations. Some terms and abbreviations

Amazonia: The Amazon region. Amazon planalto (LIT. Amazon high plain): The Plio-PIeistocene plateau land of Amazonia,usuall y at 150-200m altitude . Planicie (LIT.flat are a of very largeextent) : Allth e upland in the axial part of Ama zonia, comprising the Amazon planalto and the Pleistocene terraces. Terrafirme (LIT. stable terrain): All the upland, i.e. all non-flooded terrains, com prising the Planicie and the older geomorphologic units. Vdrzea (LIT. low, grassy land): Holocene lowland, intermittently waterlogged. Igapd (LIT. marsh): Holocene lowland, or bottom lands, permanently waterlogged. Massapi (LIT. sticking-to-the-feet): Strongly mottled, clayey soil; also used to denote terrains slightly above level offlooding, presumabl y of Early Holocene age. Belterra clay:Ver y heavy, kaolinitic clay, deposited on top of the Amazon planalto. ReworkedBelterra clay: Belterra clay nowadays on terrains below the level of the Amazonplanalto . TerraPreta (LIT. black earth):Soi lo f Pre-Columbian Indian dwellingsites . Hileia (or hylaea, LIT. the great forest): The vegetative cover of Amazonia as phytogeographic unit. Cipoal (LIT. cipo-o r lianavegetation) : Forest type composed largely of creepers and climbers. Tabocal (LIT.taboca- o r bamboo vegetation):Fores t typecompose d largely of Guadua species. Campo (LITfield):. Typ eo fsavannah . Campina (LIT. small campo): Type of savannah. Campina-rana (LIT. false campina): Type of savannah-forest. Caatinga amazonica (LIT.Amazo n caatinga, i.e.ope nforest) :Typ eo f savannah-forest. Capoeira (LIT. the forest that was):Youn g secondary forest.

FAO: Food and Agriculture Organization ofth e United Nations SPVEA: Superintendencia do Piano de Valorizacao Economicada Amazonia, the Federal Brazilian Development Board for the Amazon Region IAN: Instituto Agronomicodo Norte, the Federal Brazilian Agricultural Institute for Amazonia (as from October 1962, IPEAN: Instituto dePesquisas e Experi- mentacoes Agropecudrias doNorte) IQA: Instituto de QuimicaAgricola, th e Soils Laboratory of the Federal Brazilian SoilsCommissio n

I. The Environmental Factors

1.1 The Geographic Location

The extent ofth e Amazon region, later referred to asAmazonia , varies according to the criterion used. Most of the catchment area of the Amazon river system issituate d within Brazilian territory, but it also comprises considerable parts of Bolivia, Peru and Colombia. Compared to this, the phytogeographical area of Amazonia is smaller to the south and moreextensiv eo nth e northern hemisphere, where it includes the Guianas and a part of Venezuela (Fig. 1).

Fig. I Algunsdados geogrdficos da Amazonia

80 W 70 60

Fig. 1. Somegeographical data of Amazonia

11 As a geologic basin Amazonia is much smaller. It consists of a relatively narrow stretch atth elowe r part of the river system, and afan-lik e area towards theAndes . This study deals with the Brazilian part of Amazonia. The limits of Brazilian Amazonia are partly arbitrary. In this publication, the transitions from the belt of equatorial forest toth esavannah s of North-Eastern and CentralBrazi lan dt othos eo f Rio Branco - British Guiana are taken as boundaries. Brazilian Amazonia thus con sists of all of the StatesPari ,Amazona san d Acre,th efedera l Territories Amapd and Rondonia, and parts of the States Maranhao, Goias and Mato Grosso. References to the Amazon river in general include the Solimoes, which is the name given to the upper part of the central river in Brazilian territory. The axial part of Brazilian Amazonia may be divided into four areas, as follows:

1. the Upper Amazon region, from the frontier with Peru-Columbia to Manaus. 2. the Middle Amazon region, from Manaus to the boundary between the States Amazonas and Para (Faro). 3. the Lower Amazon region,fro m Faro to the mouth of the Xingii river. 4. the Estuary region, including Marajo island, from the mouth of theXingi irive rt o the Atlantic Ocean.

1.2Th e Climate

The climate of Brazilian Amazonia is humid and hot. The number of weather recording stations is small and they are far from evenly distributed. In the huge watershed areas at the north and the south side of the Middle and Lower Amazon region they are veryrare .Nevertheless , the climate init sentiret y isfairl y well-known. Much relevant data and many mapsar e produced ina recently published book onth e geography of Brazilian Amazonia (GUERRA, 1959). Therefore only the main charac teristics are described.

The rainfall in Brazilian Amazonia is generally high. There is nevertheless much variation in total annual rainfall, as well as in the annual distribution. The data for total annual rainfall aregive ni nFig .2 .Highes t recordings (over300 0mm/year )ar ei n the extreme east and the extreme west, namely in the north-eastern part of Amapa Territory and in the north-western part of Amazonas State. The Figs. 3an d 4 show the number of dry months - less than 50 mm and less than 100 mm rainfall per month -, and the central month of the dry season.

With regard to the temperature, Brazilian Amazonia showshigh , but not excessively high values. The mean annual temperature varies from 23.5°C to 26.9°C. The rainy season (inverno) is generally slightly less warm than the dry season (verdo). The annual amplitude inth emea n temperatures ofth e months ishoweve r verysmal l at all stations of the region, namely below 5°C, and in most places even below 2.5°C . The

12 Fig.2 Os tiposclimdticos da Amazonia brasileira conforme a classificacao de Koppen, e a precipitacao anual total (de GUERRA, 1959; cf. 1.2)

o--

10" s Fig.2 The climates of Bra zilian Amazonia as per the classification of Koppen, and thetotal annual rainfall (from GUERRA, 1959; cf. 1.2)

Fig. 3 Numero de meses si- cospor ano, e mis centralda estacao sica, sendo criterio de urn mis sico uma preci pitacao de menos de 50 mm (coligidode dadosde GUER RA, 1959)

10- Fig. 3 Numberof drymonths 5 per year, and central month ofdry season, takingless than 50 mm rainfall as criterion for a dry month (compiled from data of GUERRA, 1959)

Fig.4 Numero de meses se- cos por ano, e mis centralda estacao sica, sendo criterio de urn mis sico uma precipi o"-- tacao de menos de 100 mm (coligidode dadosde GUER RA, 1959)

Fig.4 Numberof dry months per year, and central month of dry season, taking less than 100 mm rainfallas crite rionfor a dry month (compi led from data of GUERRA, 1959)

13 temperature of the coldest month iseverywher e above 18°C.Th e absolute maximum temperature ever measured is only 44°C (Tefe). There is a distinct difference in temperature between day and night; high amplitudes are known for instance for Betem (9.6°C), Manaus (8.7°C), and S

The relative humidityi s high. The variation over the region in the annual mean is between7 3% an d94% .I nth ecentra lpar t of BrazilianAmazoni a thevalue sar eabov e 80% everywhere. The highest percentages are found near the Ocean coast, and in Acre and western Amazonas State.

Thewinds ar eusuall yweak ,excep tfo r thecoasta lregio nwher ethe y mayb emoderat e (ventogeral). Storms such as typhoons are unknown, but gusts of strong winds are frequent before afternoon rain showers. Exact data are lacking. Figures for evaporation, cloudiness and dew are not available, as no systematic recording has been carried out.

With KOPPEN'Sclassification , Brazilian Amazonia hasthre etype so fclimat e (Fig. 2). Type Af (hot and humid, without dry season) is found in the north-eastern part of Amazonas State,an d ina pocke t ofa s yet undetermined sizearoun d Belem. Type Am (humid and hot, with a short dry season)comprise sth e main part of theregion ,whil e type Aw(humi d and hot, with a pronounced dry season) occurs on the south-eastern fringes of forest covered Amazonia, in the boundary area with British Guiana, and in a part of the Lower Amazon region.

1.3Th e Geology

The geology of Amazonia was studied already at the end of the nineteenth century, for instance by DERBY (1877) and KATZER (1903). In recent years much more has become known, especially from the extensive surface and subsurface prospections of the national oil company Petrobras. On the Figs. 5 and 6, this Company's data are givenschematically ,whil eus ei sals o madeo fth emos t recentgeologica l mapo f Brazil 1 (LAMEGO, I960). The Amazon valley constitutes a low, sedimentary area between the shields of Central Brazil and the Guianas. These shields consist of crystalline basement com plexeso fol dage ,namel yo fth ePre-Cambria nperiod .Th erock sar eprincipall ygranites , gneisses and mica schists. For the Guiana shield north of the Lower Amazon region, KATZER (1903) indicates that the granites are concentrated at the frontier zone with the Guianas, the gneisses occur in a band along this zone and the mica schists- with

') The area of the Brazilian shield within Amazonia is very little explored. Omitted from Fig. 5 are several parts indicated as Cretaceous on early geology maps, and mainly as Carboniferous on the 1960 map (areas along the middle Tapajos, the middle Xingii and the lower Tocantins river).

14 intrusions of granites and syenites- occupy the southernmost part of this crystalline shield.I nnorther n AmazonasState , as well as along the Tapajos river, granites seem to be dominant. Thecomparativel y young Andean mountain range,whic h isborderin g Amazonia on the west side, consists of folded sedimentary rock, predominantly of Paleozoic or Mesozoic age. Much volcanic activity took place here. The sedimentary part of Amazonia is not one basin, but consists of several: the basin of Acre, the large basin of the Amazon proper, the basin of Maraj6, and the basin of Maranhao, which latter has two sub-basins, namely those of Sao Luis and Barreirinhas (cf Appendix 5). The first three basins are relatively deep, while the Maranhao basin isshallow .Th e Amazon basin isnarro w in itseaster n part, but very wide upstream, in the western part. The schematised stratigraphic sections of the main basins are given in Appendix 8. It can be seen that both in the Amazon, the Maranhao and the Acre basins a part of the beds was deposited under marine or lacustrine conditions. In the Maranhao basinth e depositsar elargel y Paleozoic or Mesozoic.The y outcrop over extensive areas. The youngest deposits of considerable thickness are of Late Cretaceous or Tertiary age (Itapecurii or Serra Negra beds). They occupy the nort hern part of the present surface of the basin. In theAmazo n basin proper, the depositsar efo r alarg e part Paleozoico r Cenozoic. The Paleozoic deposits only outcrop on the edges of the basin in its narrow eastern part. At one place they form a Dome (Monte Alegre). Apart from the edges, the basin surface consists of Late Tertiary deposits (Alter do Chao or Barreiras beds). The Acre basin consists of Cretaceous and Tertiary deposits for a large part. Their composition isdifferen t from that of the deposits of the same periods in the Amazon basin proper. The Marajo basin has only deposits of Cretaceous and younger age.

Seen in its entirety, the sedimentary part of Amazonia has at its surface a large proportion of Cretaceous or Tertiary sediments. They are of varying textures and colours,consis t of kaolinitic claysan d quartz sands, and are little consolidated. Thethicknes san d theexten t ofth e Pleistocene depositsproper , whichconsis t alsoo f kaolinitic claysan d quartz sands,ar eapparentl y limited,especiall y inth ewester n part of Amazonia. At Nova Olinda, near the confluence of the Amazon and the Madeira rivers, the Pleistocene sediments comprise only several metres. On the banks of the Amazon river between Manaus and Monte Alegre, the Tertiary Alter do Chao or Barreiras beds are well exposed. The contact of these beds with the overlying Plei stocene sands is found at varying heights above the river. The thickness of the Plei stocene deposits reaches a maximum of 50 m, namely at Monte Alegre (SAKAMOTO, 1960). The approximately flat terrains east and south of Belem (zona Bragantina, zona Guajarina) are covered with Pleistocene sediments, but only in a thin layer (5-10 m).Th e only thick Pleistocene deposits are inth e Marajo area. Herethe y reach about 250 m thickness and consist of grey silts (Appendix 8).

15 Fig.5 Mapageologico esquemdlico daAmazonia brasileira. Domapa preliminar da PETROBRAS(1961) e o ultimomapa geologico doBrasil (LAMEGO, 1960). O Holocenofoi esbocado demapas bdsicospreli- minares daAA F (1942),mapas deInventdrio Florestale observacoespessoais

E. AMAZONAS

-1000 -hVtW, -2ooo-r\>%:< -3000 - -4000 - -5000- •'•I* -6000•* ' + t + + + + + M I 1 1 1 1 1 1 0 300 KM

Fig.6 Corte transversaldas principals bacias sedimentares da Amazonia. Do mapa preliminarda PETROBRAS(1961) .Para alocacao doscortes transversals, vejaa Fig. 5

16 1. Pre-Cambrian/Pri-cambriano 40 2. Prc-Silurian/Pre'-siluriano

3. Silurian/Siluriano

4. Devonian/Devow'ano

5. Carboniferous/Carfrom/ero

6. Jura-Triassic//ura-7Wdsi'co

7. Jurassic (diabase)//«rdiico (diabase)

8. Mesozoic undifferentiated, probably largely Cretaceous/ Mesozoiconao diferenciado, provdvelmente mormenteCreldceo

9. Tertiary and Pleistocene/ Terciario e Pleistoceno

10. Cretaceous, Tertiary and Pleistocene/Cre/a'ceo, Terciarioe Pleistoceno

11. Holocene and inland water/ Holoceno e dgua no interior Fig.5 Schematicgeological map of Brazilian Amazonia.From preliminary map O/PETROBRAS (1961), andthe latestgeological mapof Brazil (LAMEGO, 1960).The Holocene sketchedfrom AAFpreliminary basemaps (1942), ForestInventory maps, andpersonal observations

MtCOGUKUM M A R A J 0

1000 VW- -2000

>V+*+Vt •- -400 0 * * • •• ^*vv^^»>w.;^v.v* + * * • • • , t.Moo

i& mni2nrj s en*

1. Pre-Cambrian//W-ca/M&r('an0 2.Pre-Siluria n and Silurian;Pre-silurianoe Siluriano 3. Devonian/ Devoniano 4. Carboniferous/Carfom/ero 5. Jura-Triassic//Mra-7>j'dj/co 6. Jurassic (diabase)/ Jurdsico(diabase) 7. Cretaceous/Cre/dceo 8. Tertiary and Quatemary/7Vrc/a'/7oe Quartendrio

Fig.6 Schematiccross-section of theprincipal sedimentary basins of Amazonia.From preliminary map O/PETROBRAS (1961). Forlocation of cross-sections see Fig. 5

17 For geologists, truly Pleistocene deposits are therefore restricted in occurrence. But it will be shown that Pleistocene eustatic changes in sea level and Pleistocene river regimes haveha d aver ygrea t influence in reworking of the Tertiary deposits and the topography of the Amazon land (cf. 1.4). It should be borne in mind that considerable parts ofAmazoni a which are indicated ascrystallin e or Paleozoico nth egeologi cmap ,are ,i nfact , covered withthi nlayer so f Tertiary or Pleistocene sediments.Thi s applies,fo r instance,t o thelowe r Gurupi area and the whole region of the Rio Negro.

The Holocene sediments consist largely of silts and clays.The y cover the floodplain areas along the Amazon river proper and the lower parts of its main tributaries, and are also found in the Estuary region (Fig. 5). However, the Holocene areas, all together, comprise only about 1-2% of the total land surface of Amazonia. This is much less than assumed by early explorers. As they only travelled on the rivers they rarely sawan y upland, and therefore obtained a wrong impression on the expanse of the floodplains.

1.4 The Geomorphology

1.4.1Sketc h of proto-Amazonia

Due to the absence of major orogenetic movements sinceth e Pre-Cambrian period, other geomorphogenic factors wereallowe d to play a marked role inth e development of the landscapes of Amazonia. The evolution of Amazonia as discussed below, is illustrated in the sketches of the Figs. 8an d 9. A very extensive denudation and peneplanation1 must have taken place after the Pre-Cambrian. The Andean mountain range did not then exist. Instead, a Peruvian trough was present. Because of the existence of marine Paleozoic deposits in the above mentioned Amazon sedimentary basin, DERBY (1877) assumed that there was no superficial connection between thetw o crystalline shields in those early periods. The shields would haveconstitute d two separate islandsa t the northern and southern sideo f an Amazon trench2, atrenc h whichwoul d havebee n connected withth e Peru vian trough and the proto-Atlantic until the Tertiary. KATZER (1903), however,con cluded that such an Amazon trench had no connection with water masses east of present Brazil. He stated that there must have been a superficial connection between the eastern parts of the shields, running over the present-day Estuary region. The existence of such a connection has been recently confirmed by the presence of the

') It may concern actually 'pediplanation', as this term is applied more recently, for instance by KING; theexistin g geomorphologist's controversy on this subject is evaded by adhering to the old term 'peneplain' for an erosional surface, whatever its mode of formation. *)'troug h of subsidence' according to SOARES (1956).

18 Gurupa arch (Fig. 6), and the fact that the Maraj6 sedimentary basin constitutes a Graben which hasbee n acting only from the Cretaceous onwards. Aconnectio n with the Peruvian trough willhav eexisted , although for several Paleozoicperiod s sills and land barriers are supposed to have occurred (KATZER, 1903; SAKAMOTO, 1960). During the Paleozoic the erosion products of the shields were transported from the north and the south toth eAmazo n trench, and ina western direction to the Peruvian trough. Transport to the Maranhao basin also took place. Permean, Triassic and Jurassic deposits are not found in the Amazon basin, except for somebasi cintrusions . KATZER(1903 )suppose stha t duringthes eperiod sther ewa s asuperficia l connection between the Guianas and Central Brazil over a broad front. Asrelated ,fo r instance,b y OLIVEIRAet al. (1956), agenera lsubsidence ,accompanie d by an intensive levelling,too k place at the end of the Jurassic or the beginning of the Cretaceous.Cretaceou sflat lyin g sand-stones, of continental origin, are found wide spread inth e interior of Brazil. From their reddish colour, and the presence of eolian elements, SAKOMOTO(1960 )conclude d thatthes esediment swer eforme d ina relatively warman d dryclimate ,an d spread overth eplain sclos et othei rsource .Th esand-stone s now apparently still cap many of the plateaux of the watersheds between the main rivers. Examples are the Chapada de Parecis,Serr a do Roncador, Serra do Cachimbo on the Brazilian shield, and parts of the Serra deTumucumaque ,th e Serra de Acarai and the Serra de Pacaraima on the Guiana shield. These plateaux are often at an altitude of 400-600 m. Apart from capping these plateaux, Cretaceous sedimentsar e found in parts of the basins, where they may be marine (Acre, Maranhao). SAKAMOTO(1960 )an d VARGAS(1958 )distinguis hals o anEarl y Tertiary peneplanation surface on the crystalline shields, at 250-400 m altitude. To this level for instancs would belongth e Campos de Ariramba (90k m north of 6bidos), the Campos Geraie de Obidos (200 km north of 6bidos), and the table lands of Almeirim-Prainha' amongst which isth e Serra de Paranaquara.1 The savannah areas inth eregio n of the upper Paru river, near Surinam, probably belong also to this Early Tertiary pene planation surface (cf. quotation of KATZER on p. 164), so alsod oth esavanna h areaso f the middle course of the Tapajos (Campos de Cururii, Campos de Mucajazal) that arereporte d by LECOINTE (1922).Th e geologic material of the floor of this surface is not necessarily Pre-Cambrian crystalline. SIOLI and KLINGE (1961) mention that the substratum of the Campos deAriramb a consists of sand-stones of possibly Devonian age, and that of the Campos de Cururii of sand-stones of Cretaceous age. Little can be said with certainty about the above mentioned peneplanation surfaces. The correlation with the surfaces in Central and Southern Brazil, as established by

') From the geologic description of this table land by KATZER (1903) it seems likely that itsto p issimila r to that of theplanalt o south of Santarem(Belterr acla y overIpixuna-lik e plinthite concretions; cf.1.4. 2an d1.4.5) .I t seems therefore moreprobabl e that thetop s of the Serra de Paranaquara and the adjoining table lands belong to the Plio-Pleistocene Amazon planalto level, and that their relatively high altitude is due to a Post-Tertiary relative uplift of the area between Almeirim and Prainha. This is in agreement with SIOLI (1957) who deduced from the absence of rias (cf. 1.4.3) in this area, that such an uplift occurred.

19 KING (1957),als oawait sfurthe r study. Iti spossibl etha tth eabov ementione d Amazon ones are related to KING'S Post-Gondwana and Sul-Americana levels respectively. It isgenerall y accepted(cf. OLIVEIRAet al., 1956 )that , from the Middle Tertiary on, a general uplift of Amazonia took place, except for the Maraj6 area (Graben of Maraj6, seeabove) . The extraordinarily extensive Tertiary sediments of the Amazon basin,know na sth eAlte r do Chaoo r Barreiras beds,wer eprobabl y deposited largely during the Miocene, under continental conditions. Until recently it wasassume d that this had been done by the then westward flowing proto-Amazon river. BARBOSA (1959), however, casts doubt on the truly fluviatile character of the deposition. He states that the sedimentation took place with an endoreic drainage, related to what seems to have been a semi-arid climate of great extent, and that the Amazon river system proper originated only afterwards, during a subsequent humid climate. Contrary to the general situation, the Tertiary deposits in the Acre basin, as well as those in the area east of Belem, were for a part laid down under marine conditions (Pirabas beds, Pebas beds.) To sum up, the watershed regions on the north and the south sides of the Amazon river system may be divided into the following geomorphologic units: 1. Undulating terrains with outcropping crystalline basement. 2. Undulating terrains with outcropping Paleozoic, Mesozoic or Early Tertiary deposits. 3. Two peneplanation surfaces within the area designated as crystalline on the geologic maps, supposedly of Cretaceous and Early Tertiary age respectively.

1.4.2 The Planicie. The Plio-Pleistocene Plateau or Amazon Planalto

Throughout theeaster n part ofAmazoni a canb efoun d aflat, terrace-lik e surface, at 100o r moremetre sabov eth eleve lo fth emai n rivers.Th euppe r layerso fthi s surface, about twenty metres thick, consist of uniform, yellowish, very heavy, kaolinitic, sedimentary claywithou t any visible stratification. After exposure to the weather, the clay falls apart into characteristic fine structure elements that are very resistant to destruction (see Photo 1).Thi s particular material is scarcely mentioned in literature about Amazonia.Th e material willhencefort h becalle d Belterraclay, after one of the localities whereit soccurrenc e isver yconspicuous . Before dealingwit hth eproble mo f itsorigin ,a nextensiv edescriptio n willb egive no fth eoccurrenc e of plateau landwit h Belterra clay cover.

1.4.2.1 Description of Occurrence ofBelterra Clay{cf. Fig.7 ) A Belterra clay covered flat surface is very widespread in the upper Capim - upper Gurupiarea ,i nmos tplace sa ta naltitud eo fabou t 200m ,bu t inth esouthernmos t part, theSerr ad eGurupi ,a ta naltitud eo fabou t35 0m (cf. cross-section so fth eBR-14high - way,Appendice s4 an d5) .I nth eregio nbetwee nth eriver sXing uandTocantins,th esur facei spresen tapparentl yonl yfa r southo fth eAmazo n river;i toccur swes to fTucuruf ,

20 Foto I Argilade Belterra exposlaa superficie. Aposser exposta as condicoes climdticas,a argila caolinitica tnuilo pesada, denominada deBelterra, desagrega-se em elementos macicosbem finos e muito resistentesa destrucao

Photo 1 Belterraclay exposed at the surface.After exposureto the weather, the veryheavy, kaolinitic clay, termed Belterra clay,falls apart into veryfine blocky elementsthat are veryresistant to destruc tion ata n estimated altitude of 160 m (Serra deTrocara) .Betwee nth eriver sTapajo s and Xinguhoweve rth esurfac ewit hBelterr acla ycove ri sfoun d neart oth eAmazo nriver . Directly south of Santarem the altitude isabou t 130m (GOUROU, 1949). At Belterra Estate,abou t 40k msout ho fSantarem ,th ealtitud ei sabou t 175 m.A tth e Curua-una centre, about 80 km east-south-east of Santarem, the altitude is 180 m.1 Near the Xingu the altitude isles san d the surface seemsi nmos t placesno tflat. Fo r theregio n west of the lower river Tapajos HEINSDUK (1958c),wh o knew its characteristics well from a survey inth e Curua-una area, reportstha t the Belterra claycovere d surface is also present, although much attacked. The plateau fragments in the area (Serra de Parintins, Serra de Balaio,an d others) are reported to occur at about 150m altitud e (LECOINTE, 1922). On the north side of the Amazon river the Belterra clay is also found. The clay occursi nth esouther npar t of Amapa Territory,namel y in the headwater region of Igarape" do Lago (about 60 km northwest of Macapa), at an estimated altitude of 100m .Th e clay has been reported to occur at 'some distance' north of Prainha. The presence of Belterra clay was also established at a short distance north of the towns

*) Inthi sarea ,th esurfac eha s asligh tdi pi na neasterl y direction.

21 Fig.7 Ocorrencia de argilade Belterra

:io'

x Locations//ocofoei 30 Locations with established altitude (above sea level)/locacoes com allura determinada (sobre niveldo mar) + Height above high-water of local river or rivulet/a/fwa sobre nivel do rio ou igarapi local

Fig.7 Occurrence of Belterra clay

Terra Santa and Faro, between the rivers Nhamunda and Trombetas, there at a re latively low altitude, about 60 m. It seems probable that the terrain further north, which forms a plateau - Serra de Chinelo, Serra de Cunurii- at about 200m altitud e (KATZER, 1903) has also a cover of Belterra clay. A similar plateau is very conspi cuously present east of the river Uatuma, forming the Serra de Santa Rita, Serra de Santa Rosa and others (VARGAS, 1958). The sameautho r mentions aplatea u of 150 m altitude directly north of 6bidos. Large parts of the stretch traversed by the recently constructed Manaus-Itacoatiara road (AM-1) have a cover of Belterra clay. The altitude of the approximately flat sections of plateau in this area, locally called chapadas, isabou t 120m in the central stretch of the road (about 40k m north of the Amazon river). Near both Manaus and Itacoatiara the terrain tops are less flat and their altitude is somewhat lower. There are few data about the occurrence of Belterra clay on plateau land in the western part of Amazonia. MARBUT and MANIFOLD (1925) report the existence of a plateau east of Humaita on the Madeira river. Southeast of Porto Velho large flat stretches occur, at about 40m abov e local river level,wit h acove r of Belterra clay.I t has, however, a slightly lessheav y texture than that in the eastern part of Amazonia. It seems probable that the western part of Acre State has a terrain cover similar to that near Porto Velho.Indication s from AAF and CNG topographical maps are that

22 the terrains in the central part of Amazonas State (Jurua-Purus-Madeira) have very little topographical differences and are relatively low lying, both with regard to altitude and to height above local river level. The textures of the surface materials seem to be mainly heavy, for instance around Tefe. MARBUT and MANIFOLD report that from the Catua river(abou t 250k mwes t of the mouth of the river Purus) to the west, a dissected plateau is present. It seems to extend westward to at least within 50 km of the frontier (1925, p. 642), and also far southward (1926, p. 440). They estimate the level of this plateau to beabou t 30m abov e the river Solimoes. For this regionthe yrepor tsoil stha tar esimilar ,als otexturally ,t othos eo fth eplatea usout ho f Santarem. In the frontier region there are apparently more topographical differences than inth eJurua-Purus-Madeir aregion .Fo rth eare asout ho fBenjami n Constant,o n the frontier, the existence is reported of sections of plateau similar to that of the Curua-una centre(interna lreport sFAO/SPVE A Mission, Belem).Th eonl yavailabl e data about the adjoining parts of Peru are those of MARBUT and MANIFOLD,wh o mention a similarity with the Brazilian part of the region. For the Madre de Dios areai n Boliviawidesprea dpresenc eo fplatea u land isestablishe d (DREWES, 1961).N o reliable data are available on the texture of the coveringsediments . In summarisingthes edata ,i tma yb esai dtha t overal lo f the axialpar t ofAmazoni a flatplatea u land occurswit h a cover of Belterraclay .Th e limited number oftextura l analysisdat asugges ttha tth eheavines s of the clayslightl y increasestoward sth e east. Moreover, the lower part of the Belterra clay layer seems to be less heavy textured than the upper part.1 In the eastern part of Amazonia the blanket of Belterra clay on plateau land is underlain bya layero fplinthit e( = laterite)concretions .Th eplinthit elaye r seemst o constitute the weathered upper stratum of the sediments of the Barreiras or Alter doCha obed swhic hcompose sth ebas eo fth esection so fplatea u (cf. 1.4.5).

1.4.2.2. Originof Belterra Clay (cf. Figs.8 an d9 ) Neither datingno rexplanatio no fth eoccurrenc eo fth e Belterracla yo nplatea ulan d hasbee nfoun d inliterature .Th edepositio no fth ecla ymus thav etake nplac eafte r the first uplift of the Andean Cordillera and the flattening of the Amazon land by de-

*)Th e percentage of clay (separate <2[JL) varies between 85 and 95 %. Within the silt fraction (2-50|i),ther ei sapparentl y noclea r top; two samples gaveth e following granulometric data (by courtesy of DR. FAVEJEE of the Geology Department of Wageningen University): >50[x 50-32j x 32-16[ x 16-8 JX 8-2n <2[i (BR-14, km 366; 5^H)c m depth) 4.2 2.2 1.6 2.7 2.7 87.0 (Curua-una, km 14;95-15 0c m depth) 2.6 2.2 1.6 2.8 4.9 85.9 The percentage of silt and/or fine sandi sslowl y increasing with depth, at the cost of the percentage of clay. This is illustrated on the lowest horizons of Profile 24 (III.2), as well on sampleXI-1- 6 of CATE (1960) from 280c mt o 470c m depth on theplanalt o of Curud- unaCentre .Thes esample smoreove rsho wtha t thelowe r part of the Belterra clay 'blanket' is indeed kaolinitic in character, as was already believed because of the similarity in morphometric characteristics between the upper and the lower parts of the layer of sediment. It therefore does not concern a kaolinitisation by soil formation.

23 Fig. 8 Esbdfo que demostra aformacao do planaho amazonico

PLIO-PLEISTOCENE (CALABRIAN) PLIO-PLEISTOCENO

Gurupa arch/Arco Gurupd Marajo 'graben'/'Craftcn' Je V-++ Afarq/rf Shoreline[Lin/ia da costa inland sea mar interior ^,~^.-.^-

PRESENT ATUAL •*m. D ==! Devonian/Devoniano C = Carbonifcrous/Car6on//ero Tr =;Triassic /Triasico Cr = Cretaceous/Cretdceo

Outcropping crystalline shields I „j Paleozoic-Mesozoic deposits Contrafortes crislalinosatiorantes ^=??j undifferentiated Depositos Paleozoicos-Mesozdicos. naodiferenciados f ETTTT3 Early Tertiary peneplanation surface i p»....[ on thecrystallin e shields Superficiede peneplanacao doTerciario Plio-Pleistocene Amazon planalto Inferior nos contrafortes crislalinos (broken hatching: subsided) with Belterracla y Planalto amazonico doPlio-Pleistoceno Crystalline covered with thin layer of (sombreamento quebrado: a fundado) ++++ Pleistocene deposits comargila de Belterra Cristalina coberta de camada delgada de depositos do Pleistoceno Pleistocene terraces (6 + 7: Planicie) Terracosdo Pleistoceno (6 + 7; /£'/A Cretaceous peneplanation level and EH Planicie) '/'//A deposits Superficie de peneplanacao e depositos Holocene seaiments do Cretdceo Sedimenlosdo Holoceno

Fig. 8 Sketch showing theformation of the Amazon planalto

24 Fig.9 Cortes transversals esquemdticos da Amazonia nos sentidosNordeste-Sudeste e Norte-Sul res- petivamente,demostrando as vdrias unidades geomorfologicas (para a legenda e a locacdo dos cortesveja a Fig. 8)

4 +++++ 400-600m

250-400m

150-200m

+400-600m

250-400m

150-200m

I Cretaceous peneplanation surface/Superficiede peneplanacdo Cretdcea II Early Tertiary peneplanation surface/'Superficie depeneplanacdo do Tercidrio Inferior III Plio-Pleistocene Amazon planaho/iVa/w//o amazonico Plio-Pleistoceno IV Pleistocene terraces/'Terracos do Pleistoceno (III + IV:Planicie) S = Silurian/Siluriano;D = Devonian/Dfvoniano; C = Carboniferous/Cor6om/ero;C r = Cretaceous/ Cretdceo;T = TertiaryITercidrio Fig.9 Schematic cross-sections of Amazonia in nortiWbt - southeast direction and north-south directionsrespectively, showing the various geomorphologic units (for legend and situationof sections see Fig. 8)

25 position of Tertiary (Miocene) sediments of the Barreiras or Alter do Chao beds,an d also after the upper stratum ofth e latter sedimentsbecam elaterise d in at least part of the region. The composition of the Belterra clay and its horizontal and vertical very gradual changes in texture lead to the assumption that thissedimen twa sdeposite d in a huge, shallow lake or sea bay, and that the flow of suspensed material was in an easterly direction. The tentative conclusion drawn is the following: Kaolinitic, little consolidated depositsuplifte d inth eAnde sregion ,erodin geasily ,gav esediment s that were spread widely throughout the enormous inland sea or bay that a large part of Amazonia constituted during a time when the sea level was high. The coarsest sediments were laid down in thick layers near the foot of the Cordillera, and the minuscule kaolinitic clay particles were suspended in the bay water. Whilst being borne towards the east these particles were gradually deposited, thus forming a very regular thin blanket on the shallow and flat bottom of the inland sea (see the upper sketch of Fig.8) . Thepresen taltitud eo fth eplatea u withth eBelterr acla ycove ri susuall y 150-200m i n theeaster n part ofAmazonia .Thi scorrespond sapproximatel y withth eCalabria n sea level as given in Table 1.Th e deposition of the clay therefore probably took place in the Pliocene (according to ZEUNER'Sclassification ) or in the Early Pleistocene (accor ding to the classification of WOLDSTEDT). In the western part of Amazonia altitudes are generally lower. There a subsidence must have occurred after deposition of the clay. In the transitional areas between the Amazon trench proper and the crystalline shieldsa continuin g gradualuplif t ofth eshield sha saffecte d thealtitud e ofth eplatea u with Belterra clay. The latter is illustrated in the southern part of the BR-14 cross- section (Appendices 4 and 5). The non-eroded plateau with Belterra clay cover is referred to in the following as Amazonplanalto. As may be seen from the description of occurrence of Belterra clay, there are in the eastern part of Amazonia also terrains with Belterra clay at an altitude considerably lowertha n theplanalt o of 150-200m ,namel y at 50t o 100 m(Igarape " doLago ; Faro; the western bank of the lower Xingii; near Manaus). These terrains are usually not quite flat and often occur near the main rivers. It is assumed that the clay in these places was re-deposited, with little or no admixture of lighter textured sediments, at these lower levels during the Early Pleistocene (Sicilian and Milazzian sea levels mainly; cf. Table 1 and the discussions in 1.4.3).A good example of such a terrain is the km 130-km 190sectio n ofth e Guama-Imperatriz area (Appendices 4an d 5).Th e Belterra clay on such terrains willb e referred to in the following asreworked Belterra clay.

1.4.3 The Planicie. The Pleistocene Terraces

1.4.3.1 Originof Terraces In the Early Pleistocene the river Amazon and its tributaries originated. They provided for drainage of Amazonia to the Atlantic and started cutting into the Plio-

26 Table 1 Coastal terraces of Pliocene-Pleistocene-Holocene age (compiled from WOLDSTEDT, 1958 and ZEUNER, 1959)

Chronology/crono/o^/a Sea American West European British Strand lines levels (m) Americano Europeio Ocidental Britanico linhas costeiras niveis domar (m) WOLDSTEDT ZEUNER

F\andrianjFlandriana +2 ca. | Recent Holocene Holocene Holoceno

PLEISTOCENE/P/ra/0£WK> -100? \ Wisconsin Epi Monastirian + 3.5 ca.) Wurmian Last Glaciation Epi-monastiriano Ultimaglaciacao Late Monastirian + 7.5 ca. \ Monastiriano superior Sangamon Rissian-Wiirmian Last Interglacial Main Monastirian + 17.5 ca. J Interglacial Ultima interglacial Monastirianoprincipal (Eemian) -200? Ulinoian Rissian Penultimate Glaciation Penultima glaciacao Tyrrhenian +30-40 Yarmouth Mindelian-Rissian Great or Penultimate Tirreniano Interglacial Interglacial (Holsteinian) Grande oupenultima interglacial .? Kansan Mindelian Ante-Penultimate Glaciation Ante-penultima glaciacao

Milazzian +60 ca. Aftonian Giinzian-Mindelian Ante-Penultimate Milaziano Interglacial Interglacial (Cromerian) Ante-penultima interglacial Nebrascan Gunzian Early Glaciation Primeira glaciacao PuocENE/ZV/oce/w Sicilian + 100 ca. Danubeian-GUnzian Siciliano Interglacial (Waalian)? Pre- Danubeian Villafranchian Calabrian + 180ca . Nebrascan Biberian-Danubeian Vilafranquiano Calabriano Interglacial (Tiglian)? Biberian?

Tabela 1 Terracos costeiros de idade Plioceno-Pleistoceno-Holoceno (compilado de WOLDSTEDT, 1958 e ZEUNER, 1959)

27 Pleistocene Amazon planalto. Below the level of the planalto, the land was shaped into terraces of various levels above the rivers. It is generally accepted that during the Pleistocene glacials of Europe and North America these aleve lwa smuc hlowe rtha ntoday , sincemuc h ofth eworld' s waterwa s tied up as enormous masses of ice in the polar and subpolar regions. During interglacial ages, on the other hand, the sea level was higher than at present. Compar isons and correlations of ancient strand lines all over the world have been made. Mention ismad e of thestudie s of ZEUNER (1959) and WOLDSTEDT(1958 )wh o arrived at severalspecifi cheights ,whic h theynamed .Thes eheights ,an d their correlation with thevariou s ageso fth ePleistocene ,ar egive ni nTabl e 1.Man ydiscussion s havearise n ast o whether or not glacials coincided with pluvials- i.e. times with humid climate - in the equatorial and subequatorial regions, and whether interglacials with interpluvials - i.e.time s with dry climates - in these regions. For Africa and the Middle East much relevant data has been assembled. Many authors have accepted the above mentioned coincidences for these regions, but therear eother swhos e opinions differ. In comparison, little is known for South America. Recently VAN DER HAMMEN and GONZALES (1960) established from pollen analysis at the Sabana de Bogota that for this region of South America acoincidenc e of glacialswit h pluvialsexists . WILHELMY (1952) mentions a pluvial,correlate d with thelas t glacial,whic h would have occurred from Venezuela to Parana. The glacio-eustatic oscillations in sea level during the Pleistocene may be traced on ancient strand linesan d marinecoasta lterraces. 1Oscillation si nth ese aleve lcaus eals o changes in the erosion base of rivers. At the same time differences in climate related with these oscillations normally cause differences in discharge and load of the rivers. Loweringo fth eerosio n baseo f a river (regression time)normall yresult si nentrench ment in its lower course until the river reaches a new equilibrium in its longitudinal profile (graded profile). The surrounding land, the valley bottom, is modelled into a floodplainwit h the same longitudinal profile. Subsequent raising of the erosion base (transgression time) requires normally another graded longitudinal profile of the river. In that case another, differing floodplain isformed . Along the lower course of the river, sedimentation on the old floodplain takes place normally, whilst more up stream entrenchment of the former floodplain may occur, resulting in terrace for mation.I ntha t case,present-da yremnant so fupstrea mterrac edeposit sar epossibl yo f ageswit h low erosion base of the river-or glacials-.while present-day remnantso f downstream terrace deposits are generally of ages with high erosion base - or interglacials. Therefore, terrace levels of different age may cross, at a certain point alongth e rivercours e ('terrace intersection point', cf. schemeo f Fig. 10).B ymean so f careful studies, which bear in mind the above factors, it may be possible to trace the changes in sea levelno t only from coastal terraces, but also fromfluviatile terrace s as far as these latter have been formed under influence of changes in the erosion base.

l) Reference may be made to the study of RUELLAN (1945),wh o distinguished the following levels in the surrounding of Rio de Janeiro: 80-100 m, 50-65 m, 25-35 m, 15-20 m, less than 15-20 m.

28 Fig. 10 Conexaopossivelde terracos fluviais, na situacao em que urn nivel baixo do mar precede um nivel aliodo mar

High Sea Level (Interglacial) niveldo mar alto Present Sea Level niveldo mar atual Low Sea Level (Glacial) nivel domar baixo

la, lb Fluviatile terrace deposits of interglacial time Depdsitosde terracofluvial de epocainterglacial, naspartes interiores doperfil transversaldo vale * 2a Buried fluviatile terrace deposits of glacial time Depdsitos de terracofluvial de epoca glacial, cobertos 2b Fluviatile terrace deposits of glacial time,a t the outer sideso f the cross profile of the valley Depdsitosde terraco fluvial de epoca glacial, no exteriordo perfil transversal do vale 3 Terrace intersection point;terrac e depositso f glacialan d inter-glacial timeso n the sameleve l Ponto de intersecq&ode terracos;depdsitos de terracode ipocas glacial e interglacialno mesmo nivel

Fig.10 Possible relationship offluviatile terraces, with lowsea level preceding a high sea level

These studies may be carried out at the different depths of the present burial of the deposits,o r atth e stillpresen t remnants of terraces atdifferen t heightso nth e upland. Conversely, if Pleistocene sea levels are taken as established data, then the age of present coastal andfluviatile terrac e deposits may beassessed , which iso f importance for discussion of soilgenesis . River surroundings far upstream, above resistant sills in the river course which are well above the highest ancient sea levels, are not affected by changes in the erosion base of the river. Here, the conditioning factors for terrace formation are only the water discharge and the load of the river on the spot. These factors are very much dependent on the climatic conditions. The terraces concerned are *climatic' fluviatile terraces, in contrast to the fluviatile terraces of rivers downstream that are formed under influence of movements ofth ese aleve l('thalassostatic ' terracesi nth eterminol - gy ofZEUNER, 1959). It will be shown below that Pleistocene terraces of one kind or another have a great extent in Amazonia. As already mentioned in 1.3, the sediments which for geologists are truly Pleistocene are apparently of limited expanse in Amazonia. The changes in the regimes of the rivers during that epoch have in many parts resulted apparently only in the re-working of Tertiary and/or older deposits.

29 1.4.3.2 Terrace Levels in theGuamd-lmperatriz Area Duringth e reconnaissance soilsurve y of the Guami-Imperatriz area it was possible to study in detail the geomorphology of an extensive Tertiary-Pleistocene region, namely along the upper part of the newly opened BR-14 highway, which forms thespin e ofth e surveyarea , 470 km long. For aconsiderabl e part ofthi ssection , the data of the constructors (Rodobras) regarding the original topography of the line of the highway were studied. Forth eothe r parts the altitudes giveno n the kilometre posts were copied, and the height differences between these posts estimated. This permitted the drafting of aslightl y schematised cross-section of the traversed terrain (Appendix4) ,whic hwa sfurthe r schematisedi nAppendi x 5.Fo rdiscussio n thestretc h isdivide d into anorther n (km0-195 )an d a southern half (km 195-415). Thenorthern half, with land-units Santana, Candiru, MedioGuamd and Alto Guama, belongs to the catchment area of the rivers Guama and Capim. These riversjoi n at their lowercours e and make a large curve to the West before they reach their present erosion base in the Baia de Marajo. The southern half, with land-unitsCunhantd, Gurupl-mirim, Planalto and Itinga, belongst o thecatchmen t area ofth e Gurupi river, which drains in almost straight line into the Atlantic Ocean.

In the northern half, the following levels can be distinghuished (very concretionary topsexclude d because ofthei r resistance to erosion): A completely flat terrace, covered with Belterra clay, ispresen t from km 195on , at an altitude of about 200 m (planalto terrace,cf. 1.4.2). North of km 195ver y gentle undulating terrain - also covered with Belterra clay- ispresent , whichflattens out ,i n the north, to a terrace of about 100m altitude. It is found, for instance, around km 120, but onlyeas to fth eroa d itselfi ti sconspicuousl yflat an d extensive.Furthermore , a terrace may be discerned at about 65-70m altitud e (km 5 - km 15; km 70-km 90). Remnants of a somewhat lower terrace, of 40-45 m altitude, are also visible, for instance around km20 ,k m3 3 and km45 .Al lthes eterrace shav especifi c altitudesan d appear to be unrelated to the present day drainage pattern. However, parts of the latter terrace may be,i nfact , aterrac e related toth e present day levelo floca lrivulets , allaffluent s of therive rGuama .Considere da ssuch ,th eleve lo f thesepart si s2 0m +- 1 The terrain directly around km 172ma y also represent this terrace levelo f 20m +. . Atvaryin g altitude,bu t positively related to thepresen t day detailed drainage system is a terrace at 8-10 m +. It is conspicuously present along all the rivulets between km 75 and km 160. Between km 15an d km 75 another well-defined terrace can be distinguished. It is situated in narrow bands along the rivulets, at 3-4m +. At the same height as the present rivulets the valley bottom exists (varzea or igap6). In summarising it can be said that in the km 0-km 195 section the following terrace levels are present: 200 m altitude; •* 100 m altitude;65-70 m altitude; 30-35 m altitude (20 m +); 8-10 m+;3-4 m+.

') With + is meant the height above the high-water level ofth eloca l rivero rrivulet .Wit h 'altitude'(alt. ) theheigh tabov e sealevel .

30 In the southern half, the section between km 195 and km 415, the situation is as follows: The completely flat planalto terrace is present at 200-220 m altitude in the northern 100 kms, and gradually increasing in altitude from 220 m to 360 m in the southern22 0kms. 1Th eterrace s belowthi shav ea varyin g altitude.The yar e apparent lyrelate d toth e present day levelo fth eloca lrivulets ,al laffluent s ofth e river Gurupi. Terrain tops occur at about 70-80 m +, in the stretch km 320-km 405. They are probably the remnants of a terrace at this or a slightly higher level.Anothe r terrace occurs at 30-40m +. This one isfairl y well-defined in the southern part (km 320-km 405), but poorly defined inth enorther n part (km 195-km 320).I n thelatte r part also a terrace levelo f about 10m + may betraced , but this is also fairly poorly defined. Terraces at still lower levels are absent. In summarizing, it can be said that in the km 195-km 415 section the following terraces are present: 200-220 (-*• 360) m altitude; 80m+; 30-40m +; 10 m+.

Atentativ e correlation willno wb egive n ofth eestablishe d terracelevel si nth e Gua- ma-Imperatrizare a withpas t sealeve l oscillations. Of the northern section,th e upper four terraces, with their constant altitude, are most probably marine coastal terraces of which the deposits were laid down during ages with high sea level.Th e 65-70 m one mayb efo r apar t deltaiccoasta l because on this terrace there wasfound , in the km 70-km 90section , a horizontal alternation of light and heavy textured sediments.Th e three lower terraces arefluviatile terraces . To establish whether the deposits of the latter are of interglacial age (high erosion base)o ro fglacia lag e(lo werosio n base),i tshoul d beconsidere d that therive rGuam a hasa t km0 a rock y bottom, consisting of hard sand stoneso fprobabl y Paleozoicage , which isexpose d at lowtides .Owin gt oth epresenc eo f thissill ,lo wse alevel swil lno t themselves have influenced the longitudinal profile of the Guama and its tributaries above km 0 (all the rivulets that drain to the east). Only sea levels higher than the presenton ewil lhav epermitte ddepositio no fsediments .Therefore , theterrac edeposit s ofth erivulet sdrainin gt oth eeas t areo f interglacial age.Fo rth elowe rCapi man d the Guama downstream of the confluence of the latter, resistant sills are unknown. Low erosion base deposits might therefore be present along the rivulets that are dis charging via the Capim (all those running to the West, between km 60an d km 150). If so,the n terrace intersection points,i f present, would probably belocate d upstream ofth e road, invie wo f the smallgradien t ofth erivulet sfro m theroa d on.Th eJaboti - maior-at-km-75-Capim-Guama-Baiad eMaraj o section has a fall of only 30 m over 340 km. The Ipixuna-at-km-107-Capim-Guama-Baia de Marajo section has a fall of 35 m over 390 km, and the Candiru-Acu-at-km-145-Capim-Guamd-Baia de Maraj6 section a fall of 65 m over 420 km. It is therefore probable that also the

J) It should bementione d that thealtitude si nth esouther n half of theBR-1 4cross-sectio n are not quite reliable, due to the uncertainty of the altitude of Imperatriz (probably assessed inexcess) , and the incongruency in the altimetric data for the central part of the stretch,a sprovide d byth e Rodobras survey teams;on eworkin gfor m km0 , onefro mk m 467. The relative heights overshor t and intermediate distances, however, may betake n as certain.

31 terrace deposits of the rivulets draining to the West are of ages with high erosion base. Thepreliminar y conclusion istherefor e that allterrac edeposit si nth ek m0 - km19 5 section are of ages with high erosion base, or interglacials. Comparing the terrace levelswit hth eestablishe d interglacialse alevel s(Tabl e 1),a rathe r striking conformity isfound . Thislead st o atentativ edatin go fth eterrac e depositsi nth esectio n showni n Table 2.

Table2 Ageof the terraces in the northernpart ofthe Guamd-Imperatriz area

Terrace level (m) and Origi n Sediment High sea levels (m) Age nivel do terraco (m) e sedimento niveis domar idade origem altos (m)

3-4+ more or less sandy 3.5 Lower interstadium of fluviatile//7«v/a/ matsou menos Epi Monastirian Wiirmianglacia l(Wisconsin ) arenoso Epi-Monastiriano Interstddioinferior da glacial Wurmiana 8-10+ more or less sandy 7.5 Upper part of Rissian- fluviatile//7«v/a/ mais ou menos Late Monastirian Wurmian interglacial arenoso Monastiriano (Sangamon) superior parte superior da interglacial Rissiana-Wurmiana (20+) more or less sandy 18 Lower part of Rissian- fluviatile//7uv/a/ maisou menos Main Monastirian Wiirmian interglacial arenoso Monastiriano (Sangamon) principal parte inferiorda interglacial Rissiana- Wurmiana 40-50 alt. more or less sandy 32-45 Mindelian-Rissian inter marine coastal maisou menos Tyrrhenian glacial (Yarmouth) costeiro marinho arenoso Tirreniano Interglacial Mindeliana- Rissiana 65-70 alt. more or less sandy 60 Giinzian-Mindelian inter marine coastal or deltaic maisou menos Milazzian glacial (Aftonian) coastal arenoso Milaziano Interglacial Cunziana- costeiromarinho ou deltaico i Mindeliana -* 100 alt. reworked 100 ca. Plio-Pleistocene marine coastal Belterra clay Sicilian (Pre-Nebrascan) costeiro marinho argila de Belterra Siciliano Plio-Pleistoceno remodelada 200 alt. Belterra clay 180 Plio-Pleistocene (planalto terrace) argila de Belterra Calabrian (Pre-Nebrascan) (terracode planalto) Calabriano Plio-Pleistoceno marine-lacustrine marinho-lacustrinho

Tabela 2 Idadedos terracos naparte Norte da area Cuamd-Imperatriz

32 For thesouthern section of the Guama-Imperatriz area (km 195 - km 415),drainin g to the Gurupi, the situation is as follows: In the lower Gurupi a rocky bottom of Pre-Cambrian age occurs, slightly above present day sea level, hampering navigation duringtim e oflo wtide si nconjunctio n withlo wwate r (GLERUM, 1960).Becaus eo fth e existence of this resistant sill,lo w sea levels inthemselve s willno t have influenced the longitudinal profile of the middle and upper Gurupi and its tributary rivulets. Only sea levels above the present one will have permitted deposition of sediments in the upper Gurupi system. Therefore, the deposits of established terraces in the km 195— km 415sectio n willb eo f interglacial age.N o attempt will be made to date them fully as was done for the km 0-km 195section , because of an apparent relative raising of the Gurupi head-water region, as is deducible from the position of the planalto terrace there.1 The height of the terraces above the present rivulets would not have been as high without such a raising. The 30-40 m + and 10m + terraces of the km 195-km 300sectio n mayb erelate d to theTyrrhenia n and Late Monastirian levels respectively of Table 1.

1.4.3.3 Terrace Levelsthroughout theAmazon Valleyproper Large scale topographic maps with detailed annotation of the contours do not exist for Amazonia. Severalearl yexplorer shav e however,writte n notes onth elevel so f the upland alongth erivers .I nrecen tyears ,mor esystemati cstudie shav ebee n undertaken, for instance by GOUROU (1950)an d GUERRA(1954 , 1959).Bot h SAKAMOTO(1960 )an d VARGAS (1958) attempted to correlate their observations with the various Pleistocene ages. The observations on terrace levels throughout the Amazon valley proper are listed in Table 3.I n this table,man y of the written notes are combined with personal observations and those deducible from the maps and reports of the FAO/SPVEA Forest Inventory Group.Th e latter surveyed a broad stretch of land at the south side of theAmazo n river,fro m the river Madeira inAmazona s Statet o the river Maracas- sumd in Maranhao State (Fig.22) . Many difficulties arise in attempting to correlate all the levels mentioned and deter mine their dependence on past sea level movements. First of all,th e data mainly concern scattered observations only. It is therefore not quite possible to establish whether the terrace levelconcerne d has a constant altitude or follows the river level. If the terrains concernfluviatile terraces , absolute altitudes havelittl evalu efo r ourpurpose .Th esam eapplie st orelativ eheight sabov eth epresen t river level of coastal terraces. The data themselves are often not very reliable. When altitudes are given, they are estimations in part. If instruments are used, errors of several metres are quite con ceivable in view of the vast distances (remarkable are the differences in indicated altitudes for Amazon towns, when comparing several descriptive atlases). When relativeheight sar egive ni ti sno t alwaysclea rwhethe r theyrelat et o maximum annual high-water or high-tide level, to mean annual high-water or high-tide level, to the

*) Seenot eo npag e 31.

33 Table3 Observations on terracelevels throughout the Amazon Valleyproper

Location Estimated terrace Source locacao levels (m) fonte niveis avaliados de terraco (m)

SCATTERED OBSERVATiONS/observacoesespal/wdas North of Monte Alegre/ao Norte de MonteAlegre 80 + VARGAS (1958) North of Prainha/00 Norte dePrainha 80 + North of6bido s(Serr ad eEscama)/a oNorte de dbidos 80 + KATZER (1903) (Serra de Escama) Guajarina: central course Capim, upper course Mojii, 100 alt. HEINSDUK (1958b) Acara-grande and Acari-mirimlGuajarina: curso centraldo Capim,curso superior do Mojii, Acard-grande eAcard-mirim Tome-Acu 50 + VARGAS (1958) Curud-una (cf. Fig. 11) (90) + Bragantina region/regiao Bragantina < 60 alt.

Manaus; along thelowe rcourse so fth e rivers Negro, 35^0 + GOUROU(1949) Solimoes,Tocantin s and Madeira/00 longodos cursos inferioresdos rios Negro, Solimoes, Tocantins eMa deira

Santarem (terraco de Santarem) 30 + GOUROU (1949) Southernpar t ofCanhum a and Mauesregion s(slightl y 20-30+ HEINSDUK (1958c) undulating)/pa/7e Sul das regioes de Canhuma e Maues (levemente ondulada) Curud village (70 kmEas t of Santarem)//»ovoafaode 15-20+ VARGAS(1958 ) Curud(a 70km ao Leste de Santarem) Maraj6 island (highest parts), surroundings Belem, 15-20+ GOUROU (1949) Manaus (Ponte Pe\ada)/ilha de Marajo (partes mais alias), arredores de Belem, Manaus (PontePelada) Paric6 (Monte Alegre) andTome-Acu/Par/p o (Monte 8 + VARGAS(1958 ) Alegre)e Tome"-Ac u Belem,Icoroaci , Gurupa andothe r places/Belem,Ico- 6-8 + GOUROU (1949) roaci,Gurupa eoutros lugares Cameta and Parintins, Maues, Codajas, Rosarinho 8 + MARBUT and (mouth ofth e river Madeira),opposite Itacoatiara : the MANIFOLD(1925 ) 'Maraj6 leveY/Cametd e Parintins, Maues, Codajas, Rosarinho (bocado rio Madeira), emfrente de Itacoa tiara:o 'nivelde Marajo'

Northern part ofCanhum a andMaue s region/aparte 5-8 + HEINSDUK (1958) Norte das regioes de Canhuma ede Maues Between the Baia de Pracupi and the Camaraipi 5+ HEINSDUK (1958a) (slightly undulating)/en/rea Baia de Pracupie oCama raipi(levemente ondulada)

34 Table 3 ContinuedjTabela3 Continuado

Location Estimated terrace Source locagdo levels(m ) fonte niveis avaliados de terraco (m)

Between theOeira s andth elowe r Tocantins (slightly 4-6 + HEINSDUK (1958a) undulating)/en/reo Oeiras eo curso inferior do Tocan tins (levemente ondulada) South of Igarape-mirim (northwestern part of the 5-10 + HEINSDIJK (1958b) Guajarina zone)/ao Sul de lgarape-mirim (a parte Noroesteda zona Guajarina) Narrow strips along the rivers, e.g. Santarem; villages 1-4 + on the islands of theEstuar y region; along therive r Para; the lowest course of the Tocantins; patches (tesos) within thelowlan d of eastern Maraj6 island/ faixas estreitas aolongo dosrios,p.e. Santarem;povoa- coesnas ilhasdo Estudrio;ao longodo rio Para;o curso inferior do Tocantins; tesos dentro da baixada do Leste da ilha deMarajo

Marajo (4 levels; if taken on local high-water level 20 alt. GUERRA (1959) these heights are less)/Marajo (4niveis; caso de serem 15-16 alt. determinados emnivel de enchente local,estas alturas 10-12 alt. saomenores) 4 alt. Amapa Territory (upland banks along the Amazon 4-7 + GUERRA (1954) and the Ocean coast)/Territoriodo Amapa (falesias ao longodo Amaionas e nacost a do Oceano)

TOPOGRAPHIC STUDIESO F COHERENTLARG E AREAS/ estudos topogrdficos devastas areas coerentes Caete-Maracassume area (cf. Fig.19 ) DAY (1959) Maracassume association (3levels) / 10-15 +;5-8 +; associacao deMaracassume (3 niveis) 1-3 + Pitor6 association (3levels) / 10 + (partly to 15+) ; associacao dePitord (3 niveis) 1-2 +; onein-betwee n jum entreestes Gurupi association/oMocrafao de Gurupi 2-5 + North ofPitor 6(a n 'old relatively high terrace',als o 30 + existing in thema puni t 'Agricultural Soils, undif ferentiated')/^ Norte dePitord (urn 'terraco velho relativamentealto', que tambemexiste naunidade de mapeamento 'Solos Agricolas, indiferenciados') Savannah region north of Macapa/rcyrao dos campos 80-100, 50-65,25-35 , GUERRA (1954) ao Norte de Macapa 10-15,5- 7 alt. Manaus andsurroundings/ Manaus e arredores 58 +.31 +.6 + SOARES(1956) ,dat ao f KuuXAN/dadosde RUELLAN

Tabela3 Observacoessobre niveis de terraco detodo o vale amazonicoprdpriamente dito

35 floodplainlevel , or to the annual low-water or low-tide level. Upstream, the differ ences between the annual high and low-water levels may be enormous (the mean difference at Manaus is 11 m). Differences between high-tide and low-tide level are often several metres in the Estuary region (2-3 m in the Baia de Marajo). Another difficulty lies in the possibility of raising or subsidence of areas during the Quaternary. In the Estuary region one may expect a tendency for the terraces to drown, owingt oth eexistenc e ofth e Marajo Graben (cf. 1.4.1).I n the Upper Amazon region a general gradual Quaternary subsidence is probable, in view of the low position of the planalto terrace in this region (cf. 1.4.2). If not subject to general subsidence,th e Middlean dth eLowe rAmazo nregio nhav eprobabl y experienced local vertical movements (cf. SOARES, 1956; SIOLI, 1957). STERNBERG (1950) assumed the existence of minor faults from the pattern of drainage around Manaus. The strikingly low level of thefloodplain, an d the abundanceo f lakes,i nth e Itacoa- tiara-Oroximina" area may be due to some special subsidence in this part. On the edges of the crystalline shields a tendency to gradual raising may exist, as illustrated on the southernmost part of the Guama-Imperatriz area.

As explained above, the question of the interglaciality or glaciality offluviatile terraces deposits and the possible crossing of terrace levels at some spot along the longitudinal profile ofth e river,ha s to betake n intoaccoun t aswell . At present, the longitudinal profile of the Amazon river proper isextremel y flat. Entering Brazil, 3000 km from the Atlantic Ocean, the river has an altitude of only 65 m. At Manaus, 1500 km from the Ocean, the altitude of the river is 15m during the low-water season. Consequently, tidal differences are detectable to beyond Santarem, about 700k m from the Ocean. In view of the assumed gradual subsidence of thecentra l part of Amazonas State,th e longitudinal river profile was probably less flat duringth e Pleistocene.Th edifferenc e howeverwil lno t surpass 100 m(whic h isth e estimated difference in altitude between the planalto terrace in Para and that in the central part of Amazonas State) for the oldest part of the Pleistocene. The entire Brazilianpart ,therefore , belongst oth eflatter lowe r part of the longitudinal profile of the river, wherefluviatile terrac e deposits are liable to be of ages with high erosion base, i.e.th e interglacials. Additional support for the interglaciality ofth e river terrace deposits isgive n byth e existenceofrias. Thisi sth enam efo r thestrikingl ywide ,funnel-lik e mouthso fman yo f thetributarie so fth eAmazo nriver ,fo r instance theTocantins ,th eXingii ,th e Tapajos, the Rio Negro, the Trombetas. Several of these mouths are very deep, for instance that of the Tapajos (80 m-) . An adequate explanation of the existence of the riasi s that they were excavated during the last glacial (the Wisconsin, Wurmian), when the sea level dropped to 70-100m belo w the present-day one (Table 1).Th e erosion base of the tributaries, i.e.th e bed of the Amazon proper, also will have been much lower during that age. Due to the very restricted load of several of the tributaries in Holocene times, contrary to that of the Amazon proper, these excavated mouths are only partly silted up (SOARES, 1956; SIOLI, 1957; SAKAMOTO, 1960). Having been so

36 many tens of metres lower, the Amazon cannot have left glacial deposits above its present level. As yet, the data available onfluviatile terrac e levels are not systematic and reliable enough to correlate and date them fully, in relation to the Pleistocene high sealevels . VARGAS (1958)mad ea n attempt, but iti sthough t that both hisgroupin g oflevel s and their dating are not quite correct. He assumed the highest levels,thos e of 80-50m +, to beo f Yarmouth age,an d thelowe r ones,thos e of20-3 0m + and of4-1 0 m +, to be of the two sub-phases of the Sangamon age. Only MARBUT and MANIFOLD'S 'Marajo level'( 8m +) seemst o becertain , andwidespread .Thes e authors report this levelt o extend as far as the mouth of the Japura on the Solimoes, and Borba on the Madeira. The Marajo leveli smos t probably identifiable with the Late Monastirian phase of Table l.1) It islikel y that alarg eproportio n of the terraces inth e Estuary region and along the Oceancoas t are marinecoasta l or deltaiccoastal .Thi s iscertainl y the casefo r at least the upper ones of GUERRA'S terraces of the savannah region of Amapa Territory. It is not yetpossibl et odra wth e geographic boundary between coastal and fluviatile terraces.

The Plio-Pleistocene Amazon planalto and the Pleistocene terraces of the Amazon valleyprope r will bereferre d to asPlanlcie when considered in conjunction.

1.4.3.4 Terrace Levelsin Relatively HighSituated Areas Very few data are available on the occurrence of Pleistocene fluviatile terraces in areas situated well above the highest Pleistocene sea levels, and separated from the lower course of rivers by erosion resistant sills. On the latest geologic map of Brazil (LAMEGO, 1960), extensive areas of Pleistocene deposits are mapped inth eregio n ofth euppe rXingi ian d alongth eAraguaia .Durin g the reconnaissance soil survey of the Araguaia Mahogany area, large tractso f flat landwer efoun d ata leve lo f5-1 0m abov eloca lriver san drivulets .The yar ecompose d ofver ysand y sediments and arefull y under forest cover(cf. Appendix 6,soi l mapping unit KLS, F). There can be little doubt that a Pleistocene terrace is concerned. The whole survey area is located at 150t o 300 m altitude, separated from the Amazon valley proper by a long chain of rapids, both in the Araguaia river itself, and in the Tocantins river below the confluence of the Araguaia. Hence Pleistocene sea level movements cannot themselves have exerted an influence on the formation of the terrace.Th e terrace isprobabl y a'climatic fluviatile' terrace .Th esediment smus t have deposited under a river regime that favoured deposition of coarse textured sediments over large expanses. This will have been possible only if the protecting forest cover inth e regionwa sabsent , which indicates a relatively dryclimate .Th eterrac e material

*) SOARES (1956) refers to an unedited conference of RUELLAN in 1954 at Riberao Preto, in which the latter links terraces at 80-100m ,50-65 m ,25-3 5 m, 18-20 m, 8m an d 4 m, in the vicinity of Maraj6 and Belem with the Bas Monastirien and the Pre-Flandrien of Europe.

37 therefore willhav e beendeposite d during, or atth een d of, an interpluvial, when there was not enough rainfall to sustain forest growth, but torrential masses of water occurred during the wet season, whichfiercely erode d uplands further upstream and deposited the coarse fraction of the erosion products in the area under discussion.

1.4.4 The Holocene Terrains

Next to the uplands of Pleistocene age, there are also, in a limited area within the Middle and Lower Amazon regions, terrains only slightly above normal high-water level (1-2 m +), which have clayey and reddish mottled soil. These massapes are found, for instance,o nth e savannah areas between Oroximina and Faro (SUTMOLLER eta!., 1963). VARGAS (1958) reports theirexistenc earoun d Parentinsan d Barreirinhas. The terrains are not recent, if only because the soils are fairly strongly weathered (compare Profile 3o f II.3.2). Already VARGAS (1958) suggested that these clays were deposited in a time between the deposition of the Pleistocene uplands and that of the recentfloodplain. The yar eassume d to have been laid down during the Atlantic age of the Holocene, when the sea level was relatively high (Flandrian, Table 1). The lowlands of eastern Marajo island are probably also of Early Holocene age, because they consist of sub-recent deltaic sediments. This can be deduced from the patternso frecen tan dancien t drainage,th e meso-relief, and the soilcharacteristic s on these terrains (see the section of an aerial photograph in SUTMOLLER eta!., 1963 ,an d SOMBROEK, 1962b). The floodplains and the valley bottoms are certainly of Holocene age. They are commonly dividedint otw ogroups :th evdrzeas an dth eigapos. Although theliteratur e concerned is not quite unanimous, they may be defined as follows: The vdrzeai s lowland that is intermittently waterlogged, while the igapdi s lowland that is per manently waterlogged. Defined as such, the former is found usually on extensive lowland areas, and the latter usually near or within upland areas.Th e varzeas mayb e subdivided according to the character of the variation of the water level and the quality of the water, namely: 1. Lowlands, along rivers, where the variation in water level is due to differences within the year of the river discharge (vdrzea do rio). 2. Lowlands, away from rivers, where the variation in water level is due to differ ences within the year of the rainfall (vdrzea dachuva). 3. Lowlands, inth e Estuary region,wher e freshwater tides causeth e variation in the water level (vdrzea domare). 4. Lowlands, along the Ocean coast, where sea or brackish water tides cause the variation in water level (vdrzeado mar).

Thevarzea sar epredominantl y heavytexture d(silts ,silt yclays ,clays) ,th eigapo s often peaty.Togethe rwit hth emassape san d thelowland so feaster n Marajo, the floodplains aretherefor e strikingly different inthei r sedimental composition from the Pleistocene

38 uplands in eastern Amazonia. The latter contain coarse sandy, and in several places, even stony elements. The floodplains have the customary pattern of river basin lands (back swamps or vdrzeasbai.xas) andlevee s(rive r ridgeso r vdrzeas altas). Thelatte r are sometimesver y narrow and then often run parallel (point bars or restingas). Elevated floodplains that arealmos t well-drained are rare.The yhav e beenfoun d only alongth e upper Madeira and along the lower Araguaia. This scarcity is related to the fact that at present the load of the Amazon river system is small (cf. LECOINTE, 1945), although with the enormous water discharge the total amount is enormous (SOARES, 1956). Several tributaries are practically without load, for instance the Rio Negro and the Tapajos. These are rivers of humic acids containing black water (rio deagua preta), or clear greenish or bluish water (riode agua limpa).I n contrast, rivers such as the Amazon proper andth e Madeirawhic hhav ea considerabl eloa d ofsediment san d are therefore turbid, are often called rivers ofwhit eo r yellowish brown water (riode agua branca); Siou(1951, 1956).

1.4.5 Levels of Occurrence of Fossil Plinthite (Laterite). Its Grouping and Its Dating

Throughout Amazonia, and especially inth eeaster n part, layerso freddish ,iron-ric h stony material can befound . They occur at varying altitudes and heights above local river levels, and are often accompanied by layers of highly reddish mottled, compact and soft material. Bothmaterial s are known aslaterite, and more recently asplinthite. Theorigi n ofth eAmazo nplinthit ewil lb ediscusse d inII.3 .Th emateria li sconsidere d to beth e result of the accumulation and segregation of Fe (and Al)largel y in subsoil horizons under influence of intermittently imperfect drainage. In the same chapter it will be considered how to decide whether such plinthite is still in formation or is of fossil character.Also ,i fth elatte ri sth ecase ,whethe ri toriginate din situ o rwa scarrie d along, colluvially or alluvially, from other places. With thisi nmind ,th e study offossi l hard and fossil soft plinthite mayhel pt o form a pictureo fforme r land surfaces and climaticconditions .Thi si sdeal twit hi nth epresen t sub chapter.

1.4.5.1Fossil Plinthite inthe Guamd-Imperatriz Area The situation of Amazon fossil plinthite wasstudie d in detaili nth e Guama-Impera- triz survey area, namely on the BR-14 highway proper, which had many deep and fresh road cuts and ofwhic h a complete cross-section was available (Appendix 4).I n this area, five main types of concretionary elements of hard plinthite were dis tinguished, each with a certain establishable area of occurrence1. The distinctions

') not included are the very small plinthite concretions ('shot')foun d in small quantities in a few soil profiles just above a non-plinthitic B horizon {cf. Profile 38, III.2). They are apparently of recent age.

39 between the types were only made according to morphological field characteristics, sincen o laboratory analyses are available on relative contents of elements as Fe, Al, Mn and Ti, and on the form in which these elements occur. The names employed are taken from local settlements. The types are listed in Table 4 (see also Photo 2):

Table4 The five main types of concretionary elementsof hardplinthite inthe Guamd-Imperatriz area

Name/nome MAE DO RIO IPJXUNA PARAGOMINAS LIGAQAO CAMPINHO

Size of elements small to very medium very small medium rather small (diameter) large (0.5-50cm ] )(2-1 0 cm) (0.5-2 cm) (2-10 cm) (0.5-5 cm) tamanho dos pequeno ate medio muito pequeno medio bastante elementos muito grande (2-10 cm) (0.5-2 cm) (2-10 cm) pequeno (didmetro) (0.5-50 cm) (0.5-5 cm)

Form of platy, angular prismatic to subangular irregular irregular elements or subangular reticulate blocky blocky, subangular blocky, or usually blocky to prismatic reti subangular prismatic culate forma dos laminar, em prismdtico emblocos emblocos emblocos elementos blocos angulares ate"reticular subangulares irregulares, subangulares ou subangulares, usualmente ate"prismdtico, ou reticular-pris- subangular irregularmente mdtico Grainage fine to very fine, with fine medium usually fine coarse ('iron- scattered cemented quart- medium sized zite') quartz grains granulacao fino ate" muito fino com graos fino medio usualmente fino grossoCquartzito dispersos de cimentadode quartzode ferro") tamanho medio

Colour red, dusky red red and light dusky red black or very dusky red or black red dusky red cor vermelho. vermelho vermelho preto ou vermelho vermelhoescuro evermelho escuro vermelhomuito escuro oupreto claro escuro

Arrangement (if usually vertical not specific not specific not specific not displaced) horizontal arranjo(caso de usualmente vertical nao especi- nao especi- nao especi- naoser deslocado) horizontal ficado ficado ficado

Tabela 4 Oscinco tipos principals de elementos concreciondrios de 'plinthite' duro da areaCuamd- lmperatriz

For discussion, it ispropose d to divide the studied section into the northern and the southern half.

40 Foto2 Osprincipals liposde elementos concreciondrios de 'plinthite'duro da area Guama-lmperatriz

ipixuna

wftto **?$* S*®.«

ligafao paragominas campinho iVioto2 Themain types of concretionary elementsof hardplinthite of the Guamd-Imperatriz area

NORTHERN HALF In the northern half, km O-km 195,thre e types of plinthite concretions are found, namely the Mae doRio, the Ipixunaan d the Paragominas type.

Maedo Rio typ e concretions are present only north of km 78.Th e concretions form almost everywherea layer of about 1 mthickness ,an d have a horizontal arrangement inth elowe rpar t ofthi slayer .Belo wthis ,sof t plinthite occurs,t owhic hth e transition is gradual, and which has predominantly a laminar pattern (Photo 15).Th e concre tionary layer is commonly overlain by a layer of fairly sandy sediments without concretions. In the latter layer no 'pedologic' A2 horizon, immediately above the concretionary zone, can be distinguished. It may be concluded that the concretions were formed in situ.A former Ground Water Laterite soil profile must be concerned which was truncated and afterwards buried with sediments.Th elatter , in their turn, were stripped off in severalparts .Th e topographic data'of thecross-sectio nindicat etha t the upper levelo fth econcretion s is at about 65-70 m altitude.

Ipixuna type concretions occur mainly between km 75 and km 150. They form a layer of 0.5-1 m thickness, and have a vertical arrangement in the lower part. This

41 layeri sunderlai n bya thic k layer(u pt o 10 m)o fsof t plinthite,t owhic hth e transition isgradual . The soft plinthite has acoars e reticulate to prismatic-tubular pattern. The concretions are often outcropping, but locally are overlain by a thin (less than 5m ) layer of sediments without concretions, in most places of very heavy texture (re worked Belterra clay). In this sedimentary cover no 'pedologic' A2 horizon, immedi ately above the concretions, can be discerned (Photo 18). It may be concluded that the concretions were formed insitu, and that a former Ground Water Laterite soil profile is concerned. This profile was truncated and soon afterwards buried with Belterra clay.Th e latter,i nit sturn ,wa sstrippe d off in many places.A n exception has to be made for a few stone-lines of the concretions, which are found in the sediments of a relatively young terrace in the section concerned (Photo 3). These will be of colluvial-alluvial origin. The stone-lines occur at varying depth and may therefore be designated as 'geologic stone-lines'.

Foto 3 Concrefdesdo tipo Ipixuna que parcialmente cobrem 'plinthite' macio e parcialmenteformam lin- hasdepedra em terra fridvel enao mosqueada. As concrefdes daslinhas de pedra sao depositos aluvianos oucoluvianos. As concrefdes a esquerda, pore'm, originaram-seprincipalmente insitu de 'plinthite'1 macio anterioracima do 'plinthite'macio atual(ahaixo a esquerda).Eprovdvel que ocorreualgum movimento lateral destas concrefdes, vistaa delimitacao bastante pronunciada entre as concrefdes e o 'plinthite' macio(BR-14, km 125)

-•tAU-' Photo3 Ipixuna-typeconcretions, partly capping softplinthite, partly forming stone-lines in unmottled, friable earth. Theconcretions of thestone-lines were deposited alluvially orcolluvially. The concretions at left, however, mainlyoriginated i n situfrom former soft plinthite ontop of thepresent soft plinthite (bottom left). Some lateralmovement of the latterconcretions islikely tohave takenplace owing tothe fairly sharp boundarybetween theconcretions and the soft plinthite (BR-14, km 125)

42 In a few places both Ipixuna type and Mae do Rio type concretions occur (km 78, km 63, km 30). There, the Ipixuna type always overlies the Mae do Rio type, and is sometimes vertically separated from it by soft plinthite. This indicates that Ipixuna typeconcretion soriginate d ina les sancien ttim etha n theothe rtype . The upper level of Ipixuna type concretions is at about 75 m altitude at km 78, gradually increasing to 190m altitude at km 195.

Paragominas type concretions are found predominantly south of km 150, although they have a scattered occurrence north of this (thin layers - less than 0.5 m thick - above the Ipixuna type; usually not indicated inth e cross-section). These concretions form a layer of 0.5-1.0 m thickness.Thi s layer is usually underlain by soft plinthite, to which the transition is gradual in most places. The soft plinthite is present only in thin layers - generally less than 5m- and has a fine reticulate pattern. The Para gominas concretions are partly outcropping, partly overlain by Belterra clay. In the latter case no 'pedologic' A2, immediately above the concretionary zone, can be discerned. The Paragominas concretions may therefore be taken to have predominantly originated insitu.

At places where the Paragominas and the Ipixuna type concretions occur together (for instance at km 171,k m 194), the former overlie the latter, indicating that the Paragominas ones are of later date. The Paragominas type concretions are believed to have originated mainly within the Belterra clay. Their size and form is similar to the structural elements of the clay. Recent reddish mottling (i.e. the pre-stage of concretions) in the Belterra clay is related to these structural elements, by being fine spotty or fine reticulate. This type of mottling is found in a few places above the layer of Ipixuna concretions, and very distinctly in the Bhorizo n of present Ground Water Laterite soil profiles in Belterra clay (cf. Profile 8 in II.3.2.1). In fact, transitions between Belterra clay mottling and the loose Paragominas concretions are found at several places. The position and the form of the Ipixuna type concretions lead us to assume that they did not originate in Belterra clay, but in the upper layers of the planalto base material(mediu mt o heavytexture dsediment so fth e Barreiraso r Alter doCha obeds) . The upper level of the Paragominas concretions is varying. Where they are not underlain by Ipixuna type concretions, the level is between 110an d 170 m altitude, 30-50 m above local river level.

SOUTHERN HALF In the southern half of the stretch, between km 195-km 467, Ipixuna, Ligagao, Campinho and Paragominas concretions are found. In general, the amount of con cretions and soft plinthite is much less than in the northern stretch.

Ipixuna type concretions are outcropping at the planalto scarps in the km 195-

43 -km 250section 1.The yoccu rimmediatel y belowth eBelterr aclay ,a tabou t 10m belo w theleve lo fth ecentra l sectionso fplanalto .The yar efoun d inlayer so f2- 3m thickness , and are underlain by their soft plinthite (see under northern section). It may be concluded that they were formed insitu.

South of km 250, Ligaqao type concretions are found at the scarps of planalto,1 also immediately below the Belterra clay, at about 20 m below the level of the flat central sections.Thes econcretions , forming a layer of 1-2 mthicknes s or less,ar e not underlain by soft plinthite (except at a few spots around km 270an d km 405),bu t by uniformly reddish coloured sediments of sandy clay loam to sandy clay texture. Normally the boundary isshar p (Photo 4).Th e Ligagao concretions are more or less rounded, and have no pattern of arrangement. It is therefore concluded that the Ligagao concretions are ofalluvia l (fluviatile) origin in mostplaces . The level of the Ipixuna or Ligagao type of concretions isincreasin g gradually from north to south, together with the surface of the planalto terrace. At km 195 the altitude2 is 190 m, at km 412 it is 330m .

Paragominas type concretions are found scattered. Sometimes they occur in thin layersjus t above the Ipixuna or Ligagao types at the planalto scarps, where they are without soft plinthite.A t lowerlevel s Paragominas typeconcretion s may form thicker layers,an d arethe ni nsom eplace sunderlai n bythei rsof t plinthite (seeunde r northern section). At these latter places, for instance at km 269, they were certainly formed insitu. Concretions ofvariou stype sar efoun d collectively on severalslope san d terraintops , below the level of the planalto terrace (Ipixuna + Paragominas; Ligagao + Para gominas).A tthes eplace sth econcretion sar ever ymixed .The yma yoccu ri na scattere d pattern in the sedimentary layers, but more often they are concentrated in thin layers(0. 5m )a tth etransitio n zonefro m theyellowis h Ahorizo n toth e red Bhorizo n of the soil profiles (most probably they are concentrated there under influence of soil biological activity, and may therefore be designated as 'biotic stone-lines'). Soft plinthite is lacking. There is every indication that colluvia-alluvia from the planalto scarps are involved.

In the stretches km 195-km 220an d km 250-km 260 concretions of theCampinho type are found. They are partly outcropping, partly covered by sediments that are usually of sandy clay loam texture. These concretions form layers of about 0.5 m thickness and, according to a limited number of observations, they are underlain by

') Theconcretion s and thesof t plinthite areno t confined only to thescarps .A t several deep road cuts it was established that both plinthite layers continue horizontally towards the central sections of planalto.Ther e are, however, indications that the layer of soft plinthite has its greatest thickness near the scarps. It also seems that the Paragominas concretions, when occurring on top of the Ipixuna or Ligacao types,ar e concentrated at the scarps. *)See , however, note at page31 .

44 Folo 4 Concrecoes dolipo Ligacao que cobrem terranao mosqueada e fridvel.A ausencia de qualquer 'plinlhire' macio,a delimitacao inferior pronunciada e irregularda capade'plinthite'duro, a forma subangular dos mesmos elemenlos concreciondrios e a ausencia de qualquer arranjo destes elementos, em conjunto indicam que a capade'plinthite' duro naofoi formada in situ,mas i deorigem aluvial(BR-14, km370 in. ou m.)

Photo 4 Ligacao typeof concretions capping unmottled,friable earth. The absenceof any soft plinthite, thesharp andirregular lowerboundary of thelayer of hardplinthite, the subrounded form of the con cretionary elements themselves, andthe absence ofany arrangement ofthese elements, together indicate that thelayer of hard plinthite was not formed insitu , butis of alluvial origin (BR-14, km370 ca.)

soft plinthite of only 1-2 m thick. It may be assumed that the Campinho type was formed insitu an d isth e relic of a Ground Water Laterite soil on a land surface that was at about 30 m above the present local river level. South of km 415,concretion s are sparse.The y areo f varying type (at several places platy, similar to the Mae do Rio type of the northernmost stretch), and may grade into some soft plinthite. The observations made in this section are however too limited to warrant deductions as to the origin and level of theconcretions .

An attempt will now be made to relate the different types of fossil plinthite with times of Ground Water Laterite soil formation, i.e. with the former existence of approximately flat land surfaces with intermittently imperfect drainage. The scarcity of reliable data on the surface geology in this region of unconsolidated sediments, is a seriousdrawback . There are,however , a few geological notes on outcropping older

45 deposits at the extreme north and south sides of the part of the BR-14 studied1. Personal observations on the characteristics of the substratum were made at deep road cuts,an d there are the conclusions about the age of terrace deposits (cf.1.4.3.2) . These together enable a sketch to be made of the geologic-geomorphologicgenesi so f the region.Thi s isdon e in Appendix 5,fro m which the age of the fossil plinthite may be established as follows:

After deposition during the Miocene of quartz sands and kaolinitic clays of the Barreiras or Alter do Chao beds, a time of stability existed, when neither sedi mentation nor erosion took place. During this time, an approximately flat land surface, which would be nowadays at 65 to 70 m altitude, permitted Ground Water Laterite soil formation inth e northernmost part of the area (km 0-km 70)becaus eo f imperfect drainage of the surface. This was followed by a time of erosion during which these profiles were truncated. The Mae do Rio type concretions originated by hardening ofth esof t plinthite atth e surface:Late Miocene (?). Therear e indications, for instance at km 30,tha t the Mae do Rio concretions are the relics of at leasttwo , superimposed, Ground Water Laterite soil profiles. This would indicate the existence of more than one time with imperfect drainage of former land surfaces before, or atth ebeginnin gof , thePliocene .Suc ha situatio nwoul dals oaccoun tfo rth evariabilit y in the characteristics of the Mae do Rio concretions. After this truncation a new deposition of sediments took place, similar to those described above, during a transgression time: Pliocene (?). Subsequently, there was another time of neither sedimentation nor erosion, but soil formation. Imperfect drainage of an approximately flat land surface (which would be nowadays at varying altitude; generally increasing from north to south: 70 m altitude at km 80, 120 m at km 145, 145 m at km 195, 320 m at km 405) resulted informatio n ofGroun d Water Laterite soilprofile s inth epar t km70-k m300 , and possibly around km 400.Thi s wasfollowe d by an erosive time (probably short, because of the flatness of the concretionary layer), during which truncation of these profiles and formation of Ipixuna and Ligagao concretions took place. At the same time, the land surface in the section between km 300 and km 400 was covered with alluvial Ligacao concretions:Late Pliocene (?). This was followed by a regression time, during which in several parts all the concretionary land surface was stripped off, especially in the km 150-km 195sectio n (Alto Guama):Late Pliocene or Earliest Pleistocene: Biberian (?). A new transgression time, with lacustrine - marine conditions, caused the de position all over the area of very heavy textured sediments (Belterra clay) in a thin layer (20 mthic k at km 400, 10m thic k at km 200):Plio-Pleistocene (Calabrian);cf. 1.4.2. Subsequently, a regression time caused strong erosion, resulting in a landscape of table lands (the present day sections of planalto terrace). This erosion was most

*)Fo r details cf. SOMBROEK (1962a).

46 pronounced in the central and southern parts ofth e area.Th e sharp edgesar e due to the presence of the erosion resistant layer of concretions. The same erosion gave gentle undulating terrains in the stretch km 150-km 195 (Alto Guama) where this resistant layer was largely absent: Early Pleistocene: Danubeian (?). Renewed raising ofth egenera lerosio n base resulted inflattening o fth eBelterr a clay covered land surface until it became a terrace of about 100 m altitude (re-worked Belterra clay): Sicilian level of Early Pleistocene; cf. 1.4.3.2, Table2) . Resumed erosion during a regression time hereafter resulted in sharpening of the tablelan dfeatur eo fth egenera l landscape: glacial ageof Pleistocene: Nebrascan (?). Alternating transgression and regression times during various interglacials and glacials of the Pleistocene followed. During the transgressions, sedimentation took place at various levels (cf.1.4.3.2) . In one of these ages, possibly the Yarmouth interglacial, a short time of stability existed during which a land surface with imperfect drainage in the stretch km 200- km 260, which would be nowadays at about 30 m above the local drainage system, gave Ground Water Laterite soil formation. These profiles, after truncation by a subsequent erosion, resulted in Campinho type concretions.1 At the same time, or shortly before, the main part of the Ground Water Laterite soil profiles that gave Paragominas concretions,developed , predominantly inth e stretch km 150-km 195,a t a level 30-50 m above the present day local river level. The age of the latter type is, however, difficult to assess as a whole, because it is found in some places also im mediately above the Ipixuna or Ligacao concretions at the scarp of unattacked sections of planalto. For these, Plio-PIeistocene age seems plausible. On the other hand, the Paragominas concretions appear to be formed in present times too, as pointed out before.

Summarising, it may be said that a large proportion of the fossil plinthite in the Guama-Imperatriz area was formed insitu, during and after at least two epochs of the Late Tertiary (Miocene and Pliocene) with approximately flat land surfaces of imperfect drainage. During a part ofth e Pleistocene plinthite of limited expanse was produced.

1.4.5.2Fossil Plinthite throughout Amazonia The classification and dating of the fossil hard plinthite elsewhere in the region requires further research. For full comparison with the situation in the Guama- Imperatriz area, itwoul d also benecessar yt o uselaborator y data onfossi l plinthite as a means of identification. Provisionally the following may be said:

') Their formation during a time of the Early Tertiary or even of the Late Cretaceous might be considered. This old age is however more likely to be true for the blocks of 'iron cemented quartzite' that are encountered on a fewplace si n the stretch km 195k m - 220, but not along the highway itself. Should this be the case, comparable fossil plinthite would occur deep down (about 100 m) under the sections of planalto terrace. This could not be checked.

47 1. The fossil plinthite situation in the region southeastof Santaremwa s studied to someextent . For Curua-una centre,th edat a ofth e road from therive rt oth e planalto are giveni n Fig. 11.Concretionar y elements are outcropping atth e scarp of planalto. This hard plinthite is found immediately below a 15 m thick cover of Belterra clay. Thelaye ri sabou t 1 mthick , and hasi n manyplace sa gradua ltransitio nt o underlying soft plinthite, 1-2 m thick. The concretions are of a type very similar to the Ipixuna type of the Guama-Imperatriz area. Their arrangement is vertical in the lower part, sometimes strikingly disturbed in the upper parto f thelayer .Immediatel y above the Ipixuna-like concretions, namely in the lowest part of the layer of Belterra clay, a thin layer (about 0.5 m) of concretions resembling the Paragominas type are found. On the lower slope of the planalto escarpment, Ipixuna-like concretions are also present, but inthin ,irregular , superficial layers.Thes ear ewithou t soft plinthite,bein g underlain by homogeneous reddish, unconsolidated sandy sediments (planalto base material, Barreiras or Alter do Chao beds). Throughout the length of the road before the planalto escarpment, plinthite is completely absent in any of its forms. Only in close proximity of km 0 a degree of plinthitic material isfound . At an upper level of about 16m above the river level,a superficial layer ofhar d plinthite of about 1 mthicknes si spresent . The concretionary elements are resembling more or lessth e Campinho type concretions of the Guama- Imperatriz area. The concretionary layer has a gradual transition to a layer of under lying soft plinthite, which is 1.5 m or more thick and has afine reticulat e pattern of mottling. Nearby, but at a lower level, namely 9 m above local river level, also concretions are found at the surface, however of quite another type. They resemble portions ofth e Mae do Rio type ofth e Guama-Imperatriz area. No arrangement was observed. Whether they are underlain by soft plinthite or not, was not established. It may be concluded that the Ipixuna-like concretions at the upper planalto scarp were formed insitu, an d are probably of Late Pliocene age.Thos e at the lower slope are colluvial. TheCampinho-lik e concretionsnea r the river were formed also in situ, possibly during some time of the Pleistocene. Their formation at this spot may be due to an impervious substratum of olderconcretions ,tha t possibly wereforme d - or deposited - there at sometim e of the Tertiary, before the Pliocene. Similarity in type as defined in the foregoing does,no t necessarily imply an identic time of formation of fossil hard plinthite. The similarity between the whole fossil plinthite situation at Curua-una and that inth e Guama-Imperatriz area, in particular the stretch around km 200, does however suggest that with morphometric studies correlations can be made over large areas within Amazonia. The conditions as they exist on the planalto scarps in the Guama-Imperatriz area and at Curua-una centre, werefoun d also at Belterra Estate alongth e lower river Tapajos, and behind Tucurui on the riverTocantins . Onei sforce d to believetha t allo fth eorigina l planalto terrace in eastern Amazonia has, below its blanket of Belterra clay, a layer of fossil plinthite with Ipixuna (or Ligacao)-like concretions. The conditions may be different for the remnants of planalto in the western part of Amazonia. Two examples will be given. In the Manaus-Itacoatiara area, the lower

48 Fig. 11 Mapa de topografia e solos da area ao redor do centro de Curud-una e corte transversal AB da eslrada para o planalto

Altitude , . o/turo H 190T

>/ KLS = Kaolinilk Lalosolic Sand /fl loamysan d AreiaLalosiHica Caolinitica '* ore/o barrenta KYL .- Kaolinitic Yellow Latosol. very heavy textured LatosoloAmarelo Caolinitico. de textura rnuitopesada loamysan d oreioiorrent a

yS loamy sand f oreio barrenta

V" "'^CA. It?clay ?'•>• •g argila Distance(km)/feifometragem 5 6 -t—i—i—i—I—i—i—i—i—I—I—I—i—I—I—I—I—I—i- SURFACETEXTURES IRON CONCRETIONS: HARD PLINTH1TE Texlwas de superficio Concrecoesferruginosas: 'Plinlhite' duro I I light textured lii'i I'I| prismatic-vesicular, fine grained (eolpixuna) textura leve prismatico -vesicular, granulos ftnos

medium textured "7*1 pisolithic, fine grained (ooParagominas) textura media pisolitico, granulos finos

rather heavy and heavy textured \."/", | blocky, fine grained (ooCampinho) textura meio pesada e pesada em blocos, granulos finos

very heavy textured | ssg: | platy, coarse grained (coMaedo Rio) textura muito pesada laminar, granulos grossos soft Plinthite 'Plinlhite' macio

Itit" 11'! 'Terra Preta' Terra Preta

Fig 11 Map of topography and soils of the area around Curud-una centre, and cross-section AB of the road to the planalto

49 boundary of the Belterra clay isonl y locally marked byconcretionar y elements,ther e as thin (1-3 dm) stone-lines, often without any soft plinthite. The concretions have similarity to the Paragominas type. The absence of Ipixuna-like concretions may be related to local re-working of theBelterr a clay duringth eEarl y Pleistocene(cf. 1.4.2). East of Porto Velho, at cascalheiras alongth e highway BR-29,thic k (3-5m )layer so f hard plinthite were observed. The concretionary elements thereof resemble the Ipixuna type, and often have a (sub)vertical arrangement. Soft plinthite is seldom found. The hard plinthite ispresen t at a slightly higher leveltha n the planalto at the spot (cf.1.4.2) ,thu s forming hilly outcrops (cf. GUERRA, 1953).

2. Inth eBragantina area, located directly north of the Guama-Imperatriz area, fossil plinthite is very frequent. The concretionary elements here often have the aspect of quartziticsan dston ewit habundan t ferrugenous cementing('iro n cemented quartzite'; gres do Para or pedra-Pard of several geologists) and seem to be predominantly similar toth e Maed o Riotyp eo fth eGuama-Imperatri z area.A si nth elatte r area,th e present day land surface is undulating to rolling where this type of fossil plinthite occurs at a shallow depth. This in contrast to the almost flat surface inth e stretch of the Ipixuna or Ligacaotypes .I t istherefor e likely that in the Bragantina area a large proportion of thefossi l plinthite is of Miocene and/or older age. Here also, the com plexity of the characteristics of the concretionary elements indicate more than one epochs of Ground Water Laterite soil formation in those epochs. On one location, namely Pirelli Estate,abou t 20 km east of Belem,eve n a complete and undisturbed buried Ground Water Laterite soil^; A2; B2of-soft-plinthite ) was found, at 4-5m depth. On top of this Mae do Rio-like concretions were found, grading downward into soft plinthite.

3. In the area of the upland savannahs of AmapdTerritory, plinthite both hard and soft, isals over yfrequent . CARNEIRO(1955 )produce d photographs oflaterita vesicular, laterita pisolitica, arenito lateritica (gresdo Para) an d argila lateritica, all from this area. Also GUERRA (1954) studied the crusts and sub-surface horizons of plinthite in this Territory, calling them canga cavernosa, picarra or arenito-Pard. From a limited number of personal observations in the area it is concluded that, although existing Ground Water Laterite soils occur, most of the plinthite is fossil. Part of the fossil hard plinthite wasforme d insitu becaus e it isunderlai n by related fossil soft plinthite. After truncation, it was partly buried by Pleistocene sediments. However, fossil hard plinthite of alluvial-colluvial character also occurs. The fossil plinthite in Amapa Territory is probably of various ages in the Tertiary and the Pleistocene. A small proportion of the hard concretionary elements there is apparently not fossil but recent:th evesicula r blocksa tth efalesias o f Macapa fortress and someothe r locations are the laterally exposed and therefore hardened parts of plinthite which is still in formation. The same seems to be the case with much of the pedra-preta on other beachesan d banks inth e Estuary region (cf.II.3.2.1 ; plinthite-in-formation belowth e solum). GUERRA (1959,p . 34-37) gives data on the occurrence of several horizons of

50 fossil plinthite, below the present day ground water level, in SESP drillings at Soure (Marajo island). From his limited descriptions of the horizons encountered, it is deduced that the fossil plinthite involved partly constitutes truncated Ground Water Laterite soil profiles and partly alluvial deposits.

4. Fossilplinthit eseem s to be frequent on the strip of uplands in the Lower Amazon region directly north of the river. The age is probably diverse, in view of the varia bility in type. This may however be only related to the outcropping of Paleozoic deposits of varying character in this region.

5. Other areas of frequent occurrence of fossil plinthite are indicated to be the Early Tertiary and Cretaceous peneplanation surfaces (cf. 1.4.1) in the watershed regions of Amazonia. The fossil plinthite would conservate the table land feature of these terrains. On the presumably Cretaceous plateaux on the watersheds of Para, Rhondonia and Mato Grosso, fossil plinthite seems to form a sub-superficial crust, outcropping at the edges (cf. notes of GUERRA, 1953, about the Serra de Parecis in Rhondonia Territory). Further data on these plinthites isno t yet available.

1.5 The Vegetation

The vegetative cover of Amazonia in its entirety is often called hileia after VON HUMBOLDT. DUCKEan d BLACK(1954 )mentio n manyfamilies , generaan d speciesthat , to a greater or lesser extent, are characteristic for the hileia. AUBREVILLE (1958) specifies some of the differences in structure and composition of the hileia in com parison withtropica lforest s inAfrica . DUCKEan d BLACKfoun d thelimi to fdispersio n of the genus Hevea, its most famous trees,t o constitute a well-fitting boundary of the hileia (Fig. 13).Th e hileia is present not only in northern Brazil, but covers also the main part of the Guianas, the northern part of Bolivia, the eastern parts of Peru and Colombia, and a part of Venezuela. Within Brazilian territory the hileia comprises about 3.500.000 km2.Th e latest determination of its southern and eastern boundaries was carried out by SOARES(1953) , from aerial photograph analysis. Contrary to many popular descriptions of Amazonia, the hileia is largely far from impenetrable.O nth eupland si nth eaxia lpar t ofth eregion ,th eforest s predominantly consist of high trees. Below the tree layers, the vegetation is relatively open. Shrubs, vines and palms are generally sparse. Consequently, it isno t difficult to traverse the forest. The atmosphere below the canopy isfairl y pleasant because of itscoolness . Although Amazonia contains the world's largest expanse of primeval tropical forest, there are within the region several parts with savannah or savannah-forest. Near the population centres there are terrains under cultivation or covered with secondary shrub and/or-forest. This isth ecas enotabl y inth e Bragantina area (east of Belem), along the lower Tocantins river (Cameta), around Manaus, Santarem and Amapa (Fig. 12).

51 Fig. 12 Mapa provisional da vegetacao da Amazonia brasikira It" w -t

Upland forests Fresh water swamp forests HH Mangrove Matas de terrafir me Matas de vdrzeaou igapo Manguais

Upland palm forests Uplands with agriculturean d secondary forests Cocais Terra firme cultivadae capoeiras

| 1 Upland savannahsan d savannah-forests Lowland savannahsan d savannah-forests Campos, campinase campina-ranas Campos devdrzea Fig. 12 Provisional vegetation map of Brazilian Amazonia

Apart from the collected information on the phytogeography of Amazonia by DUCKE and BLACK (1954),th e FAO/SPVEA Forest Inventory Program has provided lately for many systematic data (cf. HEINSDUK and MIRANDA BASTOS, 1963). Con tinuously from 1954t o 1962, exploratory surveyswer ecarrie d out overth evas tare ao f 200.000 km2. The areas constitute together an almost uninterrupted, longitudinal strip, 1700k mlong ,a tth esout hsid eo fth eAmazo nriver ,fro m thetributar y Madeiras in Amazonas State to the river Maracassume"i n Maranhao State. A latitudinal strip was surveyed along the forested, upper part of the newly opened BR-14highway . As

52 well as this, two special surveys were performed. They were to determine the dis persion andth etimbe rvolum eo ftw ovaluabl especies .On etoo k placeo nth elowland s along the lower Tocantins river for the species Ucuuba branca (Virola surinamensis), and the other along the lower Araguaia river for the species mahogany or Mogno (Swietenia macrophylla). A provisional vegetation map is given in Fig. 12.

1.5.1 The Forest Vegetation

The coverage of primeval forest is far from homogeneous. The most conspicuous differentiating factor is the drainage condition. The following formations may be distinguished:

1.5.1.1 LowlandForests The forests of the terrains with poor drainage may be divided into three main formations:

1. Mangue: mangrove forest. The forest of the lowlands that are under influence of sea or brackish water (vdrzea domar; cf. 1.4.4). The formation is present along the Ocean coast and in narrow bands along the Baia de Marajo and the Rio Para. Its composition isa par t ofth egenera lflora o fth ecoast so ftropica lAmeric a(Rhizophora mangle, Avieennia sp.) and does not have relation to the composition of the hileia.

2. Mata de vdrzea:fres h water marsh forest1. The forest on the lowlands that are intermittently water logged with fresh water (vdrzea do rio,vdrzea da chuva, vdrzea do mare; cf.1.4.4) .I t hasfewe r species than the dryland forest, and the timbers are often lighter (morefloaters). Salien t speciesar efo r instanceth e Sutnauma (Ceibapentandra) and the Andiroba (Carapd guianensis),an d in some areas Ucuuba {Virolasp.) . Several subformations may be distinguished. They are determined by the length and frequency of flooding and the quality of the flooding water. Upstream, and on the highest parts, the levees and point bars, of the floodplains in the lower courses of rivers,wher efloodin g occursonl ydurin gth epea ko fth e high water time,th e quantity of species is largest. The formations concerned are nearing the dryland forest as regards their composition. In the region with tidal influence the quantity of speciesi s low, especially on thefloodplain part s that areflooded daily . The forest on the latter have a rather closeresemblanc e to the formation 3.Ther e isals o a difference in com position and timber volumeo fth eforest s on thevarzea salon griver swit h transparent water {dguapreta o r dgua limpa; cf.1.4.4 )an d those on the varzeas along riverswit h

l) The applied distinction between swamp and marsh corresponds with the distinction between vdrzea and igapo.I t is in agreement with the classification of BEARD (1955), with the exception that lowlands flooded only during high tides (vdrzea do mare) are included with the marshes in the present publication.

53 turbid water (dgua branca). The former generally have species of harder and better distinguishable kernwood (DUCKE and BLACK, 1954).

3. Mata de igapd: fresh water swamp forest: The forest on the lowlands that are permanently water logged withfres h water (igapo;cf. 1.4.4) . Thisfores t formation has fewtre especies ;palm sar eofte n frequent, sometimesconstitutin gth eentir evegetation . Characteristic speciesar e for instance the Ucuuba (Virola sp.)an d the Anani(Simpho- niaglobulifera) andth epalm sAcai( Euterpe oleracea) and Paxiuba (Iriartiaexorrhiza). The boles and crowns of the trees are usually laden with epiphytes.

1.5.1.2 Upland Forests (matasde terra firme) These are the forests of the freely draining terrains of low altitude, in general all the land older than Holocene. These dryland or upland forests are characterised by an enormous variation of tree species, mainly of heavy timber. Palms are generally sparse, although many species of them are known (AUBREVILLE, 1958). Theforest s areno tquit euniform .Ther ear edifference s inmea ngros s timbervolume , and severalspecie shav ea distinc t optimum insom epar t ofth e immense region or are even restricted to specific areas. DUCKE and BLACK (1954), although admitting 'we cannot establish phytogeographical subregions in the hileia, because the flora of the high land between the navigable rivers is still almost completely unknown', made a tentative and schematic division of the hileia into subregions (Fig. 13). Each of their subregions has specific distinguishing species. Their sub-region 'Norte' is richest in species,bu t the trees are generally lesshig h and the leaves smaller and darker than in the other regions.Th e subregion 'Norteste' contains several stretches of relatively dry and low forest. The FAO/SPVEA Forest Inventory Program, already mentioned, has provided much quantitative data on the 'Sul' and 'Sudeste' regions. Several units, with geogra phical nomenclature, have been separated and mapped (cf. HEINSDIJK and MIRANDA BASTOS, 1963,an d Fig. 22). An inventory unit was defined by: 1. the presence or absence of one or more tree species, in contrast to the situation in surrounding units. 2. the difference in mean gross timber volume, and 3. the degree of anthropogenic influence.

Thecharacteristic s for each unit wereobtaine d by grouping ofon ehectar e sampling units,whic hwer e laid out scattered (unbiased), as much aswa s possible at themod e of penetration. For the exact tracing of the geographic boundaries of a unit, the topography and the soils of the area with the most representative sample units were also taken as critera to some extent (for more details on the inventory system used cf. IV, Introduction). It mayb esee ntha t the inventory units 1-15 do not vary greatly in mean grosstimbe r volume.Ther e isn o clear change in the botanical composition of these fifteen units. HEINSDIJK (1960), after noting the tendency to increase and decrease respectively in

54 Fig. 13 A subdivisao provisionalda hileiaem regioesfito-geogrdficas conforme a DUCKE and BLACK (1954) e os limitesfixados por estesau lores da dispersao dogenero Heve a

0--

V Hevea undifferentiated Heveabi'ntliamiana A Hevea ntio diferemiada Hevea brasiliensis Fig. 13 Theprovisional subdivision of thehileia into phytogeographical regions according to DUCKE and BLACK (1954), andthese authors' boundaries of the dispersion of thegenus Hevea

frequency of several botanical families from east to west,-state s 'a regular change in representation of the families in the forest types sampled, in one direction can, except for the questionable examples given, be regarded as non-existent'. In fact, these units form together one forest type, called 'Pouteria association' by HEINSDUK, after itsmos t stableconstituent . Thevariou s inventory unitsar efacie s ofthi son etyp e (for their naming cf. HEINSDUK, 1960). Nearing the south-eastern boundary of the hileia (units 16-24),ther e are more differences, not only in mean grosstimbe r volume but also inth e botanicalcomposition .Th elatte r unitsfo r apar t form othertype stha n that established for the centre of the valley.Fo r the units 16,17an d 18, for instance, the name 'Eschweilera Pouteria association' is proposed. In the transition area which islocate d along the lower Araguaia river, unit 26,wit h soilsdifferen t from those ofth e other inventory units,th ecompositio n of the forests is strikingly different from those near the Amazon river {cf. GLERUM and SMIT, 1962b). The number of tree species is considerably smaller. Parts of the area bear an almost

55 pure stand of mahogany {Swieteniamacrophylla), whichtre e isreporte d to occur also in other parts of the southern and southwestern transition zone of the hileia {cf DUCKE and BLACK, 1954).

1.5.1.3 SpecialUpland Forest Types 1. Cipoal(creepe r forest). This is a kind of upland forest that is very difficult to traverse, contrary to the predominant situation. It consists of a dense clew of many thick creepers and climbers, with only few, big,emergin g trees,ofte n in small groups and ladenwit hclimber s('climber-towers') .Th etransitio no fthi styp et oth ehig h forest is called cipoalicforest {cf.Phot o22) . Cipoal proper ispresen t in patches on the central sections of the planalto south-east of Santar^m(withi n units Planalto I and Planalto II), in rather extensive areas on the central sections of the planalto in the Guama-Imperatriz area (units Ligacao, Acailandia), and in very extensive areas between the rivers Jaracii and Xingii (units Planalto II baixo, and especially Planalto II baixo cipoal;cf HEINSDIJK, 1957an d the Figs.2 2an d 24).Th e inventory unit Upper Guama consists mainly of cipoalic forest. A vegetation somewhat comparable to the cipoal isfoun d in the western part of the Araguaia Mahogany area. There however creepers and climbers are generally thin, the field layer is dense and palms abound. There are indications that other parts of southern Pari and parts of northern Mato Grosso have also cipoal-like forest. The cipoal may be taken to represent the real 'jungle'.

2. Tabocal (bamboo forest). This is a type of upland forest which consists predomi nantly of Bambu and/or Taboca, both Guadua species.The y form an almost impene trable cover of 3-10 m high. Emerging trees are practically absent, except for small patches. The tabocal vegetation is present, in broad strips, along the headwaters of the river Gurupi. {cf GLERUM and SMITH, 1962a; cf. Fig.23) . Bamboo species are lacking elsewhere in area of the hileia, except for the extreme southwestern part {Guaduasuperba ofAcr e State; cf. DUCKE and BLACK, 1954).

3. Cocal(pal m forest). In part of the transition zone between the hileia and the savannahs of North-Eastern Brazil, palms are very frequent. Often these palm forests arepredominantl y oTBabacii (Orbygniaspeciosa). Thi suplan d forest typei sknow na s cocal, its area as zonados cocais.

1.5.2 The Savannah and Savannah-Forest Vegetation

In this publication, 'savannah' isconsidere d to beth e vegetation type which consists of grasses with or without scattered small trees or clumps of trees ('open savannah'), as well as the vegetation type which consists mainly of shrubs, of variable density ('shrub savannah'). 'Savannah-forest' is considered to be that vegetation type which consists of trees that reach maximally about 20 m height and have thin boles (less

56 than 0.25 m diameter at breast height). It seems actually to be more equivalent to 'savannah-wood' than to 'savannah-forest' as these vegetation types have been re defined recently. LINDEMAN and MOLENAAR(1959), for instance, describe the savannah- forest as having two storeys of which the upper one reaches about 25-30 m height, and the savannah-wood as having one storey, 8-15 or 20 m high, the trees having all thin boles. The vegetation type here considered as 'savannah-forest' is discussed together with the savannahs, in view of its edaphic relationship with the latter (cf.Chapte r IV). The savannahs and savannah-forests of considerable expanse are mapped in Fig. 12. Only the boundaries of the ones near the main rivers are traced with a degree of accuracy, because only of these parts reliable maps exist, from interpretation of aerial photographs (AAF, 1942). The boundaries of the savannahs of Marajo island, Amapa Territory and the Lower Amazon region are taken from AAF topographic maps, personal observation (cf. SUTMOLLER et a!., 1963), and local information. The savannahs and savannah-forests of the lower Gurupi and those of the Rio Para and the lower Tocantins (App. 7) are copied from forest inventory maps (HEINSDIJK, 1958a; 1958b; GLERUM, 1960). The savannahs and savannah-forests between the medium courses of the Purus and the Madeira are copied from BRAUN and RAMOS (1959). The savannah areas in the sur roundings of Manicore along the lower Madeira (Fig. 26) and those south of the Rio Negro, upstream of the confluence of the Rio Branco (Barcelos), are not reported in any publication known, but sketched in from indications on AAF maps.Th e savannahs of Rio Branco Territory are copied from GUERRA (1959). The savannah areas between the Lower and Middle Amazon regions and the Guianas are copied from the map of LECOINTE (1907),wh o bases his mappings on many observations of his own and those made earlier. Those of the medium course of the Xingu are taken from BOUILLENE (1926). The boundaries of the latter and, for that matter, even their existence, are however far from sure. The savannahs along the middle course of the Tapajos are reported by BOUILLENE (1926) as well as by LECOINTE (1922) which latter called them Campos de Cururii and Campos de Mucajazal. Also DUCKE and BLACK (1954) mention the existence of savannahs in this region, namely near Cachoeira de Manga- bal. The whole southern and southeastern boundary of the hileia is copied from SOARES (1953). The savannahs and savannah-forests within Amazonia are far from homogeneous. A well defined classification with full reference to similar vegetation types in other countries is however lacking. DUCKE and BLACK (1954) give a preliminary classifica tion with local nomenclature. Of part of the Amazon savannahs and savannah- forests, the FAO/SPVEA Forest Inventory Program has provided for the geographic coverage. Although such areas were by-passed during the forest inventories, and any kind of'key' to their identification is lacking, a preliminary classification of them is given, based on aerial photograph analysis. SUTMOLLER et al. (1963) give many data on the habitat and the floristic composition of several savannah areas, in parti cular lowland savannahs.

57 For the following short and provisional subdivision of the Amazon savannahs and savannah-forests, the classification of DUCKE and BLACK istake n as a basis, and the units of the forest inventories are tentatively fitted therein.

1.5.2.1Lowland Savannahs and Savannah-Forests Under this heading is grouped the grassy, shrubby, or low forest vegetation on lowlands with intermittent or permanent waterlogging. Implied are the compos de vdrzea of DUCKE and BLACK and the campo alagado, campo com arbusto alagado, arbusto alagado,floresta com arbusto alagado or 'grassland on waterlogged soils' of forest inventories.Thre e main subformations exist, namely:

/. Campo devdrzea dorio. Thi s is the grassy vegetation on river floodplains that are half-yearly covered with turbid river water (dgua branca; cf. 1.4.4). A very characteristic grass is the floating Canarana (Echinochloa polystachya,a.o. ) Shrubs and trees, especially on the highest parts (levees or point bars), are rather frequent. This type of vegetation is present in the Lower Amazon region.

2. Campo de vdrzea da chuva. This is the grassy vegetation on lowland that is half- yearly submerged with rain water. On such lowland, trees and shrubs are usually absent (campolimpo). This type of vegetation occupies for instance a large part of eastern Marajo island.

3. Campo de vdrzea do mar.Thi s isth e grassy vegetation on lowland that is regularly floodedwit h sea or brackish water. The vegetation type is associated with mangrove forest.

A special savannah-like lowland vegetation is still formed by the mondongos, which are lowlands in a part of eastern Maraj6 island that are permanently submerged, and therefore constitute a special type of igapo. Tall aquatic plants, like Aninga (Mont- richardia arborescens), abound here. On the coastal lowlands of Amapa Territory all mentioned subformations are apparently present.

1.5.2.2 Upland Savannahs andSavannah-Forests Under this heading is grouped the grassy or shrubby vegetation of non-flooded uplands oflo waltitude , ingenera l allth e Amazon land older than Holocene.Variou s subformations may be distinguished:

1. Campo. This is savannah that covers often extensive areas, but is encircled by forest. It iscommonl y located in watershed regions. Implied are the main parts of the campoan d campocom arbustoo r 'grass covered terra firme' of forest inventories. The greater part of the mapping unit 'upland savannahs and savannah-forests' of Fig. 12ar e of this type.

58 The principal plants are Gramineae. Shrubs and smalltree s may befrequen t(campo coberto), scattered (campo lavrado or camposujo), or absent (campo limpo). The trees and shrubs are stunted, have rough, thick rinds and xeromorphic leaves. Characteristic woody plants are for instance Li.xeira or Cajueiro docampo (CurateUa americana), and Muriel(Byrsomina crassifolia). Inpar t ofth ecampo sth e Mangabeira (Hancornia speeiosa) is frequent. Fires frequently sweep the terrains. In spite of the encircling byforest , thevegetatio n ofth e Amazon campos isno t related to that ofth e hileia, but resemblesth evegetatio n of the campo cerrado of Central Brazil, according to DUCKE and BLACK(1954) . The often used name campo firme is not quite adequate, because part of these savannahs are submerged by a shallow layer of rain water during the wet season (cf. IV.1.2) . Really lowpart s(botto m lands)withi n campo areas are often forested. A characteristic speciesi n the latter parts isth e palm-tree Buritl(Mauritia sp.). The extensive campos of Rio Branco Territory and the adjoining part of British Guiana are, in fact, outside the hileia region. Theirfloristic compositio n is different from thecampo s within the hileia (DUCKE and BLACK, 1954).N ofurthe r mention will be made.

2. Campina. This name applies to small savannah patches in the middle of forest, frequently with some swampy parts that are submerged by a shallow layer of rain water during the wet season. Patches and windings of white sand alternate with shrubs or small trees. Implied are some of the campo,an d campocom arbustoo r 'grass covered terra firme' of forest inventories. According to DUCKE and BLACK (1954),th efloristic compositio n of the campinas isdifferen t from that of thecampos . Gramineae are less numerous than Cyperaceae, and the species of the former are different from thoseo fth ecampos .Th ecampina shav eals oa larg ediversit yo fwoode n plants; these belong to the flora of the hileia and are related to that of the caatinga amazonica (seeunde r3) . The campinas are usually too small to be indicated on Fig. 12.Th e most numerous and typical campinas are located in the area between the lower courses of the Rio Negro and the Trombetas, where LECOINTE (1907)mappe d them.The y are especially frequent east of the Lake of Faro (which isth e widened mouth of the river Nhamun- da): Campos do Tigre, Campos de Maracana. Near Manaus several campinas were observed, occurring asirregula r stripsalon g rivulets.Example sfo r the Estuary region are Vigia, Colares, Gurupa and Porto de Moz. For the lower Tocantins area, DUCKE and BLACK (1954) mention the campina of Breu Branco and that of Arumateua near Tucurui. Also for the lower Madeira region campinas are reported: Campo Grande of Borba. The strip-likesavannah s along the Marmelosan d Manicore rivers(Fig .26 ) are probably also campinas. The nomenclature is not quite satisfactory, because sometimes the word 'campina' is used for indicating a campo of small extent. On the other hand, areas with real campina vegetation are locally called 'campo'o r even'Camp oGrande ' (cf. examples cited above).Th e situation isstil lmor ecomplicate d byth efac t that campo vegetation

59 and campina vegetation occur sometimes beside each other, for instance in the sa vannah area of Cupijo, west of Cameta, the savannah area of Mariapixi,wes t of the mouth of the Trombetas, and that of Ariramba (DUCKE and BLACK, 1954).

3. Caatinga amazonica.Thi s is a type of savannah-forest, which consists of low trees and shrubswit h interspersed high trees, or shrubsan d very lowtree so f approxi mately uniform height. The woody plants are practically all evergreen. Its flora definitely belongs to that of the hileia. According to DUCKE and BLACK (1954),th e caatinga amazonica isth eriches t in species of any vegetative formation of Amazonia.Thi ssavannah-fores t occursi nstrips ,an d isfoun d inth ecatchmen t areao f the upper and middle Rio Negro (tributaries Curicuriari, Icana and especially Uaupes)i n itsmos t typical form. Another regionwit h strips of caatinga amazonica is the Upper Amazon region near Sao Paulo de Olivenca.

4. Campina-rana. This is a type of savannah-forest which is isolated amidst the high forest, or occupies transitional areas between campo or campina and high forest (cf.Appendi x 7). Most of the savannah-forests of Amazonia fall into this grouping. Implied are the arbusto, as well as the floresta comarbusto or 'caatinga forest' c.q. 'savannah-forest' of forest inventories. AUBREVILLE(1958 )refer st oforets bassesetfourres amazoniens sur sables blancs. Both thecaating a amazonicaan dth ecampina-ran a areapparentl ygroupe d underthi slatte r heading.

1.6 Paleo Climate and Paleo Vegetation

Very littlei sknow n ofth epale oclimat ean d thepale ovegetatio n ofAmazoni a asyet , contrary to those of Africa (cf. D'HOORE, 1959). It is often accepted that the climate wasconstan t sinceth e Early Mesozoic,wit h onlymil dcooling sdurin gth e Pleistocene, and that also the type of vegetative cover remained almost unchanged. Great age of the forest coverage has been deduced for instance from the greatfloristic richnes s of the hileia. Recently however, doubts have arisen as to such constantness. In 1.4.1th e supposi tion of BARBOSA (1959) is already mentioned, namely that the sedimentation of the Tertiary Barreiras beds took place in a semi-arid climate of great extent. WILHELMY (1952) assumes that in the Late Tertiary a dry climate existed, during which extensive laterite crusts were formed form the Caribbean coast to 30°S , and open vegetation forms determined the landscape, including large parts of the area of the present day Amazon hileia. AB'SABER (1959) states that the presence of various terrace levels, conserved by hard laterite crusts, and the existence of island like campo in otherwise forest covered zones, presupposes that immediately before present times Amazonia had drier and inferior climates. They would be more com parable toth e Senegalesetha n toth eCongoles e climateso f today.

60 * It can be seen in IV.1. 2 that the island-like savannahs involved are edaphic pre destined, and in II.3tha tth eformatio n of lateriteo r plinthite isno t necessarily related to relativelydr yclimates .Bu t it is true that truncation and burying of such plinthite took place in times when no protecting vegetative cover existed, therefore during relatively dry climates. In agreement with the tentative dating of the Amazon fossil plinthite (cf.1.4.5 )thi s must havebee n duringa t least twoepoch s of the Late Tertiary (Miocenean d Pliocene),an d ona restricte d scaledurin ga par t ofth ePleistocene . The wholeconstructio n of theAmazo n valley proper in well distinguishable terraces of different levels and sedimentary materials, as described in 1.4.3, points also to distinct changes in climate and vegetation. Present-day deposition comprises only non-kaolinitic, mainly heavy textured sediments (cf. 1.4.4). This is quite reasonable considering the present day climatic conditions where almost all of the surrounding upland iscovere d with protecting tropical forest. Because of this coverage, erosion is verylimited .Onl yban k falls (terrascaidas) and afe wsand yarea swher eth evegetativ e cover has been artificially destroyed are involved. Byfa r the main part of the load of thelarg e riverscome sfro m far upstream, outside the forest belt. For the Amazon and the Madeira the main cource of sedimentary material isth e Andean mountain range (cf. SIOLI, 1951). In contrast, the sediments of the Pleistocenefluviatile terrace s are kaolinitic, often light textured, and sometimeseve ncontai n stony material (quartz pebbles,o r plinthite concretions as geologic stone-lines; cf. 1.4.5).I n viewo f the small grade of the rivers, the source of these sediments was certainly local. Principally involved was the basis material of the Amazon planalto: kaolinitic clays and quartz sands of the Barreiras beds, with one or more layers of fossil plinthite. The pronounced erosion of the Amazon planalto and the accompanying formation, with the erosion products, of well-defined floodplains (nowadays terraces) at lower levelssee m plausible only when the vegetative cover at that time was much sparser than the protecting cover of nowadays. The climatic conditions in the valley proper, during the deposition of the Pleistoceneterrac ematerials ,wer etherefor e probably of similar interpluvial character as that thought necessary for the building-up of the 'climatic' fluviatile Pleistocene terrace of the Araguaia Mahogany area (cf. 1.4.3).Thos e former climates must have had a long and pronounced dry season which limited the growth of vegetation, and a short but intense rainy season witha largeerosiona lforce . In viewo fth e conclusion in 1.4.3,namel y that thefluviatile terrac e deposits of thevalle yprope r werelai d down during interglacial times, the foregoing also supports the view that interglacials and interpluvials coincided in Amazonia. The tentative conclusion regarding the sedimentation of the Belterra clay in a huge inland sea during the Plio-Pleistocene (Calabrian) impliestha t all of the axial part of Amazonia wasdevoi d of any vegetation at that time.

In summarising, it isconclude d that during parts of the Tertiary and the Pleistocene the forest coverage was absent or restricted to relatively small areas. For instance, it must have been restricted to the higher parts of the Guiana and the Brazilian shields

61 during the Plio-PLeistocene at the time of the Belterra clay sedimentation. During theinterglacial s- interpluvialso fth e Pleistocene,rea lfores t vegetation wasprobabl y confined to thewester n part of Amazonia,nea r the Andes(all-yea r orographicrains) . This picture confirms the statements made by WILHELMY (1952), namely that the hileia only developed during the Pleistocene,fro m pockets of forest of Early Tertiary 1 age . Also very interesting in this respect are the conclusions of VAN DER HAMMEN (1957),tha t are based on analysis of fossil pollen of the Colombian Andes area. This author arrives at a distinct periodicity in vegetational and therefore also climatic changes during the Late Cretaceous and the Tertiary, for the N.W. part of South America.H eals onote stha tth ecentr eo fth e Guiana shield (theboundar y area ofth e Guinanas and Brazil) constitutes a centre of distribution of plant species, amongst which several genera of the palm family Mauritiaceae.

J) Quote:'Der Urwald, der heutedas Amazonasbecken erfiilli,hat sich erst im Pleislocdn auskleinen alltertidrenRestbestdnde entwickell,nachdem noch imjiingeren Tertiar offene Vegelations/ormationen im wesentlichen das Landschaftsbild Amazoniens beslimmlen. Aus seinemaquatorialen Riickzugsgebietheraus ist der Wa\dseit Beginnder Pluvialzeil allmdhlig nach S vorgestoszen, ein Prozesz,derjedoch aufden laterisier ten Biiden der Tertiaren Land- oberfldchebei den imPostpluvial wiederabnehmenden Niederschldgen nursehr langsahm und unvollkommen vonstattenging" (WILHELMY, 1952,p . 125).

62 II. The Main Amazon Latosols and Plinthitic Soils

II.1 Evolution of the Concepts of 'LatosoP and 'Laterite' or 'Plinthite'

The term 'laterite' was first used by BUCHANAN (1807), as descriptive of a highly ferruginous deposit, observed in Malabar (India). It hardens upon exposure and is therefore used as building material. Afterwards, apart from indicating building qualities, the word laterite became associated with red colour of soil material, either soft or hard. Almost every reddish coloured material on the earth's surface in the tropics and elsewhere was termed 'laterite*. The meaning of the term became more narrow after the work of BAUER (1898) on the chemical character of hardened material. The presence of free aluminum oxihydrates (Tonerde hydrat, gibbsite) was taken as the criterion. Generally speaking, such aluminum oxides were assumed to be present if the ratioSi0 2:Al203 (value Ki) of the quartz-free soil material was lowertha n 2. Laterauthor sassume d itt o begenera lcharacteristi c ofweatherin g inth e tropics that the ultimate products are iron and aluminum oxides, and that therefore even Si-Al products as kaolinite are not stable. HARRISON (1933) concluded from studies in British Guiana that on basic rock primarily gibbsite is formed ('primary laterite'), which can afterwards be resilificated into kaolinite, and that on acid rock kaolinite is formed directly. HARRASSOWITZ (1926, 1930), also adopted the criterion that 'allitic' constituents should be present to call a soil material lateritic. He could therefore speak of lateritischer (allitischer) Rotlehm,bu t was obliged to exclude iron oxide crusts that apparently had no free aluminum, for instance those formed in 'superficial' horizons of tropical savannah soil (Savanneneisensteine). HARRASSOWITZ took it as proven that lateritic crusts develop by evaporation of sesquioxide rich soil water and he made a scheme for the various 'laterite profiles' (cf.II.3.1) . The HARRASSOWITZ concept was attacked by MARBUT (1932), who based his con clusions onfindings i n Cuba and in the Amazon valley. The Cuba data concern the lateritischer Rotlehm, which HARRASSOWITZ had difficulty in placing in his scheme. TheCub asoil ,calle d'Nip eseries, 'ha ddevelope d onserpentin erock .I twa sconsidere d to be a true laterite in the sense that Ki values are below 2(0.2-0. 3 actually). It lacks however a sesquioxide crust, as well as a mottled layer and a grey zone, all of which HARRASSOWITZ considered to be essential constituents of the 'laterite profile'. In dications for the wandering of sesquioxides to the surface were not found. The intensity of the red colour in the upper part of the profile would be due merely to a greater degree of dehydration of the iron oxides than in lower horizons. MARBUT

63 concluded that normallaterite sdevelo p under conditions of good drainage, free from the influence of high ground water. Concentration of iron and aluminum oxides is mainly, ifno t entirely, mereresidua l(n o active upward movement),brough t about by the removal of silica, the alkalis and the alkaline earths. Sesquioxide segregation may or may not take place in the form of concretions. MARBUT found that alllaterit e profiles in the area of theunconsolidate d sediments of the Amazon valley have the HARRASSOWITZhorizons , but, below soilmaterial .Ac cordingt o hisobservation s the sequence only originates under the influence of fluctu ating ground water level at shallow depth, resulting in a specific, imperfectly drained soil. Sesquioxide crusts on dissected plateaux are fossil, and the relics of such a soil after erosion ofth esurfac e layer(fo r detailscf. H.3). MARBUTtherefor e distinguished: 1. 'laterite', as the lateritic red loam from Cuba; 2. 'ground waterlaterite' ,a sth eleache d soilswit h mottled,an d inpar t concretionary, subsoil of the Amazon valley.

Gradually, MARBUT'Sconcep ta st oth eorigi n oflateriti csoil san d lateriticcrust swa s accepted. It was fully adopted by the U.S. Soil Survey Staff. In 'Soils and Men' (BALDWIN, KELLOGG and THORP, 1938), a clear distinction is made between the intrazonal Great SoilGrou p ofth e 'Ground Water Laterite' and thezona l 'Yellowish Brown Lateritic', 'Reddish Brown Lateritic' and 'Laterite' Great Soil Groups. For the latter three Great Soil Groups, afterwards the term 'latosoF was preferred (Yellowish Brown Latosol, Reddish Brown Latosol and Red Latosol respectively, and someadditiona lgroups ; KELLOGG, 1949).Th e Latosolstherefor e should comprise 'all the zonal soils having their dominant characteristics associated with low silica/ sesquioxide ratios of thecla yfractions : lowbase-exchang e capacities,lo wactivitie so f the clay,lo wconten to f mostprimar y materials, lowconten t of soluble constituents,a high degree of aggregate stability, and (perhaps) somere d colour. KELLOGG proposed toconfin e thewor d 'laterite' to such ferruginous materialsa sharde n onexposure ,an d to the relics of such materials. This includes: 1. soft mottled clays that change irreversibly to hardpans or crusts when exposed; 2. cellular and mottled hardpans and crusts; 3. concretions; 4. consolidated concretions.

Such materials, especially, but not exclusively, in fossil form, may be found in several of the zonal soils, but their presence may be regarded as accidental at the categorical level of the Great Soil Groups. Only in the Ground Water Laterite soil, laterite is an essential feature. KELLOGG (1949)describe s the soila shavin ga'Gra y orgrayish-brow n surface layerove r leached 1 yellowishgra yA 2ove r thick reticulately mottled cemented hardpan at adept h of one

') But a cemented hardpan and concretions are apparently not essential. KELLOGG and DAVO L (1949) give Ground Water Laterite soil profiles of Central Africa, in which no mention ismad e ofa hardpa n character ofth eB horizon,an d ofwhic h theystat e 'may have concretions'. 64 foot or more.Hardpa n up to severalfee t thick. Laterite parent material. Concretions throughout'. Recently, namely in the VIIt h Approximation for a comprehensive system of soil classification, the U.S. SOIL SURVEY STAFF (1960) has introduced the terms 'Oxisol' and 'Plinthite'. Oxisol includes 'the soils that have been called Latosols, and many,i f not most ofthos etha t havebee n called Ground Water Laterite soils'.1 The diagnostic characteristic of an Oxisol is that it has an 'oxic' subsurface (B) horizon (cf. 11.2). The term 'plinthite' replaces 'laterite'. The clear distinction between lateritic soils or Latosols, and the material laterite or plinthite, as made by the U.S. Soil Survey Staff, is nowadays also adopted in other countries. PRESCOTTan d PENDLETON(1952 )relat e that in Australia the word 'laterite' is not used as a criterion in high-level soil classification, but retains its original geologicalmeanin g asa comple x parent material.Als o MOHR and VAN BAREN (1954), basing their conclusions on a large number of observations, mainly in Indonesia, agreewit h the distinctions asoriginall y made by MARBUT,althoug h theyobjec t to the term 'Latosol'. Recent French and Portuguese literature on classification of tropical soilsals otake sa clea rdistinctio n between plinthiteformatio n and Latosol formation. AUBERT (1954), for example, distinguishes cuirassement versusferralitisation.

II.2 Latosols

II.2.1 The Definition and Subdivision of Latosols as Applied in Brazil

The soil classification applied in Brazil by the national Soils Commission of the C.N.E.P.A.2 is based upon the U.S.A. system. The classification of red and yellow tropical soils in the U.S.A. system is not yet fully elaborated. The definition and subdivision of the 'Latosols' or 'Oxisols' order isstil lgreatl y debated, in part because ofthei r rarity insideU.S .territory . Detailed soilsurvey so fth e U.S.Soi lConservatio n Service in regions with tropical soils are those of Puerto Rico (ROBERTSet al.,1942 ; re-classification by BONNET, 1950) and of Hawaii (CLINE et al., published in 1955). For some other regions exploratory studies were made, for instance of Congo (KELLOGG and DAVOL, 1949). Soilswit hlatosoli cappearanc ear efrequen t inth evas t landmas s that constitutes the Federal Republic of Brazil. In recent years, many systematicfields studie s have been carriedout ,particularl y in the eastern and southern part of the country (cf.BARROS , DRUMOND, CAMARGOet al., 1958; LEMOS, BENNEMA, SANTOSet al., 1960).A detailed definition and classification of Latosols is being established in co-operation with, amongst others, the FAO Soil Survey and Fertility Branch. The latest trend in the U.S. system, as published in the VIIth Approximation (SOIL SURVEY STAFF, 1960)

') According to the concept of Ground Water Laterite soil as used in this thesis {cf. H.3.2.1), most of them belong to the 'Ultisols* of the Vllth Approximation. *)Centro National de Ensinoe Pesquisas Agronomicas.

65 has been considered, and the new definitions of diagnostic horizons (argillic horizon, albic epipedon etc.) and other distinctive characteristics are discussed and applied to some extent. But the new system is not taken as a basisb y the Brazilian Soils Com mission, one of the reasons being the above mentioned provisional character of the definition of Oxisols.

11.2.1.1 TheModal Latosol The first extensive description of the characteristics of the Brazilian Latosols as recognized byth eSoil sCommission , isgive n by BENNEMA, LEMOS and VETTORI(1959) . The elaborated concept, as presented at the 1SSS Congress at Madison U.S.A. by BENNEMA (1963), is fully discussed in the following: The fundamental characteristics of the Latosols lie in the nature and constitution of the mineral soil mass, indicating a thoroughly weathered soil. It consists of sesquioxides,silicat eclay-mineral shavin ga 1:1 lattice,quart zan d other mineralstha t are highly resistant to weathering. Primary silicate minerals with less resistance to weathering are either absent or are present in only a small amount. The same applies to clay minerals havinga 2: 1 lattice,an d those amorphous gelso fS ian d Altha t have high base exchange capacities. Free aluminum oxides are often present, but not always. Concretions of iron, manganese, aluminum or titanium oxides may be present. The silt content of the samples in the solum isgenerall y low. To this constitution can be ascribed a large number of the characteristics and pro perties of the Latosols. The most salient of them are the following:

1. Indistinct horizon differentiation, with often diffuse or gradual transition between the horizons (except when an Ap is present). 2. Absence, or scarcity, of distinct silicate clay skins on peds or distinct silicate clay linings in the channels. 3. Low cation exchange capacities of the clays. 4. Red, yellow or brown colours of the subsoil1(B ) horizon or part of the subsoil horizon. 5. Absence, or near absence, of electro-negative 'natural clay'2 in parts of the subsoil horizon that have a low percentage of Carbon (C/clay ratio less than 0.015). Additionally, the typical Latosols have: 6. Absence of well-developed blocky or prismatic structure. The structural elements are often veryfine granules ,whic h may bemor e or lesscoherent , forming together a porous, friable, massive soil mass. 7. Deep solum (A + B horizon). 8. Very friable or friable consistency when moist. 9. High porosity and high permeability.

*)Subsoil considere d in the sense as described in the supplement to the U.S. Agricultural Handbook No. 18, entitled 'Identification and Nomenclature of Soil Horizons' (1962). •) Natural clay = the clay obtained by shaking with distilled water, being a measure of aggregate stability.

66 Foto5 O principal Lalosolo amazonico. O perfil de urnLalosolo Amarelo Caolinilico (,Orto), de textura media, com a sua transicaocarateristicamente difusa entre os horizontes A e B, e a macro-eslrulura pouco pronunciadana seccao sob-superficial

+ •--'•

*•'-.-•.•*•• < " w

• i II • • _ •" : * r >1 * * Photo5 The main Amazon Lalosol. The profile of a Kaolinilic Yellow Lalosol (,Ortho), of medium - \i. -. ^ • *i • ^N texture, with its characteristic diffuse transition between the A and the B horizons, and the little pronounced macro-structure in the sub-superficial section \ >• -.*.

10. Low base saturation in the whole profile, or, at least, in the subsoil. 11. Relatively high anion exchange capacity and high phosphorusfixing power . 12. Relativelylo wamount so fexchangeabl ealuminum ,o r lowactiv e cation exchange capacity. 13. High resistance to gully erosion.

Soft or hard plinthite may be present in the lower part of the solum, or below the solum, but it is not an essential characteristic of a Latosol(cf. II.1) .

II.2.1.2 The Distinction of theLatosols from otherSoils; Latosolic-B versus Textural-B Besidesth e Latosols,th e other main welldraine d red and yellow soilso f tropics and subtropics are the Red Yellow Podzolic soils (including Rubrizems), the Andosols,. the RedYello wMediterranea n soils,an d the Reddish Prairie soils.O fthem ,especiall y the Red Yellow Podzolic soils are sometimes hard to separate in the field from.

67 Latosols (for thedefinitio n ofth emoda l Red YellowPodzoli csoil scf. III.2) . Latosols and Red Yellow Podzolic soils often occur in close proximity in Brazil. Transitional soilsbetwee nth emoda ltype so fbot h arefrequent . Both morphological and analytical studies are then required to decide whether the soil concerned is a Latosol, intergrading to Red Yellow Podzolic soil, or a Red Yellow Podzolic soil, intergrading to Latosol. The Brazilian SoilsCommissio n haspai d muchattentio n tofinding generall y valid quantitative boundaries between predominant latosolic and predominant podzolic character. This is described in the following:

1. Textural differentiation. In the light textured soils (frequent throughout Brazil) the presence of silicate clay skins and linings cannot always be used as a differentiating characteristic. Peds are often absent, and consequently also the clay skins even when the illuviation process is well advanced. Linings in channels also are not easily formed, due to the small amounts of clay present and the relatively short life of the channels. Also, many morphometrical characteristics of typical latosolic soils lose their diagnostic importance because they are also characteristics of many sandy podzolic soils. In this case the horizon differentiation is often the only reliable field characteristic for the separation. The changes in texture within the profile are of special importance.Th e subsoilhorizo n ofman y podzolic soilsi sclearl y distinguished from the topsoil due to a rapid or rather rapid change in texture. A number of the Latosols, namely those with relatively small percentages of sesquioxides, have also a change in texture in the profile (see below), especially the sandy ones, but in this case the change is gradual. If no other differentiating characteristics occur, for instance the presence of smallamount s of2: 1lattic e silicate clay minerals,i t hasbee n 1 proposed (BENNEMA, 1963) that the maximum texturalgradient for sandy Latosols should be asfollows : an increase of 70% over 20c m of the initial clay content which should be at least 5%. However, few data about clay gradients exists as yet because the available soil data refer predominantly to the mean of each different horizon, and not to a change over a certain distance. Therefore an approximative value is used, namely the textural ratio2 B/A. The value should not exceed 1.8 for the sandy Latosols. This ratio has to be based on analytical data. In the field, therefore, ex perience in feeling textural differences has to be relied upon. In the very heavy textured soils, an illuviation process does not lead so easily to a distinct horizon differentiation as expressable in a textural ratio B/A (cf. BENNE MA, 1963). Heavy clay soils, however, when subject to an illuvation process, show distinct clay skins.Therefore , absence ofcla y skins isth e most important morphome trical field characteristic for separation of the heavy textured Latosols from the heavy textured Podzolic soils.

') Texturalgradient = the difference in texture, especially in clay content, over a certain distance. *)Textural ratio B/A = the arithmetic mean of the clay content of the subdivisions of the B horizon except the B,, divided by the arithmetic mean of the clay content of the subdivisions of the A horizon.

68 For the very sandy red and yellow soils, a classification as Latosol or Red Yellow Podzolic soil is thought to be inadequate. In this case, the latosolic or podzolic character is very difficult to assess, and has moreover little sense. The presence of inactive quartz sand becomes by far the predominant characteristic, instead of the constitution ofth ecolloida l part.Th e structure, ifpresen t at all,i softe n weak,th esoi l massconsistin go fbridge so fsan d grainsonly .Als othos esoil so fwhic hth ever ysmal l colloidal part is iatosolic' may therefore be quite susceptible to (gully) erosion, resistance to which is one of the practical properties of Latosols. In Brazil, for such very sandy acid soilsth e nameAcid Red Yellow (Quartz) Sandi spreferred , instead of Latosol or Red Yellow Podzolic soil. For the upper textural limit of these Acid Red Yellow Sands 15%cla y sized particles in the subsoil1 horizon is taken: light sandy loam and lighter textures (cf. LEMOS, BENNEMA, SANTOSet al., 1960).Whe n soils with less than 15%cla y still clearly exhibit important 'latosolic' or 'podzolic' characteris tics,the yma yb eclassified ,withi nth edescribe d concept,a sLatosolic Sand (or Latosoli c Regosol),an dPodzolic Sandrespectively .Th eforme r name isapplie d for apar t ofth e Amazon soils (see below).

2. Mineralogic composition of the clayfraction. The absence, or near absence, of 2:1lattic esilicat ecla ymineral sca nb echecke d byX-ra y and DTAanalysis , preferably combined.This ,however ,i selaborat ean dexpensive .Dat ao nth echemica lcompositio n of the clay fraction are therefore often used as an index of the presence or absence of 2:1 lattice silicate clay minerals. The molecular ratio of the chemically determined silicium and aluminum oxides(Si0 2:Al203 = Ki)i softe n used for this purpose. The criterion for latosoliccharacte r isthe n arati obelo w2 .Objection s tothi sar e made, for instance, that a value above 2 does not necessarily mean the presence of 2:1 silicate clay minerals, but may be due to presence of uncombined, colloidal, silica (cf. MOHR and VAN BAREN, 1954). But many soil scientists working in tropical regions have found the ratio to be a reliable characterisation tool.

3. Cation exchange capacity. An additional, and often more convenient, gauge for determining the composition of the mineral soil mass are the cation exchange capaci ties of this mass, especially in view of the possible presence of amorphous gels of Si and Al. A part of the latter, and 2:1 lattice silicate clay minerals, have higher cation exchangecapacitie s than thosecharacteristi c for the latosolic soils.I n usingth e cation exchange capacities as a scale, a correction has to be made for the organic matter present which often accounts for the bulk of the determined cation exchange capacities. Also, it has to be clearlykep t in mindwhic h of the cationexchang e capa cities is being determined. The typical Latosols so far found in Brazil have a cation exchange capacity according to the NH4OAc method (pH=7) varying from 0 to 8 m.e./lOOg clay , while the ones that are intergrading to other soils have values up to 10-12 m.e./lOO gcla y (cf. BENNEMA, 1963).

') Seenot ea t page66 .

69 Table5 Principalcharacteristics of latosolicand textural B horizons(after BENNEMA, LEMOSan d VETTOIU, 1959 and LEMOS, BENNEMA, SANTOS et al., I960)

TEXTURAL-B/fl-fexrura/ LATOSOLIC-B /B-latosolico Normally very distinct contrast with the other Weak contrast with the other horizons. The horizons. Thetransition s are either abrupt, clear transitions arenormall y diffuse or gradual or gradual O contrasto com os oulros horizontes normalmente O Lontrasto comos outros horizontes i fraco. As 4 multo dist'mto. Astransicbes sao abrupta, clara transicoes saonormalmente difusasou graduais ougradual

At least 15% clay-sized mineral particles, and At least 15%clay-size d mineral particles more clayey than the A horizon Ao menos15% de fracdo argila,e mais argilosa Ao menos 15 % de fracdoargila que o horizonie A

When thehorizo n isheav y textured (clay, sandy When thehorizo n isheav y textured (clay, sandy clay,cla y loam), then the structure isstrongl yo r clay), then the structure is generally of fine or moderately subangular and angular blocky, or very fine granules, forming a porous homo prismatic,wit hwel lo rmoderatel ywel ldeveloped , geneous masswit h very weak coherence. Ftals o often continuous clay skins, and relatively low may have a weakly or moderately developed porosity subangular and angular blocky structure, the blockyelement s beingcompose d offine granules . Clay skins, when present, are non-continuous and of weak development. The porosity is generally high Se o horizonie4 de texturapesada (clay, sandy Se o horizonie4 de textura pesada (clay, sandy clay, clay loam), a estrutura 4 em blocossub- clay), a estrutura 4, geralmente, em granulos angularese angulares, ou prisma!ica,fortemente a pequenos ou muito pequenos, formando uma moderadamentedesenvolvida. Acerosidade 4forte massa porosa homogeneacom coerenciamuito oumoderada, muitas vezes continua. A porosidade fraca. A estrutura apresenta-se tambemem blocos e relativamente baixa angulares esubangulares,fracamente a moderada mentedesenvolvida, sendoos blocoscompost os de granulosfinos. Cerosidade, se 4presente, 4fraca e nao continua. A porosidade 4geralmente elevada

When the horizon is medium textured (sandy When the horizon is medium textured (sandy clay loam, loam, heavy sandy loam), then the clay loam, heavy sandy loam), then the structure structure is weakly or moderately subangular is fine or very fine granular associated with and angular blocky, with rather well developed single grains, forming a porous mass with very clay skins. Sometimes thestructur e iswea k fine weak coherence. Thetextura l ratio B/A islowe r granular associated with single grains, forming a than 1.8 homogeneous porous soil mass with little coherence. The textural ratio B/A is above 1.6 Se o horizonie4 de textura media (sandy clay Se o horizonie4 de textura media (sandy clay loam, loam, heavy sandy loam), aestrutura 4 em loam, heavy sandy loam), a estrutura i muito blocos subangulares eangulares, fracamente a mo pequenagranular associada comgraos simples, deradamentedesenvolvida, sendo a cerosidade fraca formando umamasssa porosa homogenea com a moderada. As vezes,a estrutura 4granular pe- coerencia muitofraca. A relacaotextural B/A i quena, fracamente desenvolvida e associada com em baixode 1.8 graos simples,formando uma massaporosa homo genea compequena coerencia. A relacao textnral B/A 4 superior a 1.6

70 Table5 continued I Tabela 5 continuada

TEXTVRAL-B/B-textural LATOSOLIC-B/B-latosolico The consistence when moist isfirm o r friable The consistence when moist is friable or very friable A consistencia, quando umido,4 firme ou fridvel A consistencia, quando umido, 4fridvel ou muito fridvel

The 'natural clay'conten t can berelativel y high. The 'natural clay' content is normally low. It is less than 1 % in the B4,excep t when theK i value is very low and the pH-KCI is higher than the pH-H20, or the carbon-clay ratio is relatively high(C/clay > 0.015) O conteudo de 'argila natural'muitas vezes 4 O conteudo de 'argila natural" 4 normalmente relativamente alto baixo. No subhorizonte Bx 4 menor de 1%, com excessao quando o Ki4 muito baixoe opH-KCI 6 mais alto que o pH-HtO, oua relacao carbono- argila 4relativamente alia (Cjargila >0.015)

The Ki value (SiO,/AljO, molecular ratio) is The Ki value (SiOj/AljO, molecular ratio) is normally above 1.8, less commonly between 1.8 normally below 1.8, less commonly between 1.8 and 1.6 and 2.0. O valor Ki (relacao molecular SiOjAl,Ot) i O valor Ki (relacao molecular SiOjAltOt) i normalmentesuperior a 1.8, raramente entre normalmente maisbaixo que 1.8, raramenteentre 1.8 e 1.6. 1.8 e 2.0.

Thecatio nexchang ecapacit y isofte n largertha n The cation exchange capacity is small. It is in the latosolic-B below 12 m.e./lOO g of clay (NH4OAc method) A capacidade de permuta de cations 4 muitas Acapacidade depermuta de cations 4pequena. t. vezes maiorque no B-latosolico maisbaixo que 12 m.e.{!00g de argila (me'todo de NHtOAc)

Easily weatherable primary minerals may be Weatherable primary minerals are practically present absent.The ycompris e less than4 % of the sand fraction Podemestar presentes minerals primdrios pouco Minerals primdrios sujeitos ao intemperismo resistentes ao intemperismo saoprdticamente ausente. Compreendem em conjunto menos de 4% dafracao deareia

The silt content is often higher than in the The silt content is generally low. The silt/clay latosolic-B ratio is normally below 0.25 O conteudo de silte 4 muitasvezes mais elevado O conteudo desilte 4 geralmente baixo.A relacao queno B-latosolico desiltelargila 4normalmente maisbaixo que 0.25

Tabela5 Carateristicas principals de horizontes B-latosolico e B-textural (segundo BENNEMA, LEMOS and VETTORI, 1959 e LEMOS, BENNEMA, SANTOS et al., 1960)

71 4. Weatherable minerals. It is suggested (BENNEMA, 1963) that the amount of minerals less resistant to weathering account for up to 4% of the sand fractions. The amount of minerals with small resistance to weathering should, however, remain considerably below this limit.

5. Silt content. As stated by VAN WAMBEKE (1962), the silt content of tropical soils appears to be closely related to the actual amount of weatherable minerals. Silt/clay ratios therefore are a helpful guide in distinguishing Latosols from less weathered tropical soils. In Brazil, a ratio of maximally 0.25 for typical Latosols is proposed, in agreement with D'HOORE (1960). BENNEMA (1963) however mentions that in some Latosols, for instance the Terra Roxa Legitima (see below), secondary minerals such as kaolinite and iron oxides form part of the silt fraction, as tiny concretions. In such soils the ratio may be higher than 0.25.I n handling this ratio as a distinguishing device it should also be remembered that inaccuracies in determination of the mechanical composition are frequent because of difficult soil dispersion. This is especially so in early publications. Often a large portion of that which wasdetermine d as silt is actually non-dispersed clay (cf. SOIL SURVEY STAFF, 1960, p. 53-54).

The set of characteristics inherent to a latosolic-B (as the diagnostic horizon for Latosols), in comparison with that of a textural-B (as the diagnostic horizon of Red Yellow Podzolic soil and others), is summarised in Table 5. The described concepts on latosolic-B and textural-B are not quite identical with the 'oxic horizon' and the 'argillic horizon' respectively of the Vllth Approximation (SOIL SURVEY STAFF, 1960). Latosolic-B and textural-B are mutually exclusive, which is not necessarily the case with the oxic and the argillic horizons: an oxic horizon can conceivably include an argillic horizon. With the applied maximum value for the latosolic-B of the textural ratio B/A, a part of the group of latosolic-B horizons falls under the argillic horizon. The lower textural limit (more than 15% clay) of latosolic-B and textural-B is the same as for the oxic horizon, but the argillic horizon can be lighter textured. An important difference between latosolic-B and oxic horizon is that the latter should have 12% or more free sesquioxides in percentage of the 1:1 lattice silicate clay minerals. Acceptance of the latter criterion would imply that some Latosols, for instance most of the Amazon ones (see below), fall outside the Oxisol Order. Also, the maximum percentage of weatherable primary minerals is smaller in the oxic horizon than in the latosolic-B, namely 1 %.

II.2.1.3 The Subdivision of the Latosols INTRODUCTION In 'Soils and Men' only three types of Latosol were distinguished (BALDWIN, KELLOGG and THORP, 1938): Yellowish Brown Latosols, Reddish Brown Latosols, and Red Latosol. After the reconnaissance of the Congo by KELLOGG and DAVOL (1949), this was expanded to: Red Latosol, Earthy Red LatosoL Reddish Brown Latosol, Black Red Latosol, Red Yellow Latosol, and Yellow Latosol. The distinc-

72 tionswer eprincipall y made on the basiso fcolour , structure and consistence. BONNET (1950) subdivided the Latosols of Puerto Rico as follows: Coastal Plain Latosol, Upland Latosol, Ferruginous Latosol, and Rain Forest Latosol. CLINE (1955) discerned for Hawai Low Humic Latosol, Humic Latosol, Ferruginous Humic Latosol, and Hydrol Humic Latosol. In that publication also the composition of thecla yfractio n wastake n into account in differentiating. BRAMAOan d DUDAL(1958 ) mentioned Red Yellow Latosol, Dark Red and Dark Reddish Brown Latosol, Brown Latosol, and Low Humic Latosol(o r Terra Roxa).I n presentingth efirst draf t of the 'Soil map of South America', BRAMAO and LEMOS(1960 ) mention, besides theabov e mentioned, also Rego-latosols, Pale Yellow Latosols, Concretionary Latosols, and Areno-Latosols. LEMOS, BENNEMA, SANTOS et al. (1960) report for Sao Paulo the following Latosols: Terra Roxa Legitima, Dark Red Latosol (ortho, and sandy phase),Re dYello wLatoso l(ortho ,sandy ,terrace ,shallo wphase) ,Humi c RedYello w Latosol, and 'Solos de Campos de Jordao'. The differentiating characteristics, also qua composition of the clay fraction, are described in extenso, and a comparison is made with earlier discerned units (see above). Allthi s isa stillrathe r confused picture ofth evariatio n within the Great Soil Group ofth e Latosols.Th ewhol eo fth erapidl y growingmas so ffield an d laboratory data on latosolic soils isbein g sorted out, in order to arrive at clearly defined, mappable units for all tropical regions. It forms part of the FAO 'Soil map of the World Program'. Division accordingt ocolou ralon ei sapparentl y not veryvalid , because this depends often more upon the stage of dehydration of the iron oxides than upon the total amount of Fe.Also ,man ymorphometrica lfield characteristic s varylittle ;al lLatosol s are deep, friable, porous soilswit h indistinct horizon differentiation. For subdivision, analytical data on the composition of the clay fraction have to be taken into account. This composition, in relative amounts of silicate clay minerals (kaolinite), Fe clay minerals (goethite), and Al clay minerals (gibbsite), is of much importance. It determines the stability of the structure, the natural fertility, and the effect of fertilizing. In Brazil, where a relatively simple and reliable method for characterising this composition is included in the standard analysis of all soil samples, a useful subdivision of Latosols mainly on the composition of the clay fraction has been developed. After a preliminary subdivision as mentioned by BENNEMA(1963) ,i t was elaborated at the Third Technical Meeting of the Brazilian Soils Commission (Rio de Janeiro, Dec. 1961) and presented by CAMARGOan d BENNEMA(1962) .Fo rth etentativ e scheme see Table 6.

DESCRIPTION OF THE SUBDIVISION OF TABLE 6 Ad I. To this group, the Acrox of the Vllth Approximation, and representing the most advanced'laterisation' ,belon gfo r instanceth e Nipeserie so f Cuba(cf. II.1) ,an d many of the soils around Brasilia (cf. BENNEMA, CAMARGO and WRIGHT, 1962).Th e soils are usually found on old land surfaces with flat or gently undulating relief. In thatcas eth eparen t materiali svarying .Th ecolou rals ovaries .Th emorpholog y ofth e

73 Table6 Tentative subdivision of theLatosols (after CAMARGO and BENNEMA, 1962J Molecular ratioso fth ecla y fraction relacoesmoleculares dafracao argila

SiOa/AljO, Al20,/Fe20, Other Groups and subgroups of the Latosols soil texture oxides grupose subgrupos dosLatosolos texturado solo outros medium heavy dxidos me'dia pesada (I) LATOSOLS WITH LOW PERCENTAGES OF SILICATE <1.0 CLAYMINERAL S (Acrox) Latosolos com baixaspercentagens de minerais-de- argilade silicato (Acrox) (Still tob esubdivided , on AI,Oj/Fe20, ratio and other characteristics/a ser subdividido, na relacao AltOJFe^ eoutras carateristkas) (II) LATOSOLS WITH INTERMEDIATE PERCENTAGES 1.0-1.6 OF SILICATE CLAY MINERALS(Norma l Latosols) Latosolos compercentagens intermedidriasde minerais-de-argila desilicato (Latosolos Normais) (Ha) with relatively high percentageso fF ecla y <1.7 MnO, minerals and relatively high percentages ofM n >0.10% and Ti(Latoso l Roxo) TiO, compercentagens relativamente altas de mine 4-8% rais-de-argilade Fe, e percentagensrelativa mentealtas de Mn e Ti (Latosolo Roxo) (lib) with intermediate percentages ofF e clay 2.0(?)-2.6 2.0-4.5 MnO, minerals (Dark Red Latosol) <0.02% com percentagens intermedidrias deminerais- de-argila deFe (Latosolo VermelhoEscuro) (He) with relatively lowpercentage so fF ecla y 2.9-5.5 4.6-8.0 minerals (Red Yellow Latosol) com percentagens relativamente baixas de minerais-de-argila deFe (Latosolo Vermelho Amarelo)

(III) LATOSOLS WITH HIGH PERCENTAGES OF 1.6-2.0ca. SILICATE CLAY MINERALS (Kaolinitic Latosols) Latosolos com altas percentagens de minerais-deargila de silicato(Latosolos Caoliniticos) (Ilia) with relatively high percentages of Fe 2.0 ca. 3.0 ca. TiO, clay minerals, andrelativel y high percentages 3-4% ofTi com percentagens relativamentealtas demine rais-de-argilade Fe, e percentagens relativa mentealtas de Ti (Mb) with intermediate percentages ofF ecla y 4.0 ca. minerals com percentagensintermedidrias de minerais-deargila de Fe (Illc) with relativelylo wpercentage s of Fecla y >4.6 minerals com percentagens relativamente baixas de minerais-de-argila de Fe Tabela6 Subdivisao tentativa dos Latosolos (segundo CAMARGO andBENNEMA , 1962)

74 profile ofthes e soilsi sabou t the same astha t of other Latosols.The y havehoweve ra very'earthy ' feeling (i.e.th eaggregate sfee lraw) .A reliabl edifferentiatin g characteris tic is the fact that the pH-KCl is higher than the pH-H20 if the organic matter content islow , i.e.i n the subsoil (B)horizons . Unlike other Latosols,th e typical ones of this group also have 'natural clay' in the B horizon (excess of electro-positive charges). The soils have an extremely low cation exchange capacity and effective fertilization is difficult.

AdHa. Tothi s subgroup belongth eTerra RoxaLegitima (also called Latosol Roxo) of Sao Paulo (LEMOS, BENNEMA, SANTOS et al., 1960), and probably part of the Ferruginous Humic Latosols of CLINE (1955). They develop on basalt and diabase. The crushed dry soil sample is magnetic. The texture isclayey . The usual colours of the Bhorizo n are reddish with low values and low chromas (<4), for instance dark reddish brown (2.5 YR 3/4) or dusky red (10R 3/4). The structure is fine granular (pdde cafe) and the solum deep. The soils have often relatively high natural fertility (coffee soils), but fertilization presents problems.

Ad lib. To this subgroup belong many of the Dark RedLatosols and Dark Reddish Brown Latosols1. The soils are found on igneous rocks and consolidated sediments with considerable quantities of ferro-magnesium minerals. The crushed dry soil sample isonl y slightly magnetic.Th e texture isvarying .Th e common colour of the B horizon isreddish , with low values (<3.5) but high chromas (5-7), for instance dark red (10R 3/6,2. 5Y R 3/6).Th e natural fertility isofte n somewhat highertha n thoseo f lie.

AdHe. T othi ssubgrou pbelon gman yo fth e Red YellowLatosols (BARROS, DRUMOND, CAMARGO et al., 1958; LEMOS, BENNEMA, SANTOS et al., 1960). They develop on igneous rock and consolidated sediments with only fair amounts of ferro-magnesium minerals. The texture varies, and there may be a slight difference in texture between theA an d the Bhorizons .Th ecommo n colouro fth e Bhorizo ni sreddis ho ryellowish , with high value (>3.5) and high chroma (6-8), for instance red (2.5 YR 5/8),yello wish red (5 YR 5/8) or strong brown (7.5 YR 5/8).

Ad Hie. To this subgroup belong the so-called Rego-Latosols;severa l Yellow Latosols;man y of the RegosolicYello w Latosolic Podzolicsoil swhic h are alsocalle d 'Rego-latosol Amarelo, fase Tabuleiro' or 'Tabuleiro' (BARROS, DRUMOND,CAMARG O etal., 1958);th e Red Yellow Latosol, terrace phase (LEMOS, BENNEMA, SANTOS et al., 1960).Thes e Latosolsar efoun d onrelativel yyoun glan d surfaces with unconsolidated sediments. Gibbsite is absent or practically absent, and Fe203 comprises often less than 10% weight of the clay fraction. The texture varies and the Bhorizo n is often

') The concept of these soils asuse d in Brazil is more narrow than the current concepts of Reddish Brown Lateritic soils, and the Dark Red and Dark Reddish Brown Latosols of BRAMAO and DUDAL (1957).

75 distinctly heavier than the A horizon. The common colour is yellowish, with high chroma and high value. The colour is often paler in regions without a dry season (10Y R or 7.5Y R 6/6, 6/7, 7/6) than in regions with a distinct dry period (7.5 YR 8/6 or 10Y R 5/6).Th estructur e isofte n weakly subangular blocky,th e porosity is lower than in most other Latosols. When dry then the profile ishar d or slightly hard in the transition zone between the Aan d the B horizon (A2, Bt). Plinthite is fairly common. The natural fertility is normally low, except in transition areas to arid regions. The response to fertilization is often favourable.

A further subdivision will bemad e both according toth ethicknes so f theA l horizon (weak, intermediate, pronounced1; the latter known to exist for units lib and lie, Ilia and Ilk), and the base saturation (low, medium, high; the latter known to exist for unitsI la, Hlban d IIIc). To augment the practical application of soilsurveys , also 'phases' are used which are based often on variations of mesologic environmental factors (geomorphology, vegetative cover). Inadditio n tothes e groups, there are Latosols, at relatively high altitudes, of brow nish colour and relatively thin (

II.2.2 The Main Amazon Latosol

The most prominent geomorphologic units of Amazonia are tne uplands, arranged as terraces, of the Late Tertiary and the Pleistocene which together are termed the Amazon Planicie (cf. 1.4.2 and 1.4.3).Thes e uplands are composed of unconsolidated acid sediments of varying texture.The y consist very predominantly of kaoliniticclay s and quartz sands, and locally of fossil plinthite. The well-drained soils developed on these sediments are for the main part very similar, considering the high categorical level of classification as isuse d in this publication. This similarity applies to both the morphometrical field characteristics and the analytical data. General occurring morphometricalfield characteristic s are the following (for individual profiledescrip tions cf. Chapter III; cf. also Photos 6, 7an d 8). 1. A deep, permeable solum. 2. Little horizon differentiation, with diffuse or gradual transitions between the horizons. 3. The texture is varying, from light to very heavy textured. The silt content is very low, especially in the subsoil2 (B) horizon. 4. The structure is usually weak very fine granular, composing a slightly coherent porous massive soil mass, which easily falls apart into weak or moderate fine or

') Percentage of Carbon at 20 cm depth larger than 0.3 + 0.057 x moisture equivalent. *)Se e note on page 66.

76 Foto 6 Oprincipal Laiosolo amazonko. Umavista de perto da estrutura da seccdosob-superficial do perfil do LaiosoloAmarelo Caolinilico (,Orlo), de lextura meio-pesada. Os elemenlos esldveis muito finosda micro-estrutura constituent uma massa de soloporoso pouco coerente, massa que se desagrega emblocos sub-angularesfracos a moderados (paraa escala veja opalito; fotografia Dr. J. Bennema)

Photo 6 Themain Amazon Latosol. A near viewof the structureof the subsuperficial sectionof the profile ofa Kaolinitic YellowLatosol (,Ortho),of ratherheavy texture. The veryfine, stable elements of the micro-structureform a weakly coherent poroussoil mass thatfalls apart into weak tomoderate subangular blocks(cf. theupright matchfor thescale; Photo by Dr.J. Bennema) medium sized subangular blocky elements (cf. Photo 6).N o silicate clay skins or clay linings are present, except for a few faint ones in the very heavytexture d soils.Excep t for the very light textured ones, the soils are little subject to gully erosion. 5. The consistence when moist is usually friable to very friable (loose in the lightest textured soils);whe n dry it isofte n slightly hard, to hard, especially in the upper part of the Bhorizon . In a number of the soils,th e subsoil horizon isdifficul t to penetrate with a soil hammer. 6. Theporosit y isgenerall y high,bu t apparently notquit e as high as in most Latosols of Southern Brazil. In a number of the soilsth e subsoil horizon iscompac t or rather compact. 7. The soils show practically always some degree of illuviation (lessivage or 'podzolisation').Th e subsoil horizon isnamel y slightly heavier intextur e than the surface and subsurface horizons: the textural ratios B/A are between 1.2 and 2.1 for relatively light textured profiles, and between 1.1 and 1.6 for relatively heavy textured profiles.

77 Foto 7 O principal Latosolo amazonico. A parte superior(50 cm m. ou m.) do perfil de um Latosolo Amarelo Caolinitico (,Orto), de textura muitopesada. £ notdvel o horizonteA x muito delgado (3 cm m. ou m.)

Photo 7 The main Amazon Latosol. The upper part (50 cm ca.) of the profile of a Kaolinitic Yellow Latosol (,Ortho), of very heavy texture. Note worthy isthe verythin A x horizon (3 cm ca.)

8. The colour of the subsoil horizon is mostly yellowish, lesscommonl y reddish, but has alwayshig h value and high chroma. The reddishcolou r may befoun d when fossil hard plinthite is present, or in transition zones to arid regions. 9. Plinthite, hard or soft, is rather commonly present.

The following are the analytical data: 10. The content of 'natural clay' is low; in the subsoil horizon it isver y often com pletely absent: for the A horizons the indices of structure are between 40 and 80, and for the Bhorizon s they are normally 100. 11. The soils are extremely or very strongly acid, with the pH-HaO between 4an d 5. The pH-KCl islowe r than the pH-H20, but normally lesstha n one unit. 12. The base saturation of the soils under natural vegetative cover is very low {ca. 15%) . Onlyi ntransitiona l areast oari d regionsi t mayb elo wt o medium. 13. Easily weatherable primary minerals are practically absent.Th e great majority of the non-clay sized particles consists of quartz grains. The other ones are vegetal detritus elements, some reddish coloured micro-concretions of plinthite, and cream

78 Foto8 Operfil de umAreia LatosolicaCaoli- nitica. E notdvel o horizonteA x relativamente espesso e afunda penetracao das raizes

Photo8 The profile of a KaolinilicLatosolic Sand. Noteworthy is the relatively thick A horizonand the deep rooting coloured pseudo micro-concretions of clay. Turmaline, staurolite, ilmenite and weathering biotite and felspar, when present, form together only a trace.1 14. The potential cation exchange capacity of the soils (NH4OAc method at pH7) variesfro m nearzer ot oabou t 20m.e./lO Og soi li nfores t profiles. However,th ecatio n exchange capacity of the organic matter-free soil material is always very low. Cal culated on thebasi so fth ecla ysize d particles it onlyamount s to 1.5-5 m.e./lOOg cla y (data obtained byextrapolatio n ofrelevan t analyticaldat a ofindividua l profiles given in Chapter III; see also Table 7 (other method of analysis) and V.3.1.1). The active cation exchange capacity, or the exchangeable (Al)+, of the subsurface and subsoil horizons is on the average only about 20-25% of the potential cation exchange capacity. 15. Themineralogica lcompositio n of thecla yfractio n isver y uniform. The molecular ratios between Si02, A1203 and Fe203 are very constant. For about fourty analysed

*)Onl y in the separate 2-16f i felspar may be present in higher quantities (cf.sample s 233-3, 300-4 and 210-4 of Fig. 14a- e and Table 8).

79 profiles scattered over the Planicie, the Si02: A1203 values (Ki) are, with rare ex ceptions, between 1.7 and 2.1,throughou t the profiles. Si02:(Al203 -f Fe203)value s (Kr)ar eonl yslightl y lower - usually 0.2 unit -, and A^Og:Fe 203 values correspond ingly high, namely about 9.0. Fe203 comprises 5-11 percent of the clay fraction. It is only in concretionary soils and very light textured soils that this percentage may become as high as20 . From the fact that Ki valuesar eno t above 2,i t mayb ededuce d that the silicatecla y minerals are of 1:1 lattice structure. In view of the generally low cation exchange capacity ofth e clay (cf.unde r 14),i t islikel ytha t kaolinite isinvolved , not halloysite. From the fact that Ki values are only slightly below 2 the absence, or presence only in a small percentage, of free aluminum oxides (gibbsite = hydrargillite) may be deduced. It is likely that by far the majority of the determined A1203i sderive d from the silicate clay minerals.Th elo wvalue sfo r Fe203 indicate the presence ofonl ysmal l amounts offre e iron oxides(goethit e = limonite;hematite) .I n viewo fth elo wcatio n exchangecapacitie son e might assumetha t parts ofth edetermine d Si02 and A1203ar e not from kaolinite, but from inactive gels of free Si02 and free A^Oj respectively. However, full confirmation of the above deductions was given by several X-ray diffractions, Differential Thermal Analyses (DTA), and an electron micrograph. Subsurface (A3 horizon) and subsoil (Bhorizon ) samples of four profiles of the soils under discussion were analysed with X-ray at the Netherlands Soil Survey Institute of Wageningen, Holland, together with samples of some other Amazon soils. The results (cf. samples 233-3, 303-2, 303-4, 300-2, 300-4 and 210-4 of the Tables 7an d 8an d the Figs. 14a-e ) showtha t indeed kaolinite isth e very predominant constituent (80-85%) of the clay fraction1. Iron oxides account for up to 12% and aluminum oxides up to 5% o f the clay fraction, whilst clay sized quartz accounts up to 6% .A n electron micrograph was also made of the clay fraction of one of the samples (cf. Photo 9). The pureness of the kaolinitic clay is remarkable; practically all of the sample consists of strikingly clean and pronouncedly hexagonal, relatively large kaolinite crystals.Compariso n ofth edat a for specific surfaces ofth eabov e mentioned samples with their cation exchange capacities (Na-acetate at pH = 8.2) shows that a specific surface of 100m 2 gives only about 8.0 m.e. cation exchange capacity, which applies to both the total soilmas s and to the clay fraction (cf.Tabl e7) .

') The presence of kaolinite in the coarser fractions is due to incomplete dispersion, and enclosing by cementing sesquioxides.

Fig.14a-e X rayspectra ofseparates > 80\i(I4a), 16-80( i (14b), 2-16 \L (14C), <2\X (14d), < 2(i - treated (14e) of the main Amazon Latosols (samples233-3, 303-2, 303-4, 300-2, 300-4, 210-4) and plinthitic soils (samples96-2, 96-5, 302-4, 178-4) in comparison withthe composition of someother Amazonsoils (cf. theTables 7and8 andthe electron micro-photographs 9,10,11 and12) Fig.14a-e Curvas de Roentgengramas de separados > 80\J. (14a), 16-80j x (14b), 2-16 fi (14c), < 2( x (14d), < 2\x-tratados (14e) dos principals Latosolos amazonicos (amostras 233-3, 303-2, 303-4, 300-2, 300-4, 210-4) esolos de 'plinthite' (amostras 96-2, 96-5, 302-4,178-4) em comparacao de algunsoutros solos amazonicos (cf. Tabelas 7 e 8 e micrografias eletronicas:fotografias 9, 10, 11 e 12)

80 Fig. 14a (>80\x)

Inumily

?10-i - Ommi-lmbf>tfi i Afa (KaoTWIUtowl .wryhw y uiturad) _ »ji _ •# *• Fig. 14b(16-80 \i) Fig. 14c (2-16 p) Fig.14d(< 2 jxj

K sk40(iniM M » mica 'f »MWrmadtal* O squ^rtr atbilwyu Hi • htmatita Hd »hrfr»rg»l1it t L *limon>t« Ch >cNonw Sw l« * twvUins illila

'154-2 JJAvAjili 290-2

210-X- Ouimi-lmpfatri* Af a(Kaol.YtHLalo»o4.vtr y h««v Itilurtd)

r» ( ,-^.,^.f^wU Fig. I4e (< 2 (I- treatedltratados) 550*C 550*C Foto9 Micrografia eletronica dafracao deargila (< 2p) deLalosolo Amarelo Caolinitico (,Orto) de texturamuitopesada (amostra 303-4; Curud-una)*

PAo/o9 Electron micrograph ofthe clay separate (< 2\i) ofa Kaolinilic YellowLatosol (,Orlho), very heavytextured (sample 303-4; Curud-una)*

Foto10 Micrografia eletronica dafracdo de argila( < 2 \i) de SoloLaterita Hidromorfica (amostra 96-5;Caeti-Maracassume)*

• •*.;%»* <

P/io/o /0 Electron micrograph of theclay separate(<2y.)ofa Ground WaterLater ite soil(sample 96-5; Caete-Maracassume)* Foto 11 Micrografiaeletrdnica da fracdo de argila (<2\L) de Solo Clei Humico(amostra 188-2; Vdrzeado Baixo Amazonas*

Photo11 Electronmicrograph of the clay separate (< 2\i) ofa HumicCley soil (sample 188-2; Lower Amazonfloodplain) *

Foto12 Micrografia eletrdnicada fracao deargila (< 2y.) deSolonelz, fase Costeira(amostra 154-2; Leste da ilhade Marajd)*

r' V*- - * >*••: 4% •: . -n »,»,>* > *•** •

•• fc »•»..• it *» "*™"" -»• ' -» i <« 21 JUL-*! . " rS^BBT

Photo 12 Electron micrographof the clay separate (< 2y.) of a Solonelz, Coastal phase (sample 154-2; EasternMarajd island) * * By courtesy of Dr. H. Beutelspacher.Braunschweig .W-Cermany . Por obsequio doDr. H. Beutelspacher. Braunschweig, Alemanha. Table7 Classification, locationand general analytical data of samples withspecial analysis

Depth Classification Location Sample1 profun- classificacao localizacao amostra1 didade (cm)

Humic Gley soil HG Lower Amazon floodplain 188-2*** 20-70 Solo Glei Humico Varzea do Baixo Amazonas Low Humic Gley soil, Carbonate subsoil LHG.c Lower Amazon floodplain 190-2* 15-50 phase Varzea do Baixo Amazonas Solo GleiPouco Humico,fase subsolo com Carbonato Solonetz, Coastal phase Sol, c Eastern Maraj6 island 154-2*** 20-70 Solonetz,fase Costeira Leste da ilhade Marajd Solonetz, Coastal phase Sol, c Eastern Maraj6 island 175-3* 40-140 Solonetz,fase Costeira Leste da ilhade Marajd Solonetzic Humic Gley soil, intergrade to Southern Marajd island 178-4* 55-100 Ground Water Laterite soil Sul da ilhade Marajd Solo GleiHumico solonitzico, 'intergrade' para soloLaterita Hidromdrfica Hydromorphic Grey Podzolic soil, high HP/,6, o Araguaia Mahogany area 290-2** 7-25 base saturation, Ortho Area Araguaiana de Mogno 290-4 52-100 Solo 'Hydromorphic GreyPodzolic", satu- racdo debases aha, Orto

Red Yellow Podzolic soil, low base sat. RPH, Rio Branco do Acre 320-2 3-25 Solo Podzolico Vermelho-Amarelo, saturacao Rio Brancodo Acre 320-4** 60-85 de bases baixa Ground Water Laterite soil GL Caete-Maracassume area 96-2 15-70 : Solo Laterita Hidromdrfica Area Caeli-Maracassume" 96-5*** 150-170 Red Yellow Podzolic soil, int. to Kaolinitic RP-KYL, CR Guama-Imperatriz area 302-4** 500-600 Yellow Latosol, Concretionary phase AreaGuamd-lmperatriz Solo PodzolicoVerm.-Am., int.para Latosolo Amarelo Caol.,fase Concreciondria

Kaol. Yellow Latosol, medium textured KYLm Guama-Imperatriz area 233-3** 70-140 Latosolo Amarelo Caol., texturamedia Area Guama-Imperatriz Kaol. Yellow Latosol, very heavy textured KYL„& Curod-una centre 303-2 22-60! LatosoloAm. Caol., textura muitopesada Centro Curud-una 303-4*** 95-150 t Kaol. Yellow Latosol, medium textured KYLm Belem 300-2 32-75 \ LatosoloAmarelo Caol., textura media Belem 300-4* 15O-250 Kaol. Yellow Latosol, very heavy textured KYLr* Guamd-lmperatriz area 210-4** 60-150 LatosoloAm. Caol.,textura muito pesada Area Guama-Imperatriz

0 Number offield descriptio n and of horizon/mi/nero dadescricdo decampo e dohorizonte do perfil tr.- traces//rac

88 Tabela 7 Classified%do, localizacdo e dados analiticos gerais de amostras com andlise especial

Cation exchange H Granulomere _ capacity ori" separate ^ capacidade total ,on. fracdo granulo- org- 2 Exchangeable cations detroca Specific surface r'~ metrica cations trocdveis (Na-acet. pH = 8.2) superficie especifica zonte ++ ++ + <16(xl6-80(z>80fx Ca Mg K+ Na <2000(i < 2u. <2000u. < 2U. (50 (%) (%) (%) (m.e./100g) (m.e./100g) (m'/g) C„ 82.6 16.5 0.9 0.62 5.7 9.3 7.8 0.3 0.5 20.7 40.3 111 185

C,„ 63.9 35.0 1.1 0.80 5.2 8.4 6.3 0.5 1.7 20.5 40.3 93 178

B2» 81.7 16.5 1.8 0.70 6.8 4.0 16.5 0.4 3.3 26.7 43.6 126 175

B2*„ 79.2 19.5 1.3 0.26 7.7 3.4 11.2 , 0.6 9.6 26.6 43.8 109 188

B2ij 71.7 27.2 1.1 0.52 4.3 0.3 tr. 0.3 0.3 19.0 34.3 100 168

A, 10.2 52.4 37.4 2.03 5.0 6.1 38 B*» 0.27 5.0 4.9 1.5 0.2 0.2 11.3 72

A, 1.22 4.5 8.9 51 B» 69.6 21.7 8.4 0.65 4.5 15.0 78

A, 20.5 16.9 62.6 0.75 4.6 2.6 26 B«» 65.8 5.3 28.9 0.23 4.5 0.1 tr. 0.1 0.1 5.9 45 90

c 59.1 16.7 24.2 0.05 4.7 3.0 4.6 35 59

Bi 25.3 2.4 72.3 0.32 4.6 1.8 6.5 22 73

As 2.04 4.7 6.7 65 B» 77.5 18.2 4.3 1.06 4.7 tr. tr. 0.2 tr. 5.3 4.8 50 54

Ai 0.39 5.0 1.6 22 Ba 18.6 5.5 75.9 0.13 1.2 18

B» 70.6 26.5 2.9 0.76 4.7 2.9 4.8 48 81

• Mineralogical analysis only (.cf. Table 6)!sdmenteandlise mineralogica(cf. Tabela 8) ** also X-ray curves (c/. Fig. 14a-e) lambent curvasde Roentgengramas (cf. Fig. I4a-e) *** also Electron micrographs (cf. Photos 9-l2)Itambem micrografiaseletrdnicas (cf. Fotos 9-12)

89 Table 8 Mineralogical composition of the various granulometric separates

Depth Hori- SeparateIf"racao >80u Separate//raf

188-2 HG 20-70 Cig 73 10 12 5 47 10 16 15 12 190-2 LHG, c 15-50 C\q 64 7 F== 10 16 3 60 10 12 10 8 154-2 Sol.c 20-70 B2p 60 10 12 15 3 67 7 10 8 8 175-3 Sol, c 40-140 BMg 71 10 8 8 3 64 8 10 10 8 178-4 55-100 B310 20 4 76 88 4 4 4 290-2 HPAO, O 7-25 A, 95 5 94 6 32(W RPib 60-85 B, 90 10 92 8 96-5 GL 150-170 B«» 89 3 8 62 15 15 8 302-4 RP-K.YL, CR 500-600 C 30 35 35 18 45 30 7 233-3 K.YU 70-140 B, 88 4 8 86 5 6 3 303-4 KYL„» 95-150 B2t 65 23 7 Hd= = 5 10 77 5 8 300-4 KYLm 150-250 Ba 97 3 100 210-4 KYL„» 60-150 B, 75 18 7 17 69 6 8

Amorphous matter and heavy minerals (maximum 1 %) were neglected in the calculation of the minerals (mentioned in %,)!Material amorfo t minerals pesados (mdximo 1 %) foram desprezados no cdlculo dos minerals, cujo resultado i dado em %

') cf. Table Hcf. Tabela 7

Reference is made to the publication of CATE (1960). He studied, with both X-ray and DTA, samples of all horizons of a very heavy textured and a very light textured specimen ofth esoil sunde r discussion, collected at Curua-una centre.Thi sautho r also arrived at about 80-85% kaolinite in the subsurface and subsoil horizons, when calculated on the clay fraction.

This description shows that the majority of the well-drained soils of the Amazon Planiciehav e a subsoil horizon which constitutes a latosolic-B as described in II.2.1. The main Amazon Latosols fall into group IIIc of the tentative scheme of Table 6 elaborated byth enationa l BrazilianSoil sCommission . This, because of the relatively rather weak macro-structure of the soils, the slightly hard, to hard consistence in part of the profile when dry, the textural differentiation within the profile, and especially the composition of the clay fraction as expressed inth eSi0 2: Al203:Fe203 molecular ratios. The name 'Rego-latosol' might be applied for this subgroup of Latosols.Thi s name ishoweve r misleading,becaus ei timplie srelativel yyoun gage ,withou t development of definite genetic horizons. The soils under consideration, however, have a distinctly zonalcharacter .The yar ematur e soils,becaus ethe yar edeepl yan d strongly weathered and show definite profile development, although the transitions between the horizons are diffuse or gradual(cf. DAY, 1961).I n discussions within the Soils Commission, it

90 Tabela8 Composicao mineraldgica das vdrias fracoes granulometricas

Separate//rofSo 2-1 6| i Separate//rac3o< 2(x Q K L Ht Hd F M Ch Q K L Ht Hd F M It Sw Ch

35 15 15 20 15* 20 22 3 30 10 8 7 35 15 15 23 12* 20 26 3 20 16 8 7 35 15 15 20 15* 20 25 3 20 10 15 7 37 14 17 17 15* 20 25 3 15 6 24 7 59 13 2 2 3 18 3 25 35 6 3 2 10 16 3 92 3 5 30 30 15 25 62 12 8 8 7 3 24 24 4 8 7 33 15 56 12 It= 5 10 2 2 92 6 8 55 30 7 3 85 4 8 73 20 7 3 84 5 5 3 15 71 4 10 2 85 3 5 5 68 14 3 3 12 6 84 5 5 20 63 12 5 3 80 7 5 5

J* "•Quartziauartro; K - kaolinite/rao/Z/irVa,- L - limonite (goethite) limonita (goeihita); Ht - hematite/ f'matita; Hd - hydrdrgiHilehidrarxilita; F - felspar /Wrfjparo; M - mica (.Mile)I mica (ilila); It - 'ntermediate (between kaolinite and i\\ile)!intermedidrio (enire caolinita e ilila); Sw — swelling illite ""aexpanslvel; C h - chlorite/c/or/m Some swelling illite included;incluso ahuma ilila expanslvel

has already been agreed upon that such a term should be rejected. 'Rego-latosol' will not therefore be used, and a new term is proposed in view of the extent of these Latosols. For full characterisation it is considered the best plan to use the termKaolinitic YellowLatosol (KYL), since apparently the most common colour of the subsoil horizon is of a yellow hue. As mentioned, in transition areas to arid regions the colour tends to be of reddish hue, and the base saturation may be medium. For the time being,th e term Kaolinitic Red Latosol(KRL ) is applied in this case. Another name might however fit better. The light and very light textured profiles must fall outside the Latosols, and belong to the Acid Sands. Such profiles, as far as they occur on the Amazon Plarucie, are otherwise very similar to the Kaolinitic Yellow Latosols. Therefore, the term Kaolinitic Latosolic Sand(KLS ) is applied for those profiles that have less than 15% clay in the Bhorizon . Besidesth etypica l Kaolinitic Latosols,ther ear ei nth eAmazo n Planicie comparable soils, in which however several of the characteristics of a textural-B horizon (cf. II.2.1) are present. In this instance the names to be applied will be either: Kaolinitic Yellow Latosol,intergrade to Red Yellow Podzolic soil,or : Red Yellow Podzolic soil, intergrade to KaoliniticYellow Latosol, depending upon the degree of presence of these characteristics (for details cf. 111.2).

91 II.2.3. Comparison with other Classification Systems

Important other systems for classification of red and yellow soils of tropical and subtropical uplands have been developed for Africa. A. The classification for French speaking African countries (AUBERT and DUCHAU- FOUR, 1956). B. Theclassificatio n for Angola, developed byPortugues e soilscientist s (BOTELHO DA COSTA et al.,1959) . C. Theclassificatio n for Congo,develope d by Belgian soilscientist s (SYSet al., 1961). D. The classification for thesoi lma po fAfric a south ofth eSahar a (D'HOORE, 1959).

No attempt will bemad e to givea full comparison between the classification system as developed in the U.S.A. and in Brazil, and those mentioned above. Over-all correlations are in execution, for instance by FAO for its 'Soil map of the World' program.Referenc e ismad et oth eman ycomparativ enote sgive nb y LEMOS, BENNEMA, SANTOSet al.(1960 ) on the classification applied for Sao Paulo State. In this publication, an attempt is only made to establish the place of the Brazilian 'Latosols',an d morei nparticula r ofth e Kaolinitic Yellow Latosol and the Kaolinitic Latosolic Sand, in the African systems:

A. AUBERT and DUCHAUFOUR (1956) distinguish Sols rouges miditerranes, Sols ferrugineuxtropkaux (oufersiallitiques) and Solsferralitiques. The latter are charac terised bya 'ferralitic ' B horizon,wit hindividualisatio n ofiro nan daluminum ,an d are about identical with the Brazilian concept of'Latosols'. Within the Solsferralitiques the following subgroups are distinguished: 1. Solsfaiblement ferralitiques: Ki 1 1.7-2.0 2. Solsferralitiques typiques:Ki<1. 7 3. Solsferralitiques humiques (ou humiferes):mor e than 5% organic matter in the A horizon 4. Solsferralitiques a cuirasse enplace: with hardpan of hard plinthite, formed in flat terrain 5. Solsferralitiques a cuirasse dehas depente: with hardpan of hard plinthite, formed at the foot of slopes Thesols faiblement ferralitiques are comparable with the Kaolinitic Yellow Latosol, and perhaps also parts of groups 4 and 5. No special classification is given for the sandy ferralitic soils.

B. Theclassificatio n of BOTELHO DA COSTAet al. (1959)i scomparabl e withth e French one. Distinguished are Solos tropicais semi-dridos, Solos fersialiticos tropicais and Solosferraliticos. Th e latter are approximately identical with the Brazilian concept on'Latosols'.A nexceptio nt othi s is formed by a subgroup of the Solosferraliticos

*)SiO, : A1,0, ratio - see before.

92 whichcontain s weatherable primary minerals,an d isals o called Solospdra-sialiticos. The subdivision of the Solosferraliticos prope r is as follows: 1. Solopsamo-ferralitico: sandy 2. Solofort entente ferralitico or levi-ferralitico: not sandy, Ki <1.33 3. Solomediamente ferralitico: no t sandy, Ki < 1.7 4. Solofracamente ferralitico: no t sandy, Ki 1.7-2.0 Presence or absence of hard plinthite, and possible humiccharacter , appears only in the lower categories of the classification. It is apparent that the Kaolinitic Yellow Latosol is comparable with the Solo fracamenteferralitico, an d the Kaolinitic Latosolic Sand with part of the Solopsamo- ferralitico.

C. The classification of SYS et al.(1961 ) is rather different from the two mentioned above. For hisclassificatio n he discerns B textural, Bstructural and Bde consistence, which are defined as follows: Btextural: horizon ofcla yilluviation ,whic hi sa tleas ton efifth heavie ri ntextur e than the A and the C horizons, and has clay skins. Bstructural: horizon which has no clay illuviation but nevertheless clay skins on an appreciable part of the structure elements, and which has a firmer consistence than the A and the C horizons. Bde consistence: horizon without clay illuviation or clay skins, but only with a firmer consistence than the A and C horizons. It has a granular or weak to moderate subangular blocky structure, and contains often round pseudo-concretions of clay. This Bde consistence isno t identical with the 'latosolic-B' as described in II.2.1; the Latosols proper of the Congo are for a part described as A-C profiles. The Bde consistence seems to be identical with the transitional zone between the A and the B horizon (A3, B^, with stronger consistence, of some of the Brazilian Latosols (cf ad111c of Table 6o f II.2.1, and H.2.2). Recently, SYS(1962 )define d alsoa B ferralitique fo r the Congo.Thi son ei salmost ,i f not quite identical with the latosolic-B of Brazil. In the same publication it is stated that the B deconsistence forms the upper part of someferralitic- B horizons, and that the Congolese 'C horizon may form the lower part of this ferralitic-B horizon. The Bde consistencepoudreux applied to the soil survey of Rwanda-Burundi (FRANKART, HERBILLON and VERHOEVEN, 1962) is believed to be identical with the latosolic-B.

The Congo soils fall into two large groups, namely (1) soils from recent materials, and (2) soils from non-consolidated kaolinitic materials. The second group, called Sols climatiques, is divided into Kaolisols and Kaolisols lessee's. The former have a Bstructural o r Bde consistence, the latter a B textural. The well-drained Kaolisols are subdivided in hygro-, hygro-xero-, and xero-kaolisols. The hygro-kaolisols are found under tropical forest, in regions with less than two dry months. The hygro-xero- kaolisols are found under tropical savannah, in regions with more than two dry months. Both groups areferralitiques (largely kaolinite and sesquioxides in the clay

93 fraction), and have a base saturation that is below 40-50%; in part, they are humic (humiferes)1. The xero-kaolisolsoccu r in dryer regions. They are fersialitiques (besides kaolinite also appreciable amounts of silicate clay minerals of 2:1 lattice occur), have a Bstructural, an d a base saturation that is above 40-50%. The subdivision of the hygro- and hygro-xero-kaolisols is according to the degree of weathering of the kaolinitic material, and the texture. The subsoil horizon can be ferrisolique or ferraholique. Contrary to the ferraholique horizon, the former has either:

1. anappreciabl e amount ofcla yskin s(mor etha n2 5% )o nth ehorizonta lan d vertical aggregate units, i.e.a B structural or 2. a silt/clay ratio larger than 0.2 or 0.15 (for sedimentary rocks and alluvia, re spectively igneous and metamorphic rock), or 3. more than 10%weatherabl e minerals in the fraction 50-250 micron.

Three main groups within the hygro- and hygro-xero kaolisols emerge, namely: 1. Ferrisols.Thes e are ferrisolique; gibbsite may be present in small amounts; amorphic gels of silica and aluminum are present in appreciable amounts. 2. Ferralsols. These areferraholique, and have more than 20% clay in one of the horizons above 1 m depth. Gibbsite is often present; small amounts of amorphic gels of silica and aluminum only in a few cases. 3. Arenoferrals. These areferraholique an d have lesstha n 20% clay in the horizons above 1 m depth.

It iseviden t that of thehygro- and hygro-xero-kaolisols the ones that areferrisolique fall outside the concept of soils with a 'latosolic-B' horizon (cf II.2.1).Th e Ferralsols are however comparable with the Brazilian Latosols, and the greater part of the Arenoferrals iscomparabl e with the Latosolic Sands.Th e Ferralsols and Arenoferrals are not subdivided systematically according to the composition of the clay fraction (no data on the Si02: A1203: Fe203 molecular ratios). Among the various Ferralsols (e.g. JONGENan d JAMAGNE, 1959; SYS, 1960;i n the latter publication all the Ferralsols described have a B deconsistence, and the Arenoferrals A-C profiles), the Kaolinitic Yellow Latosol is most similar to the Ferralsols desplateaux du type Yangambi and theFerralsols des bas-plateaux de la Cuvette Congolaise.Th e KaoliniticLatosoli c Sand is similar to the Arenoferrals desplateaux du type Salonga. The similarity in morphometric field characteristics between the Amazon Kaolinitic Yellow Latosol and the Ferralsols of Congo, more in particular those of Yangambi, was already noted by D'HOORE and TAVERNIER on recent visits to Amazonia. The Fe203/clay ratios of the Yangambi soilsar eapproximatel y identicalwit h those of the

') Recently, the kaolisols humiferes, occurring in mountain areas, have been set apart from, and placed at the same level as the hygro-kaolisols, the hygro-xero-kaolisols and the xero-kaolisols (SYS, 1962).

94 KaoliniticYello wLatoso lan dth e KaoliniticLatosoli cSand ,bu tthei rcatio n exchange capacity seems to be considerably higher (10-15 m.e./100 g clay; cf. DE LEENHEER, D'HOORE and SYS, 1952, p. 37-41, and SYS, 1960, p.74) .

D. Recently, several drafts have been prepared for a general soil map of Africa, by the Inter-African Pedological Service of the Commission for Technical Co-operation in Africa south of theSahar a(C.C.T.A.) . D'HOORE (1960),i n explaining the legend of the third draft of this map, distinguishes: 'ferruginous tropical soils' (or fersialitic soils),'ferrisols ' and 'ferralitic soils.'Th eclassification s mentioned abovear e therefore combined toa degree. His ferralitic soils, which closely resemble the sols ferralitiques of AUBERT and DUCHAUFOUR, the solos ferralitkos of BOTELHO DA COSTA, and the Ferralsols + Arenoferralso f SYS,ar ecomparabl ewit hth e Brazilian Latosols. D'HOORE'S criteria for subdivision of the ferralitic soils are not specifically the composition of the clay fraction, but more the colour of the soil and the parent material. The Kaolinitic Yellow Latosol is believed tocompar e most closelywit h his mapping unit Kb: 'ferralitic soils with yellow-yellowish brown as dominant colour, and de veloped on unconsolidated, more or less clayey sediments'. The Kaolinitic Latosolic Sand is the most like the mapping unit Ka: 'ferralitic soils, with yellow-yellowish brown as dominant colour, and developed on unconsolidated sandy sediments'.

II.3 Plinthitic Soils

II.3.1 Origin of Plinthite

For the benefit of the following discussion plinthite will be subvivided into two groups, namely:

Soft plinthite (mottled clay, Fleckenzone, argiletachete, horizon bariole): A layer of soft (i.e. cuttable with knife), dense, usually clayey, humus-poor mineral material with many, coarse, prominent mottles.Th e mottles are red or purple1, often with admixture of some yellow, and occur in a white or light grey matrix. In case of predominanceo fth ereddish ,th esituatio nma yb edescribe da sth eoccurrenc eo fwhit e and someyello wmottle si n are d or purple matrix.Th epatter n of mottling isvarying . It may be reticulate (polygonal), prismatic (vesicular) or platy (laminar). The centres ofth ere do r purplepart s areofte n indurated tosom eextent .

Hardplinthite (iron concretions, Eisenkruste, laterite,cuirasse, ferruginou s quartzite, canga, pigarrd): A slag-like (i.e.onl y breakable with hammer), humus-poor mineral

') Actually usually weak red in the Munsell notation.

95 material, apparently largely consisting of indurated iron oxides; as'Well as the earth between thismaterial , ifpresent .Th eindurate d elementsvar yi n colour(re d to black), size (from fine gravel to enormous boulders and crusts), shape (pisolithic, platy, prismatic, massive, vesicular), grainage (fine to very coarse elements, usually of quartz, around which the sesquioxides are cemented), and arrangement (vertical, horizontal, irregular). Thesetw oname s will also be placed, between square brackets, after relevant terms in the literature referred to below.

According to HARRASSOWITZ(1926 , 1930),a fully developed 'laterite profile' should consist of the following sequence: 1. Eisenkruste (Zellenlaterit): ironstone crust, i.e. indurated, slag-like, porous sesquioxide [hard plinthite] 2. Anreicherungszone (Fleckenzone): enrichment zone, mottled zone [soft plinthite] 3. Zersatzzone(Bleichzone): dissolutio n zone, grey zone 4. Frisches Gestein: parent material...

Asalread y mentioned in II.1 ,HARRASSOWIT Ztoo k ita sprove ntha t theconcentratio n of iron and aluminum occurs at the soil surface, due to evaporation after transport of sesquioxide-rich soil water bycapillar y risefro m the greyzone . Hetherefor e assumed that laterite [plinthite]formatio n doesnot ,a t least not fully, take place under tropical rain climate or even monsoon climate, but requires a savannah climate. The laterite [plinthite] would not even be able to support tropical forest, due to the crust for mation (Waldfeindlichkeit). MARBUT (1932), however, observed in the Amazon valley that the HARRASSOWITZ sequence, if occurring, is found below soil material. He noted the following succes sion: nr. 1:soil ,nr .2 :iro n oxide layer, porous and slag-like [hard plinthite],nr . 3: mottled layer [soft plinthite],nr . 4: grey layer, nr. 5:unconsolidate d clay and sand. Thesecond 1laye r isofte n lacking.Th esoi labov eth ezon eo firo n concentration isal ways'podzolized' ,i nth e sensetha t it has a relatively lightcoloure d and light textured surface layer (A horizon), rich in Si02. MARBUT concluded from extensive field observations throughout the valleytha t the HARRASSOWITZprofil e isdu e to a process ofsegregatio nwhic htake s placea t shallowdept hbelow th e surface, underth e influen ceo fgroun d water, and may befollowe d byth eerosio n of theoverlyin g soil material. The layers of sesquioxide accumulation and induration: mottled zone [soft plinthite] and crust [hard plinthite], constitute essentially one horizon which develops at the surface of the ground water. The thickness of the horizon depends largely on the width of the zone over which the ground water surface fluctuates during the year.

') In the publication concerned actually is writtenfourth. From its context it becomes, however, apparent that MARBUT must have meant the secondlayer . Writing or printing errorsar ecommo n in thatpaper .The y may havecontribute d considerably to the confusion which subsequently arose on the subject.

*• 96 According to MARBUT, ifth ezon e of fluctuation isdee pn o sesquioxide accumulation takes place, because of a restriction in the access of atmospheric oxygen. The crust [hard plinthite] develops in the top of the horizon where the mottled material [soft plinthite] outcrops owing to erosion of the surface layer or at escarpments, or where themateria l occursbelo wa shallo wlaye ro fsand ysoi lmaterial .Crust s [hard plinthite] on dissected plateaux which nowadays have no shallow ground water level, are con sidered to be fossil, and the relics of a mottled zone [soft plinthite] which developed before relative raising of the terrain. With hisobservation s made in regions with a tropical rain climate,an d in agreement with thisconcep to f thedevelopmen t of the crust, MARBUT found noreaso n to believe that aclimat ewit h alternate wetan d dryseason swa sa requisite .Moreover , heargued , with the help of field observations, that formation of such a crust through capillary risean d evaporation at the surface inan y case ishighl y improbable. That such a pro cessi sals o difficult toassum efo r purely physical reasons,wa safterward s discussedb y MOHR and VAN BAREN (1954, p. 371). MARBUT concedes that 'the horizon of iron oxideaccumulatio n and induration maydevelo pa tth esurface , but onlyi nth eevident lyrar ecas ewhe nth egroun d watersurfac e liesa tth eearth' s surface'. Since MARBUT'Sobservations ,man yfield an d laboratory studies havebee n published on the subject. Many of them are in agreement with MARBUT'S conclusions. THORP and BALDWIN (1940), for instance, adducing evidence from China and Thailand, are convinced that the BUCHANAN'Slaterit e [soft plinthite] develops in the lower part of a soil called 'Ground Water Laterite'. This soil evoluates under intermittently shallow ground water level,b y transport of iron compounds from superficial horizons( A ho rizon) downward to the subsoil( Bhorizon ) wherethe y form reddish mottlestha t har den on exposure. They give a picture of how bleaching of the superficial horizon occurs in such soils ('podzolisation', 'lixiviation'). They also state that 'erosion and exposure of the laterite [soft plinthite] horizon isth e true explanation of the origin of themuc hdiscusse d lateritecrust s [hardplinthite]' . PRESCOTT and PENDLETON (1952), basing their conclusions mainly upon Australian occurrences,state :'th e evidence,therefore , istha t laterite [hard plinthite] isessential lyth eexpose d illuvial horizon ofa n ancient soil'. They consider itlikel ytha t the con centration of iron in the zone of fluctuating ground water is due to both downward movement from eluvialsurfac e horizons,an d thecarryin g upwardswit ha risin gwate r tablefro m thegre yzon e(fo r whichlatte rthe yprefe r theter m'pallid 'zone) .Practicall y all laterite crusts [hard plinthite] in Australia are considered to be fossil and of Late Tertiary age.The yar e believedt ohav ebee nforme d whenclimati can d geomorpholog- iccondition swer edifferen t from thepresen t dayones . Asregard sAfrica , aver yextensiv estud yo fplinthiti cmaterial s (zonesde accumulation desesquioxides), their different modeo fformation , and theirclassificatio n ina genetic system, is given by D'HOORE (1954). According to him, the accumulation of sesqui oxides followed by hardening can berelative or absolue. The former is a carrying off (leaching) of other constituents, principally of Si,fro m the horizon concerned, whilst the latter isa n addition of sesquioxides into the horizon.Th eaddition of sesquioxides

97 at the absolute accumulation is believed to take place predominantly downward, by soilsolution s that passverticall y(i nth eprofile ) or laterally (alongslopes) ,whil esome timesF ei sadde d byflooding water .Th efirst movemen t givesth ecuirasse de Vhorizon Bet Vhorizon gleyifii, the second the cuirasse de bassepente o r cuirasse denappe. The flooding givesth ecuirasse de galerie. A similar classification isapplie d by MAIGNIEN(1958) .H esuppose stha t Alaccumul ation horizons(bauxit e horizons)generall y belongt oth erelativ eones ,sinc eA li scon sideredt o beprincipall y aresidua lmaterial ,an dtha t Fe(an d Mn)accumulatio n hori zonsar eusuall y absolute. For West Africa, plinthite formation isthough t to begreat esti nth etransitio n zonebetwee ntropica l rainfores t and desert,wher e iti spracticall y independent of the parent rock. MAIGNIEN states also 'les cuirasses [hard plinthite] s'edifientnormalement a Vinterieur des profits. La mise a Vaffleurement se fait parirosion hydriquequi decape les horizons meubles de surface. Unecuirasse affleurante est la partie superieure d"unprofit tronque. Elle represente un stade senildevolution.'

Themos trecen t publication onplinthit ei sth erevie wb y SIVARAJASINGHAM,ALEXAN DER, CADYan d CLINE(1962) . Theygiv eals oa pictur eo nth echemistr y and theminera logyo fsesquioxid econcentratio n and hardening.

II.3.2. Plinthitei nAmazoni a

Sinceth eobservation so f MARBUT,a fe wgeographers 'description shav ebee npublish ed, in Portuguese, on lateritic crusts [hard plinthite] in Northern and Central Brazil (for instance GUERRA, 1953, 1954).Bu tther e has been no large-scale checking or ela boration of MARBUT'Sfindings. A detailed report on the widespread occurrence of plinthitic materials in Amazonia, and a discussion in relation to soil classification, is therefore thought to be useful, also in view of the fact that around the world still relatively few complete data about 'Ground Water Laterite' soil are reported. The formation of plinthite in Amazonia takes place predominantly on flat land sur faces with a cover of unconsolidated sediments of Tertiary or Quaternary age. There are also a number of areas with Paleozoic-Mesozoic outcrops or peneplained Pre- Cambrian crystalline basement which show formation of the material. As will be demonstrated inth efollowing , theAmazo nplinthit eformatio n is,wit hfe wexceptions , of the typecalle daccumulation absolue b y D'HOORE(1954) , more especiallyth e forma tion ofcuirasses de Vhorizon B et Vhorizon gleyifie. Afluctuating groun d water level,o r pseudogroun d water level,i sindee dessentia l for the formation of byfa r the majority ofth eplinthiti cmaterial s ofAmazonia . Onlya fe winstance swer eencountere d inwhic ha fluctuating groun d water levelma y notb ea nessentia lfacto r inplinthit eformation . Thisconcern sconsolidate d sediments (cf. Profile descriptions 35 and 41, and the notes on Non Calcic Brown-like soil, Gravelly phase and Acid Brown Forest-like soil, Gravelly phase in III.2.). The plin thite formation in these instances, which will not be further discussed, probably falls

98 under D'HOORE'S accumulation relative. It may be a feature of a general weathering process inth etropic s under conditions of good drainage. Thepossibilit y ofaccretion ,withou t afluctuatin g ground watertabl eproper ,o f fossil plinthite layers in areas of unconsolidated sediments was suggested in one instance {cf. page 115).

11.3.2.1 Plinthite-in-formation The presence of a relatively shallow and fluctuating ground water level, or pseudo ground water level or saturation zone, is frequent in Amazonia because of various factors. There is a large expanse of flat land surfaces, which are moreover low lying with regard to the local drainage levels: areas belonging to the Early Tertiary (?) peneplanation surface, sections of the Plio-Pleistocene planalto, portions of the Late Pleistocene terraces,an d the Early Holocene terrains.Ther eare , inman y places,tem porary difficulties inth edischarg eo fth erainwater , due to seasonally high water level of many of the rivers and yearly peaks in rainfall distribution in a large part of the region. It may benote d thatfluctuations o fth egroun d water level are not necessarily linked to the occurrence of wet and dry seasons. Suchfluctuations ar e also governed bydifference s inlatera l drainage possibilities that are linked with seasonal differences in water discharge of the river system in the area. In many parts of Amazonia, this discharge is determined more by the amount of water drained from the Andes and Central Brazil, which has a large seasonal variation, than by the rain falling in the region itself. Another factor which favours the occurrence of shallow ground water levels is the presence, at a number of places, of shallow-lying impervious layers, for instance hard sand stone or fossil hard plinthite{cf. 1.4.5). Theabsenc eo fan yenrichmen t ofth eabov ementione dflat lan d surfaces throughth e deposition of sediments offlooding, o f volcanic asheso r of other windborne material is noteworthy {cf. SAKAMOTO, 1960).Thi s absence isanothe r condition favouring the formation of plinthite, which is essentially a highly weathered material.

FORMATION OFPLINTHIT EBELO WO R IN THELOWES TPAR TO FTH E SOLUM Atsevera lplaces ,notabl y whereligh ttextured , loosesediment sfor m thecove ro f the land, it was observed that plinthite is formed in a zone of fluctuating ground water levelwhic hi srelativel y deepbelo wth esurfac e (arbitrarily deeper than twometres) .I n theseinstances ,th esolu m propero fth esoi lprofil e isnormall yno t affected bythi s for mation. The soil is well or moderately well drained, and constitutes for instance a Kaolinitic Yellow Latosol. The zone offluctuating groun d water levelis , apparently, toodee pt ohampe rful l activityo froot san d soilfauna , whichshoul dhel pt ofor m and maintainth elatosoli cprofil e asi tis . Such a situation wasobserve d in several parts of the relatively low uplands that may be found in the Estuary region (predominantly Epi- and Late Monastirian terrace levels, cf. 1.4.3).Th e fluctuations in the ground water level in these parts are largely determined by tidal movements, which take place at 2t o 5m below the surface. On cliff faces, the soft plinthite formed in the fluctuation zone becomes exposed at low

99 tide. The hardening relics,whic h inloco usually have a vesicular-reticulate form, are often found onth e beaches.The yhav ea blac k surface duet oth eactio n of theestuar y water (pedra pretao f Soure,Mosqueir oan d otherplaces ;Amap atown ; cf.th enote s of MARBUT and MANIFOLD, 1926, p. 434). The following short profile description illustratessuc ha situation:

Profile 1. (KAOLJNITIC YELLOW LATOSOL, medium textured - over plinthite-in-formation) Field description 185A(Sombroek ) Marajo-island, Soure, river cliff Terrace 2-3 m above high waterleve l (fluctuations in waterleve l are 2-3 m, largely due to tides). A, 0-25 cm: Dark brown (10YR 3/3) light sandy loam. Many roots, many pores. Transition gradual. A, 25-70 cm: Yellowish brown (10YR 5/6) sandy loam. Many roots, many pores. Transition gradual to diffuse. BSi 70-150 cm: Brownish yellow (10YR 5/8) light sandy clay loam. Common roots, many pores. Transition diffuse. B2J 150-210 cm: Reddish yellow (7.5YR 7/8) sandy clay loam. Common roots, many pores. Transition gradual. C,„ 210-260 cm: Pale yellow (2.5Y 8/4) light sandy clay loam, with common medium sized distinct mottleso f white(2.5 Y 8/2) andre d(2.5Y R 5/8),especiall yi nth elowe rpart .Transitio n gradual. Cm 260-350 + cm: White (10YR 8/1) light sandy clay, with many coarse prominent mottles of red MOR4/6 ) and some yellow (10YR 7/8) and pink (5YR 7/4), the latter especially in the lower part;mottlin g isi n avesicula r tocoars eprismati c pattern.O n theclif f face itself there d parts are hard (hard plinthite), but in from the bank the whole horizon issof t except forsom e tiny centres in the red (soft plinthite).

In this case, the plinthite formation is more a geological than pedolgiocal process. Most probably,ther e isn o transport of sesquioxides and/or clay sized particles from theuppe r 200 cm. toth ehorizo no fplinthit e formation. Similar horizons of soft plinthite are reported to occur sometimes in the B3 and C horizonso fa fe wsoil so f Rio deJaneir oan dSa oPaul oState swhic har ecomparabl et o the Kaolinitic Yellow Latosol (cf. BARROS, DRUMOND, CAMARGO et aL, 1958, p. 289; LEMOS,BENNEMA ,SANTO Set al.,1960 ,p . 389).

FORMATIONO FPLINTHIT EWITHI N THE SOLUM: GROUND WATER LATERITESOIL S Theformatio n ofplinthit eneare rth esurfac e wasobserve d inman yplaces .Thes ear e terrains with an imperfect drainage, on which the zone of fluctuating ground water leveli sshallo w(arbitraril y at lesstha n two metresdepth) . In these instancesthi szon e isshallo w enought o restrict the activity of roots and soilfauna . No full homogenisation,whic husuall ywoul d resulti n Latosolformation , istherefor e possible.Clay-size d particlesan d sesquioxides arecarrie d downwardt oth ehorizo no fplinthit e formation. The latter, in effect, now forms the Bhorizo n of a pedological profile. The Ground WaterLaterite soilprofil e proper, as defined by MARBUTan d successors, develops. Thefirst larg eexpans eo f thesesoil swa sobserve d by DAY(1959)i nth e Caete-Mara- cassume" area. Afterwards, Ground Water Laterite soilswer e found to occur also in many parts of the Lower Amazon region and on the Island of Maraj6 (DAY, 1961;

100 SOMBROEK,1962b) . Scatteredobservation selsewher ele dt oth econclusio ntha t thesoi l isver y common throughout Amazonia,ofte n associated with areas of natural savan nah orsavannah-fores t(cf. IV.1.2.2) . Severalphase so fth esoil ,an dintergrade so fi tt o other soils,hav ebee ndistinguishe d to date.A pictur eabou tth evariabilit y inth epro file characteristicsma yb e obtainedfro m thefollowin g shortdescriptions , partlybase d onprofil e pitsan dpartl y onaugerings . Mosto f theseprofile s wereanalysed .Th eana lyticaldat a that arethough t to be of importance for theclassificatio n of the soils are giveni nAppendi x9 . Aboveeac hprofil edescriptio n theprovisiona lclassification , asuse di nth e surveyre ports,i sprinte d betweenbrackets .I nthi spublication ,n oattemp twil l be made to give anelaborat e scheme for classification of thevariou sGroun d WaterLaterit esoils .I ti s evident,however ,tha tth edescribe d profiles allcom eoutsid eth eLatoso lo rth eOxiso l Order, because the plinthitic horizon has many of the characteristics of the 'textural- B\ c.q. the'argilli chorizon' .Th esoil ssee mt oconstitute ,i nfact , akin do f imperfectly drained phaseo f the Red Yellow Podzolicsoils .Man yo fthe m probably come under

Foto13 O perfil de urn solo Laierha Hidro- morfica.Claremente visiveis sao o horizonte-Ax coloridode humo, ohorizonte A t arenosobranco e o horizonte B densae de texlura pesada de 'plinthite''macio, comsens numerosos mosquea- I*" """-•"• dos grossos e prominentes de maliz vermelha ** -%• nnma matriz de cinzento claro. Neste perfil os »^» <:- mosqueados lent padrao reticular. Porcausa do cardtermuito alvejadode horizonte At,o perfil classifica-seentre o chamado solo Later it a Hidro- morfica,fase Baixa(fotografia Mr. Th.H. Day)

t . _ • I Vi

-, *- - J -*» - . * * • -J

Photo 13 Theprofile of aground WaterLaterite S °H. Clearly visibleare the humus stained At horizon, the while sandy At horizon, and the dense, heavytextured Bhorizon ofsoft plinthite, with its many coarseand prominent mottles of 1 red huein a lightgrey matrix. Themottles have ' - vawti. -*>;- ,„•-> fif\ a reticulate pattern in this profile. Because of thestrongly bleached character of the Athorizon, the profile falls under the so-called Ground Water Laterite soil.Low phase (Photo bv Mr. Th. H. Day) •fc—i^Aifiiii iiiiiiViiKi

101 the 'Ultisols' of the VNth Approximation, more especially the Ochraquultic Plinta- quults (SOIL SURVEY STAFF, 1960,p . 226-227). The Ground Water Laterite soil,Lo w phase seemst ob e identical toth e 'Gray Ferruginous soil'a sdefine d inth eMultilin gualVocabular y of SoilScienc e (JACKS, TAVERNIER and BOALCH,1960) .

Profile 2(GROUN D WATER LATERITEsoil , medium textured phase) Field description 124(Day , Sombroek) Lower Amazon region, 20k mNN W of Prainha (Lat. 1°.48'S; Long. 53°.40'W) Flat;remnan t ofterrac eca. 50m aboveloca l rivulet.Imperfectl y drained becauseo fimperviou slaye r of fossil plinthite atshallo w depth (outcropping at sides);uppe rphreati c level- ?m ,lowes t phreatic level below —1.1 m. Early Pleistocene sediments. Grasses and scattered shrubs; yearly burned. Surface strewn with platy slabs of hard plinthite, stratified coarse andfine grained .

A, 0-10cm: Yellowish brown (10YR 5/4) lightfine sand y loam. A, 10-30 cm: Brownish yellow (10YR 6/6)fine sand y loam. B„ 30-60 cm: Reddish yellow (5YR 6/8)fine sand ycla y loam,wit h fewt ocommo n medium sized prominent mottles ofwea k red (7.5R 4/4); also some white spots. About 15% small (2-3 cm) coarse grained plinthite concretions. B& 60-110 4- cm: Lightreddis h brown (5YR6/4 )clay , with many coarse prominent mottles of dusky red(7. 5R 3/4) ,stron g brown (7.5YR 5/8) and light grey(N7/0) .

Profile 3 (GROUND WATER LATERITE soil, medium textured phase) Field description 136 (Day, Sombroek) Lower Amazon region, 5 km Eo fTerr a Santa (Lat. 2°.07'S; Long. 56°. 27'W) Extensive flat terrain, 1-2m above local river high water level. Imperfectly drained; upper phreatic level - ?m ,lowes t phreatic level below —2.3 m.Earl y Holocene sediments. Grasses and patcheso f low trees;yearl y burned. A, 0-10 cm:Dar k grey brown (10YR4/2 ) very friable fine sandy loam, with a few fine faint mottles of strong brown (7.5YR 5/8). Transition clear. A, 10-40 cm: Brownish yellow (10YR 6/8)friabl e fine sandy clay loam, with afe wfine distinc t mottles of grey (10YR 5/1). Transition clear. Bji» 40-150 cm: Reddishyello w( 5Y R6/8 )firm fine sand yclay ,wit hman ycoars eprominen t mottles of red (2.5YR 5/8). Within the red some dusky red hardening centres. Transition diffuse. Bua 150-225 + cm: Light red (2.5YR 6/8)firm fine sand y clay, with many coarse prominent mottles of light grey (N 7/0). Within the light red several dusky red, hardening centres.

Profile 4 (GROUND WATER LATERITE soil, Low phase) Field description 269 (Sampaio) Lower Tocantins, 2k m Eo f Curucambaba (Lat. 2\9'S; Long. 49.17'W) Extensively flat; terrace, 8 mca. abov e local river level. Imperfectly drained; upper phreatic level +0.2 mca .(rai n water), lowest phreatic level below —3.3 m. Late Pleistocene sediments. Grasses and scattered low trees and palms; yearly burned.

A, 0-40 cm: Very dark grey (N 3/0)ver y friable light sandy loam. Transition gradual. A, 40-100 cm: White (10YR 8/1)friabl e sandy loam. Transition clear. AB, 100-150 cm: White (10YR 8/1)friabl e light sandy clay loam, with fewt o common medium sized distinct mottles ofstron g brown (7.5YR 5/8). Transition gradual. B2l, 150-220cm: White (10YR 8/1)firm cla y loam, with many fine distinct mottles of yellowish red (5YR 4/8)an d reddish yellow (7.5YR 6/8). Within there d very small hardened nodules. Transition gradual. B2a, 220-330 + cm:Re d(2.5Y R 4/8)firm ligh t clay loam, with many medium sized prominent mottles of white (10YR 8/1) and reddish yellow (7.5YR 7/8). Within there d small hardened nodules. »•

102 Profile 5 (GROUND WATER LATERITEsoil ) Field description 96 (Day) Caete-Maracassume area, Piria river (Lat. P.32'S; Long.46°.27'W ) Flat terrace, some metres above local river level. Imperfectly drained; upper phreatic level - ?m , lowest phreatic level below —3.0 m. Pleistocene sediments. Young secondary forest.

Ap 0-15 cm:Ver ydar k brown(10Y R 2/2) veryfriabl e loamy sand.Structureless ,o rwea k medium sized granular. Transition gradual and smooth. AI 75-70 cm:Dar k greyish brown (10YR 4/2) friable sandy loam,wit ha fe wfine fain t mottles of brown (10YR 6/8). Weak coarse angular blocky. Transition gradual and smooth. At 70-110cm: Yellowis h brown(10Y R 5/6)friabl esand yloam ,wit hcommo n mediumsize d faint mottleso f lightyellowis h brown(2.5 Y 6/4) andreddis h yellow(7.5Y R 7/8).Ver y weak coarse angular blocky. AB 110-120cm: Heavy sandy loam. Transition zone. B„, 120-140cm: Yellow (2.5Y 7/6)firm ligh t sandy clay loam, with many medium sized distinct mottles of red(2.5Y R 5/8). Weak very coarse subangular blocky. Transition abruptan d smooth. B,u, 140-150 cm: Gravelly light sandy clay loam, with same colours as B„t. 50-75% quartz pebbles (1-3 cm diam.). Transition abrupt and smooth. II B1S0 150-170+ cm: White (10YR 8/1)firm heav y sandy clay loam, with many coarse prominent mottles ofre d (10R 4/6). Inth e red some dusky red hardening nodules.

Profile 6 (GROUND WATER LATERITE soil, intergrade to GREY HYDROMORPHICsoil ) Field description 83 (Day) Caete-Maracassume area, Maracassume river (Lat. 1°.40'S; Long. 45°.52'W) Flat; terrace, some metresabov e local riverlevel . Imperfectly drained;uppe rphreati c level probably + 0.2m ca., (rai n water), lowest phreatic level probably —2.0 m. Shallow Pleistocene sediments over Pre-Cambrian granitic rock. Grasses.

A, 0-5cm: Pale brown (10YR 6/3) loose sand. Single grains.Transitio n abrupt andwavy . A„ 5-15 cm: White (10YR 8/2) loose sand. Single grains.Transitio n clear and wavy. An 15-30 cm:Yello w(10Y R 8/6) very friable to loose sand,wit h commonfine distinc t mottleso f reddish yellow (7.5YR 6/8). Weak coarse angular blocky. Transition clear and wavy. (H)B„30-40 cm: Pale yellow (2.5Y 8/4) firm sandy loam, with many coarse distinct mottles of reddish yellow (7.5YR 6/8). Weak to moderate coarse subangular blocky. About 5% small (1-2 cm diam) plinthite concretions. Transition clear and irregular. (IDBJB40-100 cm:Whit e (N 8/0) firm heavy sandy clay loam, with common coarse prominent mottles of red (2.5YR 4/8). Moderate tostron g medium sized columnar. (II)BW 100-140 + cm:Whit e( N 8/0)firm ligh t loam,wit h manycoars eprominen t mottleso freddis h yellow (7.5YR 7/8). Throughout theB horizo n a few medium sized (3-5c m diam.) angular stones ofquart z and grano-diorite(?) .

Profile 7 (GROUND WATER LATERITTsoil ) Field description 162/173 (Day, Sombroek) Cf. III.2, Profile 43.

Profile 8. (GROUND WATER LATERITE soil) Field description 199 (Day, Sombroek) Guama-Imperatriz area, km 129 E(Lat . 2°.45'S, Long.47°.25AV ) Slightly dipping part of extensive terrace, 50m ca.abov eloca l rivulet level. Micro-relief ofkauwfoe- loes1.Imperfectl y drained; upper phreatic level +0.5 mca . (rain water), lowest phreatic level below

') Also called 'pocket' micro-relief (DAY, 1961), or canaletes(SOMBROE K and SAMPAIO, 1962).

103 —1.5 m. Plio-Pleistocene lacustrine sediments (Belterra clay), reworked during Early Pleistocene. Shrubby low forest.

At 0-3cm: Light grey(10Y R 6/1)friabl e clay. Moderatefine subangula r blocky.Transitio n clear. Ajf 3-20 cm:Whit e(10Y R 8/2)friabl e tofirm heav yclay ,wit h manyfine fain t mottleso fbrownis h yellow (10YR 6/8). Moderate to weak medium sized subangular blocky. Transition gradual. B«s 20-100cm: White (10YR 8/1) firm heavyclay , with many medium sized prominent mottleso f red (2.5YR 4/8)an d someyello w (10YR 7/8). Moderate towea k medium sized subangular and some angular blocky; tendency toplaty . Afe wfain t clay skins. Transition gradual. B220 100-130 + cm:Re d (2.5YR 6/8)ver y firm heavy clay, with many medium sized prominent mottles of white (10YR 8/1)an dsom e yellow (10YR7/8) .

Profile 9 (GROUND WATER LATERITE soil) Field description 214 (Sombroek, Sampaio) Guama-Imperatriz area, km25 2 (Lat. 3.42'S; Long. 47°.29'W) Slightly dipping part in almostflat terrain ; terrace 5 mca. aboveloca lrive rlevel .Imperfectl y drained; upper phreatic level -? m, lowest phreatic level below —1.5 m. Late Pleistocene sediments. High forest, of low timber volume.

Ai 0-5 cm: Brown (10YR 5/3)ver y friable fine sandy loam. Weak medium to fine subangular blocky. Transition clear. At 5-40 cm:Ver y palebrow n (10YR 7/4)firm fine sand y clay loam,wit h common - in lower part many - fine distinct mottleso fstron g brown (7.5Y R 5/8). Moderate medium sized subangular blocky. Transition clear. B,j 40-80 cm: White(10Y R 8/2) veryfirm clay ,wit h manyfine t omediu m sized distinct mottleso f strong brown (7.5YR 5/8) and light red (2.5YR 6/8).Wea k tomoderat ecoars eangula r blocky. Common faint clay skins. About 10%rathe r soft, only partly loose plinthite concretions, small(0.5- 2c m diam.), ratherfine graine d and dark reddish brown (2.5YR 3/4) orre d (2.5YR 4/8). Transition gradual. Bw 80-150 + cm: White (2.5YR 8/2) very firm clay, with many coarse distinct to prominent mottles ofligh t red(2.5Y R6/8 ) andsom e yellow(10Y R 7/8).Massive ,t o weak coarse angular blocky.

Profile 10 (GROUND WATER LATERITE soil,ligh t textured phase-intergrad e to KAOLINITIC LATOSOLIC SAND); (SANDY GROUND WATER LATERITE soil) Field description 151 (Sombroek) Maraj6 island, 40k mN of Muana (Lat. 1°.09'S, Long. 49°.!!^ Ridge-like terrain (teso), 1-2m abov e submergeablelowland . Imperfectly tomoderatel y well-drained; upper phreatic level-1. 7 mca., lowes t phreatic level—2. 5m ca. Late Pleistocene sediments. Grasses and scattered low trees andpalms ; yearly burned; fertilized.

Alt 0-20cm: Grey (10YR 6/1) loose fine sand. Single grains. Transition gradual. A„ 20-70 cm: Dark grey brown (10YR 4/2) very friable fine sand. Structureless, to weak fine crumbly. Transition gradual. A, 70-140 cm: Pale yellow (2.5Y 7/4) loose loamy fine sand. Transition gradual. Bie 140-170cm: Brownish yellow (10YR 6/8) loose loamy fine sand, with common medium sized distinct mottles of red (10R5/8) . A fewsmal l hard plinthite concretions. Transition gradual. B2i» 170-200 cm: Red (2.5YR 4/8)loos e loamy fine sand, with many coarse distinct mottles of brownish yellow (10YR 6/8). Transition gradual. B220 200-230cm: White( N 8/0) looseloam y fine sand, with many medium sized distinct mottleso f light red(2.5Y R 6/8)an dyello w (10YR 7/8). Transition gradual. Ct 230-325 + cm:Whit e (N8/0 )loos e fine sand.

104 Profile 11 (GROUND WATER LATERITE soil, heavy textured phase); (GROUND WATER LATERITE soil, Humic phase) Field description 157 (Sombroek) Maraj6 island, 40k mN o f Muand (Lat. 1°.09'S; Long. 49°.!!^) Flat terrain. Imperfectly topoorl y drained; upper phreatic level +0.5 m (rain water),lowes t phreatic level -3.0 mca. Early Holocene (?) sediments. Grasses, scattered low trees andpalms .

A! 0-30 cm:Blac k (N 2/0),ver yfriabl e humiccla yloam .Stron gfine granular . Transition gradual. At 30-60 cm: Light grey (10YR 7/1) firm clay, with a few fine distinct mottles ofreddis h yellow (7.5YR 6/8). Transition gradual. BM 60-270 cm: White (N8/0 )ver y firm clay, with many coarse prominent mottles of dark red (7.5R 3/6) andsom e yellow (10YR 7/8). Throughout the horizon inth e dark redman y small. dusky redan dhal f hard nodules. Note: Asimilar profile in the neighbourhoodshowed in the B, thefollowing structure: Weakto moderate coarseprismatic, breaking into strong coarse subangular blocky. Presence of clay skins.

Profile 12 (GROUND WATER LATERITE soil, intergrade to KAOLINITIC YELLOW LATOSOL, or reverse) Field description 317 (Sombroek, Falesi) Porto Velho, km7 2o fBR-29 . Extensive flat terrace, 10m abov e local rivulet level.Imperfectl y to moderately well-drained; upper phreatic level -1.0 mca., lowest phreatic level below —2.0m .Plio-Pleistocen e lacustrine sediments (Belterra clay). Low forest with many creepers.

A! 0-10 cm:Greyis h brown (10YR 5/2)friabl e lightclay .Moderat efine t omediu m sized subangu lar blocky. Transition clear. A, 10-60 cm: Light olive brown (2.5Y5/4 ) friable clay. Moderate medium sized subangular blocky. Transition diffuse. B, 60-100 cm: Light yellowish brown (10YR 6/4) friable tofirm clay . Weak tomoderat e medium sized subangular blocky. Transition gradual. Bsl 100-150 cm: Light yellowish brown (10YR 6/4)firm clay . Weak to moderate medium sized subangular and angular blocky. Common faint clay skins.Scattere d (2%) small (1 cm diam.) fine grained dark reddish brown (2.5YR 3/4) rather soft plinthite concretions. Transition gradual. B829 150-200+ cm: Pale yellow (2.5Y 7/4) firm clay, with many medium sized prominent mottles of dusky red(10 R 3/4) and somedar k reddish brown (2.5YR 3/4),yellowis h brown (10YR 5/8) and white (N 8/0)". Moderate medium sized to coarse angular and subangular blocky,-with tendency toprismatic . Common faint clay skins.

Profile 13 (GROUND WATER LATERITE soil, intergrade to Low HUMIC GLEY soil) Field description 146 (Sombroek) Lower Amazon region, 20k mS o f Prainha (Lat. 2°.02'S; Long. 53°.30'W) Patch of low upland (teso), 0.5m above high water level, within floodplain. Imperfectly drained; upper phreatic level 0.0 mca., lowest phreatic level about -3.0 mca. Early Holocene (?) sediments. Grasses, scattered high trees.

Ai 0-10 cm: Dark grey (10YR 4/1) loamy fine sand. Transition clear. A, 10-25cm: Light grey (10YR 6/1) loamyfine sand , with fewfine fain t mottleso freddis h yellow (7.5 YR6/8) . Transition clear. AB? 25-60 cm: Light grey (10YR 6/1)fine sand y loam, with many medium sized distinct to pro minent mottles ofyellowis h red(5Y R 5/8). Transition gradual. Bjj 60-225 + cm:Whit e (N8/0 )firm fine sand y clay loam, with many coarse prominent mottles of red (2.5Y R 4/8),partl y bounded with yellow(10Y R7/6) .Fro m 120-150c m also some weak red (10R 4/3) mottles with hardening centres.

105 Profile 14 (GROUND WATER LATERITE soil,intergrad e to HYDROMORPHIC GREY PODZOLICsoil ,Dee p phase) Field description 289 (Sombroek, Sampaio) Araguaia Mahogany area, Bloco Piranha (Lat. 6°.01'S; Long. 48°.10'W) Almost flat terrain, 1m ca. above level of local intermittent rivulet. Micro-relief of kauwfoetoes. Imperfectly drained;uppe rphreaticleve l +0.3 m ca. (rainwater) ,lowes tphreati cleve lbelo w—4. 0m . Triassicfine sand-ston e (?). Low forest with manycreepers .

A, 0-50 cm:Ver ydar k grey(10Y R 3/1) looseligh t sandyloam . Single grains.Transitio n gradual. A, 50-140 cm: White (10YR 8/1) loose light sandy loam.Transitio n clear. AB„ 140-175cm: Whit e (10YR 8/1) friable sandyloam , with common medium sized faint mottles of brownish yellow (10YR 6/8). Transition clear. Bw 175-250cm: Whit e(5 Y 8/1)firm ligh t clayloam , withman ycoars eprominen t mottleso fdar k red (7.5R 3/8) and reddish yellow (7.5YR 6/8). Central parts of dark redhal f hardened. Transition gradual. B^ 250-420cm: (Sample d 250-320 cm).Whit e (5Y 8/1)firm heav y loam, with many very coarse prominent mottles of dark red(7.5 R3/8) , reddish yellow (7.5YR 6/8)an d some palere d (10R 6/4). Central partso fdar k red somewhat hardened.

Profile 15 (GROUND WATER LATERITE soil,intergrad e to HYDROMORPHIC GREY PODZOLIC soil, Gay stone substratum phase - partly truncated?) Field description 287 (Sombroek, Sampaio) Araguaia Mahogany area, near Bloco Piranha (Lat. 6°.02'S; Long. 48°.ll'W ) Almost flat terrain, about 2 m above level ofloca l intermittent rivulet. Imperfectly drained; upper phreatic level +0.1m (rai n water), lowest phreatic level below -1.2 m.Jurassic-Triassi c silty clay- stones. Low forest with many creepers.

Ai 0-5 cm: Palebrow n(10Y R6/3 )ver yfriabl e light loam, with manyfine fain t mottles ofyellow ish brown (10YR5/8) . Weak to moderate fine subangular blocky. Transition gradual. AM 5-30 cm:Ligh t grey(10Y R 7/2) very friable loam,wit h manyfine distinc t mottleso f yellowish brown(10Y R 5/8).Wea k medium tofine subangula rblocky . About 80% small( <1 cmdiam. ) hard,fine graine d dark red (7.5R 3/8) plinthite concretions. Transition clear. B,„ 30-50 cm: Light grey (10YR 7/1) friable clay loam, with many medium sized faint mottleso f yellowish red(5Y R 4-5/6). Moderatefine subangula rblocky .A fe wfain tcla yskins .Transitio n gradual. BJJ 50-80 cm: Verypal ebrow n(10Y R 7/4)friabl et ofirm silt ycla yloam ,wit hman y mediumsize d prominent mottleso fre d(10 R4/6 ) andwhit e(10Y R 8/2). Moderatefine angula rt o subangular blocky, compsing weak medium prismatic. Common faint clay skins. The centres of the red parts somewhat hardened. Transition gradual. Bw 80-100cm: White(10Y R 8/1)friabl et ofirm silt ycla yloa mwit hman ymediu msize d prominent mottles of red (7.5R 4/8) and yellow (10YR 7/6). Moderate finet omediu m sized angulart o subangular blocky, composing weak medium prismatic. Common distinct clay skins. Transi tion gradual. C9 100-120 + cm:Whit e( N 8/0) friable silty clay loam, with many medium tocoars e prominent mottles ofdar k red (7.5R 3/8) and yellow (10YR 7/6). Weak medium angular tosubangula r blocky structure.A fe w faint clay skins.

Profile 16 (GROUND WATER LATERITE soil, intergrade to HYDROMORPHIC GREY PODZOLIC soil, Micaceousphase ) Field description 274 (Sombroek, Sampaio) Araguaia Mahogany area, opposite Xambioa (Lat.6°.29S ; Long. 40C14'W) Narrow strip of lowland along intermittent rivulet. Imperfectly to poorly drained; upper phreatic level +0.5 mca. (rivulet water), lowest phreatic level below —1.5 m. Colluvium-alluvium from sur rounding Pre-Cambrian micaceous schists. High forest.

106 At 0-4 cm: Dark grey (10YR 4/1) very friable sandy loam. Fine crumbly, composing very weak fine subangular blocky. Some tinyflakes o fmica . Transition clear. Aw 4-20 cm:Ligh t grey (10YR 6/1) friable heavy sandy loam, with commonfine distinc t mottles of brownish yellow (10YR 6/8). Weak medium subangular blocky. Transition gradual. B2» 20-90 cm:Ligh t grey( N7/0 ) firmclay ,wit h manyfine t omediu m sized distinct mottleso fre d (2.5YR 5/8)an d yellowish red(5Y R 5/8). Moderate medium subangular toangula r blocky. Common very faint clay skins. Transition gradual. Bw 90-130 + cm: Light grey (N 7/0) firm light clay, with many fine tomediu m sized distinct mottles ofyellowis h brown (10-7.5YR5/8) .

Profile 17 (GROUND WATER LATERITE soil, intergrade to LITHOSOL - partly truncated?) Field description 145 (Sombroek) Lower Amazon region, Monte Alegre, centre ofDom e (Lat. 1°.57'S, Long. 54M0'W) Low part of gentle undulating terrain. Imperfectly drained; upper phreatic level 0.0 mca., lowes t phreaticleve l —1.0m ca. Devonia nfine sand-stone . Grassesan d manyshrubs ,locall ycacti .

A,-At 0-15 cm: Grey brown (10YR 5/2) friablefine sand y loam, with commonfine fain t mottleso f yellow (10YR 7/6). Transition abrupt. B, 15-30 cm: Reddish yellow (7.5YR 6/8)friabl e clay loam. About 80%loos e andhar d small (0.5-2 cm diam.)fine graine d very dusky red plinthite concretions. Transition abrupt. B-Q 30-80 cm:Whit e( N8/0 ) veryfirm cla y loam, with many medium sized prominent mottleso f dusky red (7.5R 3/4) and some yellow (10YR 8/6). Transition abrupt. R 80+ cm: Stratified hard sand-stone; upper part broken and ferruginous.

Profile 18 (Solonetzic HUMIC GLEY soil, intergrade to GROUND WATER LATERITE soil); (GROUND WATER LATERITE soil, Humic phase) Field description 178(Day , Sombroek) Maraj6 island, 8k mN Wo f Arariuna (Lat. 0°.59'S; Long. 49°.60'W) Flat lowland. Imperfectly to poorly drained; upper phreatic level +1.0m ca. (rive r water, almost without load),lowes t phreaticleve l —4.0m ca. Early Holocene marine-deltaic(? )sediments . Grasses and occasional shrubs.

Ai 0-15 cm: Very dark grey brown (10YR 3/2)firm cla y with some strong brown (7.5YR5/8) . Moderate medium sized prismatic, breaking into moderate coarse subangular blocky. Transi tion abrupt and smooth. A„ 15-20 cm: Grey brown (10YR 5/2)friabl e clay loam, with common fine distinct mottleso f yellowish red(5Y R5/8) .Moderat e medium sized subangular blocky. Transition abrupt and smooth. B„-C 20-55cm: Dar k grey(10Y R 4/1)firm clay ,wit h manyfine distinc t mottleso fre d (2.5YR 4/6). Weakfine prismatic ,breakin gint ostron gfine subangula rblocky .Transitio nclea ran dsmooth . Bjij-C55-100 cm: Light grey (10YR6/1 ) very firm clay, with many fine distinct mottles ofre d (2.5YR 5/6) and reddish yellow (7.5YR 6/8). Strong very coarse prismatic. Clay skins promi nento n horizontalpe d surfaces,fain tt odistinc t on vertical pedsurfaces .Transitio n diffuse and smooth. Ba«»-C100-180 + cm: Light grey( N6/0 )firm clay , with manycoars e prominent mottleso fdar kre d (10R 5/6) and reddish yellow (7.5YR 6/8). Weak tomoderat e coarse prismatic, composed of strong coarse angular blocky. Clay skins prominent on both horizontal andvertica lpe d surfaces.

107 The main factors that cause the variations, as illustrated in the profile descriptions, areth e following: 1. the character of thefluctuation o f the ground water level (or pseudo ground water level,o rsaturatio n zone),whic hfluctuation ma yvar yi ndept h and frequency. 2. thetextur eo fth eparen tmaterial . 3. thedegre eo fpre-weatherin go fth eparen t material.

Ad 1.Th eplinthit e formation maytak e place at relatively shallow depth (theProfile s 6, 8, 15),o r relatively deep(Profile s 4, 10, 12, 14).I n the latter case,littl e influence of the ground waterfluctuation ma y befel t inth e upper horizons;th e A2o f the Ground Water Laterite profile mightthe n betake na sfo r instanceth eA 3o r the Bt horizon ofa shallow KaoliniticYello wLatoso le.g. Kaolinitic LatosolicSan d (Profiles 10, 12). Thethicknes so fth elaye ro fsof t plinthite(th eB horizon )depend so nth ethicknes so f the zone of fluctuation. The lower boundary was established only in a few instances (Profile 10); in the other cases no 'pallid zone' proper was reached during the field studies. It is evident however, that the upper boundary does not coincide with the upper levelo f the ground water. At placeswher ethi s levelreache sth e surface tempo-

Foto14 SoloLaterila Hidromorfica, desenvolvi- doem sedimentos arenosos. Nesteperfil, os mosqueadosvermelhos do horizonte B de 'plinthite' macio,ocorrem comofaixas verticalmentearran- jadas. O horizonteA arenosoiixiviado tern cor relativamente escura em sua seccao centra/, devida a alguma cumulacSode humonaquele lugar (fase inicial de formacao de Podzol Hidromorfico dentro do horizonteA de solo Laterila Hidromorfica)

Photo14 A Ground WaterLaterite soil, devel opedon sandy sediments. In thisprofile, the red mottles of thesoft plinthitic B horizon occuras vertically arrangedstripes. The leachedsandy A horizonis relativelydark coloured especially inits central section, due to some humus accumu lation there (initial stage of Ground Water Podzol formation within the A horizon of a GroundWater Laterite soil)

108 rarily, or the terrain is evenshallowl y covered with rain water for a part of the year (Profiles 2,4 ,6 , 8, 11, 14, 16,18),th e soft plinthite layernevertheles s only starts from some depth on (0.5-1.5 m). One may expect the thickest layers of soft plinthite up stream of the main rivers, where seasonal differences in river level are greatest (Soli- moes, Madeira, Araguaia). Also,sinc enea r riversth e fluctuation of the phreatic level isgenerall ylarge rtha n onwatershe d parts,on e mightfind thicke r layersa tth e former locations.Apparentl y however Ground WaterLaterit esoil spredominat ei nwatershe d regions (cf.IV .1.2.2) .Nea r the rivers the zone offluctuation i s probably usually too deep to cause plinthite formation (good drainage), or the sesquioxides are carried off toth eriver . Ad 2. Understandably, the textural differentiation between the A and the plinthitic B horizon, as expressed in the textural ratio B/A, depends on the texture of the parent material. If this texture was light throughout, then the plinthitic Bhorizo n of thegeneticall y ultimate profile would berathe rligh ttexture d (Profile 10).I fth e parent material was a clay,the n also the ultimate A horizon would be at least rather heavy textured1(Profil e 8).I t seemstha t the upper boundary of the plinthite layer,wit h the same drainage condition (for instance the ground water leveltemporaril y at the sur face, or standing rain water), depends on the texture of the parent material. If this is sandy, then the plinthite develops relatively deep (Profiles 4 and 14 versus 8). The pattern of mottling seemst o depend on both texture and structure of thesoi l material insitu (cf.U.3.2.2). Ad 3. Per definition, the 'modal' Ground Water Laterite soil isa strongly weathered soilan d itscla ymineral sar e'lateritic' ,b yconsistin g only ofsilicat ecla ymineral swit h 1:1 latticestructur e (kaolinite),an d sesquioxides (gibbsite,goethite ,hematite) .Suc ha 'lateritic'characte r ofth ecla yfractio n showsu pi n the SiO^ALOg ratios (Kivalues) , whichshoul d beabou t orbelo w2 . The parent materialca n bestrongl y pre-weathered('rego-genetic 'material) ,th e only mineral constituents being quartz, kaolinite and sesquioxides. In this instance, the 'modal' Ground Water Laterite soil can be established relatively rapidly, since only the concentration of the kaolinite and the sesquioxides in the Bhorizo n is involved. This is likely to be the casewit h allth e Ground Water Laterite soils of the Planfcie. They havea Kivalu eo f about or below 2,a low silt/clay ratioi nth e Bhorizo n and a lowcatio nexchang ecapacit y(Profile s 2,4,7,8,9,10,12). Profile 5ca n also be called a 'modal' Ground Water Laterite soil. X-ray diffraction wasapplie d toth elowe rpar t ofth e B horizon ofthi sprofil e (cf sample96- 5o fTable s 7 and 8, and Fig. 14a-e) . The clay fraction of this sample is practically completely kaolinitic (92%) , with only 6% hematite and 2% quart z present. The percentage of sesquioxides is understandably larger in the coarser fractions. They consist however onlyo firo noxide s(hematit e8-1 5%) ,no taluminu m oxides. That inth efraction s 2-16

') Thetextur e of theplinthiti c Bhorizo n isofte n estimated too heavyi n thefield, du et o thedensenes so f thehorizon .

109 micron and 16-80micro n also some other minerals,notabl y mica, are present is pro bablyrelate dwit hth efac tthat ,i nth eare ao fth eprofile ,th ecrystallin ebasemen t(mic a schists)i softe n almostoutcropping .Th ecompositio n ofth ecla yfractio n ofth esampl e iscomparable , to a large extent, with that of the Bhorizon s of the Kaolinitic Yellow Latosol profiles (samples 233-3, 303-4, 300-4 and 210-4 of Table 8). An electron micrograph (Photo 10) however shows that the kaolinite crystals are less well developed than in these latter. The difference in crystallisation also shows up in the different valuesfo r specific surface of the clayfractio n of samples 96-5an d 303-4res pectively(Tabl e7) .

When the parent material isrelativel y rich, easily weatherable primary minerals and silicate clay minerals of 2:1 lattice structure forming part of it, then the formation of the'modal ' Ground WaterLaterit e profile requiresbot h aweatherin gt okaolinit e and sesquioxides,an da regroupin g inth eprofile .Thi si slikel yt otak emor etime .Apparen - ly, the modal type is not therefore reached on the Holocene sediments nor on some Paleozoic-Mesozoic deposits. Lowlands which are still being enriched (flooding by river water with a load of sediments,dgua branca, cf.1.4.4 )usuall y have Low Humic Gley or HumicGle y soils.O n thefollowin g sites,however , the formation of plinthite wasinvariabl yobserved : 1. The lowlands intermittently covered with only rain water (except for those ineas tern-central Marajo islandwhic har ealkaline) . 2. The lowlandsflooded wit h riverwate r without load of sediments (dguapreta, dgua limpa, cf. 1.4.4). 3. The terrains,alon g rivers, only slightly above high water level(massapes, cf. 1.4.4).

Examples are the Profiles 3, 11, 13an d 18.Th e weathering in these profiles is often not yetver ystrong . Kivalue sar e stillabov e2.0 ,whic hindicate s that alsonon-kaolin - iticsilicat ecla ymineral sar epresent .Thi si ssubstantiate d bya sampl eo fProfil e 18,t o which X-ray diffraction wasapplie d (sample 178-4o f theTable s 7an d 8).I n addition to kaolinite (35% )an d sesquioxides (11%),th ecla yfractio n contains also mica(illite , 10%),intermediat e (16%) and somechlorit e (3%).I n the coarserfraction s moreover some felspar ispresent .Th especifi c surface ofth e samplei sstil lhig h(Tabl e7) . That the weathering in the profiles under discussion is still not very strong is also shownb yth e fact that often no pronounced textural difference betweenth eA an d the Bhorizon s hasdevelope d asye t(Profile s 3,11),an dtha ttextura ldifference s withinth e parent material,du et o stratification, areofte n notye tquit eerase d(Profil e 18). The Ground Water Laterite soils on relatively rich deposits of Paleozoic-Mesozoic age, and on peneplained Pre-Cambrian basement have a large textural difference between A and Bhorizons . Ki values are, however, still considerably above 2(Pro files 6,14,15,16, 17). Alsoth esilt/cla yratio sar eusuall ystil lhig h(Profile s 14,15).

Whatever the parent material, the ultimate stage of Ground Water Laterite soil for mation is believed to be represented by the following profiles (cf. GL in Fig. 15): A

110 thick, bleached and very sandy A horizon, grading sharply into a thick, relatively heavy textured, dense, slowly permeable Bhorizo n of soft plinthite. At the transition zone of the two horizons, some plinthitic material occurs that has already hardened. Such profiles, in this publication called Ground Water Laterite, Low phase, generally havea poo r vegetative cover(cf. IV.1.2) . Anylowerin go fth edrainag e basean d chan ge in climate will result in erosion of the A horizon, followed by hardening of the truncated soft plinthiteo fth eB horizon.

II.3.2.2 Fossil Plinthite INTRODUCTION In Amazonia, it is common to encounter layers of hard or soft plinthite on well drained sites, where the phreatic level occurs at many metres depth, and a shallow pseudo ground water leveldoe s not exist. In view of the above described present day formation of plinthite in Amazonia, and in agreement with the majority of the recent literature on the subject, such plinthite is considered to be fossil. It was formed in timeswhe nther eexisted ,in situ o r inth e surroundings, a land surface with a fluctu atinggroun d waterleve la tshallo wdepth . Fossil hard plinthite may occur underlain by fossil soft plinthite, or as an unconfor mable layer on top, or between non-concretionary sediments without any mottling. Fossil soft plinthitewithou ta distinc tcappin go ffossi l hard plinthitewa srarel y found.

Layerso ffossi l hard plinthiteoccurrin galone ,a scrust so nto p of, or asstone-line so f varyingthicknes swithin ,non-concretionar ysediment swithou t anymottling ,wer eno t formed in situ,bu t carried along, alluvially or colluvially, from other places.Absenc e of a regularity in the arrangement of the concretionary elements, and a more or less rounded form of the elements, are complementary in establishing their alluvial-collu- vialorigin . However,th eapparentl y mostfrequen t situationi stha t alaye ro ffossi l hard plinthite is underlain by a layer of fossil soft plinthite. In these instances, it is likely that both plinthiticlayer sconstitut eth erelic so fa Groun dWate rLaterit esoi l- oro fit sgeologi c relative, if the formation had taken place below the solum. After the conditions of imperfect drainagecease d to exist theprofil e waserode d geologically to the point that all the light textured and loose A horizon, and possibly part of the heavy textured plinthitic B horizon,wa sstrippe d off. Whilstth elowe rpar t ofth eplinthiti c B (and the C)horizo nha sremaine d soft,th euppe rpar tha sbecom ehar dplinthite .Th ere dmottle s changed and became stony elements (pisolithic pieces, vesicular blocks or massive plates,dependin g upon the original pattern of mottling),b ydessicatio nthroug h alter nate wetting and drying at the surface. The whitean d yellow parts of the densemate rialchanged , under influence of roots and soilfauna , to become friable homogeneous yellowish or reddish earth. It isnormall y only a lack of contact with the atmosphere which preventsth e hardening of that section ofth e plinthitic layer which isbelo w the present day crust. This wasverifie d at many deep cuts excavated for the BR-14high way,whic h wereexamine d during theconstructio n of the road, as wella sa t intervals

111 Foto15 O processo de endurecimeniode 'plinthite'. Apds ser exposto por menos de dois anos, osub-solo original de 'plinthite"macio do cone de estrada tern endurecido consideravelmente. Foi levado pela dgua o material caolinilico da matriz de corbranca, enquanto osmosqueados vermelhos endurecenles - aquiempadrdo laminar - ressattam docorte. Na parte superior do corte pode ver-seo 'plinthite''duro como era antes da ex- cavac&o. Partes deste 'plinthite' duro ainda mantem urn arranjolaminar (Tipo de concrecbes Mde do Rio; BR-14,km 40m. ou m.)

Photo 15 The hardening process of plinthite. After beingexposed for less thantwo years, the originalsoft plinthiticsubsoil of thisroad embankment has hardened to a considerable extent. Thekaolinitic materialof the whitecoloured matrix hasbeen washed away, while the hardening red mottles - here in a platy pattern - pro trudefrom theface of the embankment. At theupper mostpart of the embankment, the hardplinthite canbe seen as it existed beforeexcavation. Parts of this hard plinthite still have a platy arrangement (Mde do Rio typeof concretions, BR-14,km 40 ca.) during the2- 3 yearsafterwards . Whileth eplinthit ewa ssof t duringth eexcavating ,th e wandso fth ecut sbecam egraduall y hard asstone 1(cf. Photo 15). The sequence: fossil hard plinthite - fossil soft plinthite, normally indicates that the material wasforme d in situ.Bu t it iso fcours e possibletha t the hard plinthite,o r part ofit ,wa scarrie d alongan d deposited onto p ofth eplinthit ewhic hwa sforme d insitu. Certainty isobtaine d about thein situ originatio n oflayer so ffossi l hard plinthiteover lyingsof t plinthite ifth e sequence hasth efollowin g features: 1. The character of the concretionary elements (size,form , grainage) isrelate d to the character ofth ere d partso fth esof t plinthite 2. The arrangement of these elements in the lower part of the hard plinthite layer is similart oth epatter no fmottlin gi nth esof t plinthite 3. Thetransitio n between hard plinthite and soft plinthite isgradual . Afrequentl y occurringsituatio n istha t fossil soft plinthite iscappe d by related fossil hard plinthite which in itstur n isoverlai n bynon-concretionar y sediments of varying thickness. Such a situation might be confused with present-day Ground Water Late- rite formation, especially if the hard plinthite forms a layer near to and parallel with

') It islikel yhowever ,tha t the types of soft plinthite which have relatively high Ki values, for instance those constituting the Zersatzzone of consolidated sediments and rocks, need considerably more than a few yearstim e for hardening, if they harden at all.

112 thesurface . But,a sstate d above,layer so fhar d plinthite are infrequent inpresent-da y Amazon Ground Water Laterite soils and arethe n alwaysthin . Oneca nestablis h the fossil character ofth eplinthit efro m theabsence ,directl yabov eth ehar d plinthite,o fa relatively light textured layer with colours of little chroma and high value, and/or mottling(pedologicall y anA 2 horizon). Sucha narrangemen tof : sediments- fossil hard plinthite- fossil soft plinthite,ca n be explained byassumin gtha t after truncation and hardening ofa forme r Ground Water Laterite soilprofile , the stony surface wasburie d with sediments.A disturbance of an original arrangement of the concretionary elements in the upper section of the hard plinthite layer isals oa sig no ftruncatio n and subsequent burying. Itma ystil lb enote d that,wit hth esequence :sediment s- hard plinthite- soft plinthite, therei sa goo dpossibilit yfo r newGroun dWate r Lateritesoi lformatio n inth ecoverin g sediments.Thi s isth e case when the layers of fossil plinthite are impervious, thereby causing apseud ogroun d water level. Thisne wsoil ,i nit sturn , maybecom e truncated, resulting in another sequenceo f fossil plinthite aboveth e earlier one (cf.1.4.5) .I f the zone of fluctuating ground water level is located within the layer of fossil hard plin thite,the n thecondition s for the formation ofplinthiti c conglomerates are present. The above considerations on the origin of the sequence sediments- hard plinthite - soft plinthite holds for Amazon conditions. It is quite conceivable that in semi-arid regions,fo r instance in North-Eastern Brazil,th e sequence may constitute the whole of a non-eroded fossil Ground Water Laterite profile, because there the soil may fre quently dry out to a great depth, and the original A2ma y loseit saspec t ofIixiviation . Amazon fossil plinthite not truncated at sometim e in the past, and consequently not hardened over more than a few decimetres, is apparently uncommon. In the central part ofth eGuama-Imperatri zare a however, awell-draine d soilwa sfoun d which may be a fossil full Ground Water Laterite soil profile. The soil, classified under Red YellowPodzolic soil, intergrade toKaolinitic YellowLatosol (cf. for instance Profile39) , is found on terraces that are 30-10 m above the level of local rivulets. The soil has a yellowish (hue 10YR) and friable A horizon and a reddish (hue 5 YR or 2.5YR), ratherfirm an d heavier textured B horizon. Betweenthese ,a transitiona l horizon(AB ) occurs,20-5 0c mthick , ofwhic hth e upper part hasrathe r often small(1- 2c mdiam.) , hard, round plinthite concretions which comprise about 5-20% of the soil mass.Th e lower part ofth etransitiona l horizon has,withi n a matrix of yellowish hue,common , medium sized and distinct mottles of yellowish red or red (5YR, 2.5YR).Th e indica tionsar etha tth eparen t materialo fth eA ,A Ban d B isidentical .Th efeature s described therefore make it likelytha t an originally plinthitic horizon has hardened init s upper part,whil ei nth elowe r part themottlin gha sbee ndiminishe d byoxidation . Genetical ly speaking, the profile is likely to constitute a 'raised*, but not truncated Ground Water Laterite soilprofile ,i nw hic hth eformatio n of plinthitewa slimited . Inth eforegoin gdiscussio n it wasassume d that the soft plinthite1belo w a capping of

*)Th echroma san dvalue so f thematri x of soft plinthitewel labov eth egroun d waterleve l are usually less low and less high respectively than those of the matrix of soft plinthite in the Bhorizo n of present-day Ground Water Laterite soils.

113 fossil hard plinthite is all of fossil character. However, the layer of soft plinthite is sometimes of ten and more metres thickness. The lowest section of such thick layers has only a concentration of sesquioxides into mottles (the form of which is often directlyrelate d toth estratificatio n ofth eparen t material;cf. Phot o 16),no t aconcen tration of silicate clay minerals. This lowest part, after exposure, maintains also a rathersof t consistence.I t isimprobabl etha t atth etim e of Ground Water Laterite soil formation the fluctuation of the ground water level was so great that all of the layer

Foto16 A seqiiencia de 'plinthite' durofossil e 'plinthite' maciofossil queconstitui o solo Podzdlico Vermelho-Amarelo, 'intergrade'para Latosolo Amarelo Caolinitico,fase Concreciondria. A camada de 'plinthite' duroque forma a parte superior doperfil secompoe deelementos concreciondrios discretos em terrafridvel ndo mosqueada. A camadade'plinthite' macio que constitui aparte inferiordo perfil, secompoe de material argiloso fir me e denso com mosqueadosprominentes e grossosde matiz vermelho numa matrizbranca oucinzento claro.Neste perfil os mosqueados tim urn padrdo vesicular-prismdtico eos elementos concreciondrioss3o dotipo lpixuna. As separacoes ovais dentro do horizonte de 'plinthite' macio sdo compostas de terra ndo mosqueada fridvelsem concrecoes. Provdvelmente resultamde homogenizacdo local do 'plinthite' maciopor cupins (BR-14,km 115 m.ou m.)

Photo16 The sequence offossil hardplinthite and fossil soft plinthite thai constitutesa RedYellow Podzolic soil, intergrade to Kaolinitic YellowLatosol, Concretionaryphase. The layer of hardplintite which makesup the upper part of theprofile, consistsof discrete concretionary elements in unmottled friable earth.The layerof soft plinthite which makesup the lowerpart ofthe profile, consistsof dense and firm clayeymaterial with aprominent coarse mottlingof red hue in a white orlight grey matrix. Inthis profile,the mottling has a vesicular-prismaticpattern and the concretionary elementsare of the lpixuna type. Thepockets and pipes that arevisible in thelayer of soft plinthite consist of unmottled, friable earth without concretions. Theyare probably the result of local homogenisation of the soft plinthite by termites (BR-14,km 115ca.)

114 wasthe nformed . Thethicknes so fth esof t plinthitelaye rma yb edu et oa ver ygradua l lowering of the ground water level,relate d to a very gradual lowering of the drainage base.I n viewo f thefac t that the layer seemst o bethickes t at the rims of crust-capped plateau land (cf. 1.4.5, note on page 44), it is however more probable that, once a sequence of hard and soft plinthite is present in a welldraine d position, the soft plin thite may enlarge itself in a downward direction under influence of obliquely down ward movinggroun d water.Fo rsuc hpresent-da yaccretio no ffossi l plinthitea fluctua ting phreatic level proper would therefore not be necessary.

From the descriptions in 1.4.5, it can be concluded that the fossil hard plinthite of Amazonia rarely constitutes massive, impenetrable and thick slabs (hardpans,cara paces), asreporte d for many areasi n Africa, for instance N.Nigeria . Usuallyther ear e only concretionary elements between looseearth . This general weakness of the devel opmento fth eAmazo nplinthiti ccrust si sbelieve d to berelated ,i npart ,t oa genera llo w Fe content of the mainAmazo n sediments,whic hals oaccount sfo r thelo wiro noxid e content of the main Amazon Latosol (cf.11.2.2) . Thevariatio n inth e characteristics ofth e concretionary elements of fossil hard plinthite, as shown in 1.4.5,i srelate d to a variation of patterns of mottling ofth e original soft plinthite.Th e latter isbelieve d to have depended upon differences in texture and in stratification of the sediments, or upondifference s instructur eo fth esuperficia l layers(i.e. th esoil )before , orduring ,th e time when conditions of imperfect drainage caused the plinthite formation. It was discussed in 1.4.5tha t the Paragominas type of concretions is very likely connected with the structural constitution of the Belterra clay. The reticulate-prismatic pattern of the Ipixuna type may bedu e to a blocky to prismatic structure of the B horizon of theorigina lsoil ,befor e ora tth etim ewhe ncondition so fimperfec t drainage occurred, ort oactio n ofroot san d soilfaun a inthi slaye rdurin gth eplinthit e formation. Forth e Maed o Riotype ,a norigi ni nnon-homogenize d stratified sedimentsi splausible ,espe cially for those elements that are platy and alternating coarse and fine grained, or coarse grained throughout.Th emor emassiv easpect ,compare dwit hothe rtypes ,ma y indicatea formatio n below the solum, lateral enrichment with sesquioxides, or both.

FOSSIL PLINTHITE BELOW THESOLU M Fossilplinthit eoccurrin gonl ydee pbelo wth esurface ,arbitraril y below 1.50 m,i sno t diagnosticfo rclassificatio n ofth esoi la tth espot ,unles sa ta ver ylo wcategorica l level. At several places such a situation was observed. At the town of Belem fossil hard plinthite is found from about 5m depth on, under sandy sediments that constitute a Pleistoceneterrace .Nea r Itacoatiara,thi nband s(3 0cm )o ffossi l hard plinthite,i npar t underlain with fossil soft plinthite (150cm) ,ar e found at a local depth of 8m below the surface (ETA-54seringal, near river Urubii).O nth e planalto terrace in the region south and southeast of Santarem, fossil hard plinthite is found at about 15m below theleve lo fth eflat centra l sections.Als oi nth eGuama-Imperatri zare a such situations are common. In these instances the solum is usually a KaoliniticYellow Latosol (KYL), which

115 Fig.15 Esbocode algunssolos amazonicoscom ''plinthite'(para os simbolos, vej'a Tabela9)

RP-KYL.CR KYL,CR

Sediment or rock Sedimento ourocho Fig. 15 Sketch of someAmazon plinthitic soils (see for symbols Table 9) f>f/ll developed on the overlying sediments. Imperfectly drained soils may however occur, also where the land surface iswel labov eth eleve lo fth eloca lrivers ,du et oth efac t that the fossil plinthite can form an almost impervious substratum. Intha t caseth e profile constitutes normally arecen t Ground Water Laterite soil(cf. Profil e 8 ofII.3.2.1) .

FOSSIL PLINTHITE WITHIN THE SOLUM Fossilplinthit e atth e surface or at ashallo wdept h (lesstha n 1.50 marbitrarily ) is, of course,a n important characteristic of the soilin situ.I n fact, the plinthite now actsa s

116 parent material for a soil. Thecharacteristic s of the soil depend onth e degree of weathering ofth e original soft plinthite. Asdiscusse d before, thefinal produc t of this weathering isth ehar d plinthite,whic hi nAmazoni a consists,wit hfe wexceptions ,o fa mixtureo fhard , concretionary elements,an d loose, friable earth. Theconcretionar y elements areinactive ,whils tth e earth is,i nfact , similart o that oflatosoli c profileso n non-concretionary sediments in the surrounding area (kaolinitic character ofth ecla y fraction; Si02:ALj03molecula r ratios = Kivalue saround , or slightlybelo w2) . Thisweatherin gma y beonl yrelativel y shallow, i.e. thedens elaye r of soft plinthitei s still present neart oth esurfac e (arbitrarily lesstha n 1 m).I ntha t caseth eactua l solum consists ofa friable, concretionary topsoil (Ahorizon ) with subangular blockyo r granular structure, and adense ,firm subsoi l (Bhorizon ) with predominantly angular blocky structure and presence ofa few clay skins. Because ofth e morphometric field characteristics ofth e Bhorizo n thesoi lcanno t becalle da Latosol ,althoug h Kivalue s are around orslightl y below 2.Suc h soils are classified asRed Yellow Podzolic soil, intergrade toKaolinitic YellowLatosol, Concretionaryphase (RP-KYL, CR). If the weathering of the soft plinthite has progressed to agreate r depth from the sur face (arbitrarily to more than 1 m),the n practically thewhol e solum, both the A and the B horizons,consist so ffriable , concretionary earth.Therefore , thesoi lca nb eclas sified asa Latosol: Kaolinitic Yellow Latosol,Concretionary phase(K.YL , CR). This classification hast o begiven ,o fcourse ,als ot othos econcretionar y and friable soils,i n which soft plinthite iscompletel y lacking and theconcretion s wereno t formed insitu, but are ofcolluvial-alluvia l origin (stone-lines). Athicknes s ofth e stone-line ofmini mally 30c mabov e 1 mdept h istake n asa criterio n for aConcretionar y phase.

Profile 19 (RED YELLOW PODZOLIC soil,intergrad e to KAOLINITIC YELLOW LATOSOL, Concretionary phase - type ofconcretions : Ipixuna) Field description 202 (Sombroek) Guama-Imperatriz area, km 109.6 (Lat. 2°.35'S; Long. 47°.28'W) Approx. flat terrain; terrace, 55m abov e level ofloca l rivulet. Slightly imperfect drainage. Pliocene sediments,probabl y formerly covered with Belterraclay .Hig h forest, with rather high timber volume.

A, 0-10 cm: Brown (10YR 5/3) friable clay. Moderate fine subangular blocky. About 25% loose hard plinthite concretions, medium tosmal l sized (10-0.5c m diam.),prismatic , predominantly finegrained , dusky red (10R 3/4) and locally light red (10R 6/6).Transitio n clear to gradual. A, 10-70cm: Brownish yellow(10Y R 6/6-8) friable heavyclay . Moderate fine subangular blocky and locally angular blocky. Afe wfain t clay skins, largely at thesurface s of the concretions. About 75%loos e hard plinthite concretions, similar to those inA, .Transitio n gradual. B, 70-120 cm: Reddish yellow (7.5YR 7/6) firm heavy clay, with many medium sized to fine distinct mottles ofre d (2.5YR 5/6) and yellow (10YR 7/8) ina reticulate pattern; the red parts are rather hard. Weak medium prismatic, composed of moderate medium to fine subangular blocky. Common faint clay skins. About 25 % half loose,verticall y arranged, hard orrathe r hard plinthite concretions, similar tothos e in Ai. Transition gradual andirregular . B,-B, 120-350 cm: Reddish yellow (5YR 7/8)ver y firm heavy clay, with many coarse prominent mottles of dusky red(7.3 R 3/4), white (N8/0 )an dyello w (10YR 7/8); thedar k red occurs predominantly as vertical pipes,abou t 5-20c m long,whic h are rather hard. Weak medium to coarse prismatic, falling apart into moderatefine subangula r and angular blocky. Afe w faint clay skins. Transition gradual. C 350-550 + cm:Reddis h yellow (7.5YR 8/6)fir m tover y firm heavy clay, with many coarse prominent mottles ofpal e red(7.5Y R 6/6), white (N8/0 ) and yellow (10 YR8/8) .

117 Profile 20 (RED YELLOW PODZOLIC soil,intergrad e to KAOLINITIC YELLOW LATOSOL, Concretionary phase - type ofconcretions : Mae do Rio) Field description 238 (Sombroek, Sampaio) Cf.III.2 , Profile 40.

Profile 21 (RED YELLOW PODZOLIC soil, intergrade to KAOLINITIC YELLOW LATOSOL, Concretionary phase - type ofconcretions : Paragominas) Field description 205(Sombroek , Sampaio) Guama-Imperatriz area, km 175.6 (Lat. 3°.06'S; Long. 47°.18'W) Side of low hill in very gently undulating terrain. Slightly imperfect drainage. Plio-Pleistocene lacustrine (Belterra clay) andolde r sediments; upper part ofprofil e probably colluvial. High forest, of low timber volume. A, 0-5 cm:Brownis h yellow (10YR 4/2) very friable heavy sandy loam. Weak fine subangular blocky. About 60%loose , hard plinthite concretions, very small sized (0.5-1.0c mdiam.) , subangular blocky, fine grained, dusky red(7.5 R 3/4). Transition clear. A2 5-80cm: Brownis h yellow (10YR 6/6)friabl e heavy fine sandy clay. Weak medium to fine subangular blocky. About 80% loose, hard plinthite concretions, similar to those in Au but somewhat larger (0.5-2.0 cm diam.). Transition gradual. Bi 80-120 cm: Reddish yellow (7.5YR 6/8) friable to firm clay with common medium sized distinct mottles of red (2.5YR 5/8),predominantl y in a fine polygonal pattern. Moderate medium sized subangular and angular blocky. Common faint clay skins,largel y atth e surface of the concretions. About 80% loose, hard concretions similar to those ofA, , but somewhat larger (0.5-2.5 cmdiam.) . B2 120-220 cm: Light red (2.5YR 6/8) firm clay, with common to many medium sized faint mottles of light red (7.5Y R 6/6) ina fine polygona l pattern, and half hard. Weak medium sized subangular and angular blocky.Abou t 50%hal f loose, hard orhal f hard plinthite concretions, similar to those in theA, ,but somewhat larger (0.5-2.5 cmdiam.) . Transition clear. B3 220-280 + cm: Light red (2.5YR 6/8) friable to firm clay, with common coarse distinct mottles ofpal e red (7.5R 6/4),an dwhit e (N8/0). Massive, to weak medium sized subangular and angular blocky.

Profile 22 (KAOLINITIC YELLOW LATOSOL, Concretionary phase) Field description 114 (Day) Cf.III.2 , Profile 28.

Profile 23 (KAOLINITIC YELLOW LATOSOL, medium textured - over bank of fossil plinthite) Field description 337 (Commissao deSolos ) Amapa Territory, along road Macapa-Fazendinha (Lat. 0°.0'; Long. 51°.08'W) Flat terrain, about 4 m above local river level. Well-drained. Pleistocene sediments. Grasses and scattered shrubs; yearly burned. A! 0-15 cm:Ver ydar k greyishbrow n (2.5Y 3/2) friable sandy clay loam. Weak very fine to coarse granular. Transition clear and smooth. A3 15-30cm: Olive brown (2.5Y 4/4) friable heavy sandy clay loam, with many fine faint mottles of very dark grey brown (2.5Y 3/2). Weak very fine subangular blocky. Transition diffuse and smooth. Bi 30-50 cm: Yellowish brown (10YR 6/5)friabl e sandy clay. Little coherent massive, easily falling apart into fine earth. Transition diffuse and smooth. B21 50-75 cm: Olive yellow (7.5Y 6/5) friable light clay. Structure asi nB ^ Transition diffuse and smooth. B22 75-110 cm:Brownis h yellow (10YR 6/6) friable light clay. Structure as in B^ Transition abrupt andirregular . B2U 110-190 cm:Sam e material as B28 but with about 80% loose, large (upt o 20c m diam.), somewhat rounded, vesicular plinthite concretions, fine andmediu m grained and variegated reddish coloured. *»

118 The foregoing profile descriptions,an d their analytical data (Appendix 9), illustrate the variation in the characteristics of wello r moderately well-drained soilswit h fossil plinthite. Another profile of Red YellowPodzoli c soil, intergrade to Kaolinitic Yellow Lato- sol, Concretionary phase (RP-KYL, CR), located at km 30o f the Guama-Imperatriz area, wasals o studied. A sample of the soft plinthite of thedee p subsoil, beneath two separate layers of fossil hard plinthite, was analysed by X-ray diffraction (cf.sampl e 302-4o f Tables 7an d 8,an d the Figs. 14 a-e). Thecla y fraction wasfoun d to bepre dominantly kaolinite (85%), iron oxides comprising only 12%an d clay sized quartz 3% .I nth ecoarse rfraction s moresesquioxide sar epresen t(30-3 5%) ,bu tthe yar eonl y hematite. This mineral apparently kept enclosed much kaolinite (35-55%). In the 2-16micro n and 16-80micro n fractions moreover somemic a(7% )occurs .Th especi fic surface of both thewhol e soilan d thecla yfractio n isver ylow .Th ecla yfractio n is in fact quite comparable, also in itschemica l activity, to that of Bhorizon s of Kaoli niticYello w Latosol profiles (samples 233-3,303-4 ,300-4 ,an d 210-4o fTabl e7) . In the Guama-Imperatriz area practically all concretionary soils contain soft plin thitea tshallo wdepth .The ytherefor e havebee ngroupe d asRP-KYL , CR.Th eimpres sion is that the same situation exists in the majority of the concretionary soils else where in Amazonia. However, those near Belem (Bragantina area) and Amapa (Savannah area) are, for a part, deeply friable: KYL, CR. It is often difficult, during surveys,t oestablis h theuppe r levelo fth esof t plinthite,becaus eo fth estonines so fth e surface. Itca n only beeasel ydon ewher erecen t deeproa d cutsar epresent , aswa sth e case in the Guama-Imperatriz area.Besides ,fo r actuallan d useth edegre eo fstonines s may be more important than the morphometric-genetic characteristics of the subsoil which determine the classification. The genetic relationship between the Ground Water Laterite soils and the above described concretionary soilso fAmazoni a wasth ereaso nwh yth elatte rwer eprovisio nallyterme d'Groun dWate r Lateritesoil ,Truncate d phase'(cf DAY, 1961:SOMBROEK , 1962b). Such a term however refers to the genesis of theparent material of the soils, not that of the present-day soils. Moreover, it suggests a drainage condition inherent into the Ground Water Laterite soilprofile , i.e. intermittently imperfect drainage.Th e soils concerned are however well or moderately well-drained. Also,th e differences in thicknesso fth elaye rwit hfriabl e earthcanno t beindicate d asdecisiv ea si ssecure db y thedistinctio n between Latosol and Red YellowPodzoli csoil . The concretionary soils are, in fact, intergrading to Lithosols. When practically no earth occurs between the mass ofconcretionar y elements,o r the fossil plinthite forms a slab-like hardpan, a classification of the soil as 'Concretionary Lithosol' may be the most appropriate (cf. CAMARGO and BENNEMA, 1962). Such senile soils are, how ever, rare in Amazonia.

A general scheme for the different soilstha t canb edistinguishe dwit h the formation, respectively thetruncatio n and theburyin go fAmazo n plinthite isgive ni n Fig. 15.

119 III. The Soils Classified, and their Geographic Occurrence

III.l TheMethod so fStud y

III.1. 1 FieldStudie s

The soil data compiled in this publication were obtained from field studies by FAO soil scientists during the years 1957-1962,partl y in cooperation with members of the Soils Section of the Instituto Agronomico do Norte (IAN) at Betem. During this time, several reconnaissance soil surveys were executed, largelyi nth eeaster n halfo f the region. Also many observations scattered throughout the region were made (cf. Fig.16) .Togethe rwit hth edat a ofth esurveys ,the yenable d apictur e to beobtaine d of the soilso f Amazonia, and their relative importance. Atota l of about 350ful l profile descriptionswa sassembled . The reconnaissance soil surveys of the Caet6-Maracassume area, the Guama-Impe- ratriz area and the Araguaia Mahogany area (2,3 an d 4o f Fig. 16)wer e executed in cooperation with an FAO/SPVEA Forest Inventory Team,whic h made forest inven tories of the samearea s at the same time.Aeria lphotograph s wereuse d for these sur veys.The ywer eonl ypartl yvertical swit hful l coverageo fth eterrai n (Araguaia Maho gany area). The majority were trimetrogon photographs of poor quality. Moreover, they left gaps in the coverage. For the one area mentioned, aerial photograph inter pretation wasver yhelpfu l for soilmapping . Forth eothe r areashowever , wherevege - tational differences were hardly noticeable, the photographs could only be used to produce little more than topographic-planimetric maps, as prepared by the Forest Inventory Team. It wastherefor e necessary to make comparatively manyfield obser vations. This was economically and technically possible because of the cooperation with the Forest Inventory Team, which cut, for its measurements, a network of ten- kilometrestransect sal love r theare aconcerned . Details on the location of these tran sects may be found on the maps of GLERUM (1960) and GLERUM and SMIT(1962a , 1962b). Thefield studie swer epartl y done on profile pits and partly on boringswit h a Dutch soil auger, the Edelman auger. The profile descriptions were made according to the standardsestablishe d inth e SOIL SURVEY MANUAL (1951). The pH of the soil in the field wasestimate d with adaptedfluids (Soil-te xp H kit, La Motte-Morgan pH kit, or Hellige pH kit). The Munsell Soil Color Charts were used to establish soil colours. Unless stated otherwise, thecolou r noted arethos e of the soil in a moist condition.

120 Fig.16 Os estudos decampo ulilizadospara basedesta publicacao

W W H* 12* M* U»

Reconnaissance soilsurvey s Concentration ofothe r field studies Levantamentos de reconhecimento dossolos Concentracao dos outrosestudos decampo

1. Bragantina (FILHO el al., 1963); 2.Caete-Maracassum e (DAY, 1959); 3.Guama-Imperatri z (SOM- BROEK, 1962a); 4. Araguaia Mahogany (SOMBROEK and SAMPAIO, 1962); 5. Manaus-Itacoatiara (SANTOSet al., in execution/em execucdo)

Fig.16 Thefield studies of Amazon soilswhich were usedas abasis/or thispublication

III.1.2 Laboratorystudie s

Sampleso fabou t 100 ofth estudie dsoi lprofile swer eanalyse din .th elaboratory .Th e analysiso fmos tsample swa scarrie dou ta tth eInstitut od eQuimic aAgricol a(IQA )i n Riod eJaneiro ,Brasil ,unde rth esupervisio no f the chemist Dr. LeandroVettori .Th e methodsemploye d (IQA, 1949)ar e thesam ea sthos e used for thelaborator y analysis of the studies of the national Brazilian Soils Commission (cf. BARROS, DRUMOND, CAMARGOet al., 1958; LEMOS,BENNEMA , SANTOSet al., 1960),an d are asfollows :

PHYSICAL ANALYSIS 1. Apparentbulk density: Weighingo f 100 mlearth ,compacte d ina metalcylinde ro f thesam ecapacity . 2. Realbulk density: Aknow nweigh to ffine eart h( < 2.0 mm), dried at 105°C ,i spu t ina receivero f 50 ml,whic h is thenfilled u pwit habsolut eethy lalcoho L

121 3. Mechanical analysis:Sedimentatio n in a Koettgen cylinder, using NaOH as the dispersion agent. In the samples with moretha n 1 %C ,th e organic material is first destroyed byH 202. For aportio n ofth esamples ,th eseparatio n ismad ei nfou rfraction s accordingt oth e classification of the U.S. Department of Agriculture (fractions 2.0-0.2 mm, 0.2-0.05 mm,0.05-0.00 2m m and < 0.002mm )an d for the other samples according to the In ternationalClassificatio n (2.0-0.2mm ,0.2-0.0 2mm ,0.02-0.00 2m man d< 0.002mm) . Thetextura lclas sname s generally used inth efield ar e based onth e subdivision ofth e U.S. system. In caseth e data are provided according the International Classification, the percentages of the 0.02-0.05 mm fraction are therefore estimated, by graphical interpolation ona summatio ncurve . 4. Texturalrelation BjA: The mean of the clay percentages of the subhorizons of the B(exclusiv e B3),divide d byth e mean ofth e clay percentages ofth e subhorizons of theA . 5. 'Natural' clay: The percentage ofcla yobtaine d by shaking with distilled water. 6. Indexof structure: Obtained bycomparin g 'natural' clay content with clay content after dispersion withNaOH .Inde xo fstructur e = 100X (1-naturalclay/tota lclay )

7. Moisture equivalent, M.E.: Determination accordingt o the process of BRIGGS and MAC-LANE, centrifuging the moistened earth at 1000g durin g 40minutes .

CHEMICAL ANALYSIS 8. Organic carbon, C: Oxidation of the organic material with potassium bichromate 0.4N (metho d of TIURIN). 9. Totalnitrogen, N: Digestion with sulphuric acid, catalized with copper and potas siumsulphate .Afte r transformation ofal lN int oammoni a salt,thi si sdecompose d by NaOH and the distilled ammonia collected in a solution of boric acid 4% whic h is titrated withH 2S04,0.02N . 10. RatioCjN: Dividin g organic carbon by total nitrogen.

11. pH-H20 andpH-KCl: Determination withpotentiomete r ina soi lpast ewit hrati o ofsoi lt owate rapproximatel y 1:1,usin gglas selectrode .

12. Availablephosphorus, P2Ob: Extractionaccordin g to the TRUOG method, (H2S04, 0.01— 0.00 1N + (NH4)2 S04, 0.05 M)usin gth e'Unicam 'colorimeter , or according toth e BRAYmetho d (HC1,0.1N + NH4F,0. 5M) .

13. Attack byH 2SOt (d = 1.47): Under areflu x cooler,2g o ffine eart h areboile d for an hour with 50m lH 2S04, d = 1.47. After boiling,th e material iscooled ,dilute d and filtered intoa receive ro f25 0m lcapacity . a. Si02: Theresidu eo fth esulphuri caci dattac k isboile dfo r halfa n hourwit h20 0m l Na2COs, 5% .Th e mixture isfiltered,, an d ina measure d part ofth efiltrate, th edis -

122 solved silicai sprecipitate d byexces so fconcentrate d H2S04 and heating in asand - bath untilsmoking .Thi ssilic ai sdetermine d gravimetrically. b. Al203, totalaluminum: In 50m lo fth enitrat eo fth esulphuri caci d attack, the other heavy metals are separated with an excess of NaOH, 30%.A measure d part of the newfiltrate i sgraduall y neutralized with HCl and the aluminum determined volu- metricallywit h EDTA. c. Fe203, totaliron: Determination on5 0m lo fth efiltrate o fth esulphuri c acid attack byth e bichromate method, usingdiphenilamin e asth e indicator and tinchloride as thereducer . d. Ti02: Determination in thefiltrate o fth e sulphuric acid attack byth e colorimetric method of H202, after elimination of the organic material by heating with some dropso fconcentrate d KMn04. e. P2Oh, totalphosphorus: Colorimetric determination on thefiltrate o f the sulphuric acidattack ,usin gascorbi caci da sth ereducer ,i nth epresenc eo fammoniu mmolyb - date,sulphuri cacid ,an d bismuthsalt .

14. Values Ki (Si02:Al203 ratio)and Kr (Si02:(Al203 + Fe203) ratio): Calculation on molecular basisfro m the data obtained under 13.I t isassume d that Alan d Fe are determined totally with the sulphuric acid attack as described under 13.Th e data for thesetw oconstituent stherefor e represent thesum so fth eportion s ofF ean d Aloccur ringi nexchangeabl e form, theportion soccurrin ga sfre e sesquioxides- includingcon cretions- and the portions that are constituents of the silicate clay minerals.Th ede termined Sicomprise sal ltha t ofth esilicat ecla ymineral san dals otha twhic hi spresen t infre e colloidalform . Quartz and other primary minerals are not, however, attacked, evenwhe no fclay-size . TheIQ Aha sfoun d thatth edeterminatio n ofth eK ian dth eK ro nth efine eart h frac tion generally givesth e same results as the internationally used determination on the clayfractio n (VETTORI, 1959).Th e Kr may beslightl y different if concretions are pre sent. 15. Exchangeablemetallic cations: a. Thesum ofexchangeable metallic cations, S: Determination by percolation of 12.5 gfine eart h with 250m lnorma l ammonium acetate at pH 7. 100 grams of the per colate are evaporated,1calcinated , the residue dissolved in a measured excess of HCI ,0. 1 N,an dth eexces sdetermine dwit hNaOH ,0.1N .Th evalu eS represent sals o the sum ofth e separately determined C&++, Mg4-1",K + and Na+. b. Exchangeable calcium,CO++, and magnesium,Mg++: A measured part of the solution prepared for direct determination of S(cf. 15 a) isuse d for determination of Ca++ + Mg++ by EDTA, applying Eriochrome as the indicator. In another measured part of the solution, Ca++ is likewise determined by application of the indicator Murexide. With this analysis method, the Ca of possibly present free carbonates isinclude d inth evalu efo r Ca++. c. Exchangeable potassium,K +, andsodium, Mr 1":Direc t determination inth eperco lateo fammoniu m acetate witha flame photomete r (see,however , 16).

123 16. Solublesalts: I fth eamoun t of Na+, determined under 15c,i srathe r high,the n the presence offre e saltsi nth e soilsolutio n issuspected . In that case,th e quantity of free anions,referre d toa s'solubl esalts *o r'CI' ,i sdetermine db ymeasurin gth econductivit y ofth esaturatio n extract. In the presence of soluble salts,th e sum of the metallic cations as determined under 15a represents the exchangeable metallic cations together with the soluble metallic cations. It isassume d that the latter are normally Na+ for a large proportion. In this case therefore, the Na+ in the saturation extract is determined separately. This amount of Na+ issubtracte d from the amount determined under 15c,t o obtain the real amount of exchangeable Na+ on the adsorption complex. The value S, as deter mined under 15a, isals odiminishe d withth eamoun tofNa + inth esaturatio n extract, to obtain the approximately correct sum of exchangeable metallic cations on the ad sorption complex. 17. Exchangeablealuminum, (Al) +: Determination by shaking 10g fine eart h with 200 mlKC1, 1N ,followe d by decantation, and titration with NaOH, 0.1 N, in the pre senceo fbrome-thymo lblue . In fact, this determination givesth e 'active acidity'. Experience at IQA has however shown that the data for exchangeable (Al)+, determined colorimetrically with alumi- non after extraction with KC1,1N , are practically alwaysequa l to the data for active acidity. 18. Potentialacidity, H+ + (Al)+; Determination by extraction with normal calcium acetatea t pH7 . This value represents the 'potential acidity'. Its includes the 'active acidity*, or the exchangeable (Al)+(se e 17).H + alone representsth epH-dependen t acidity. 19. Potentialcation exchange capacity, T: Not determined separately but obtained by the addition of S(se e 15a ) and H+ + (Al)*(se e 18).A s such, it isequivalen t to the potential cation exchangecapacit y accordingth e NH4OAcmetho d at pH 7. 20. Basesaturation, V:Obtaine d bycomparin g Swit hT . V = 100S/T . The value represents the base saturation percentage of the potential cation exchange capacity.

MlNERALOGICAL ANALYSIS Mineralogic analysis by IQA involves only a few selected samples,o f which only the coarse sand fraction is studied. After separation of the minerals in groups with the Brogger funnel, using bromoform and a mixture of bromoform and chloroform as a separationfluid, th e material isanalyse d under thebinocula r loupean d the polarizing microscope.Amount sles stha n 0.5 %weigh t are indicated as'traces' .

Samples of a few selected profiles were analysed in laboratories elsewhere, namely that of UNSF British Guiana Soil Survey at Georgetown, British Guiana, that of the Royal Tropical Institute at Amsterdam, Holland, and that of the Netherlands Soil

124 Survey Institute (STIBOKA)a t Wageningen, Holland. For adescriptio n ofthei r ana lysismethod s cf. PLAISANCEan d VANDE R MAREL(1960) ; DAY etal. (1964).

111.2 TheSoil sClassifie d

The soils described in this publication are classified according to an international system, so as to permit comparison of the Amazon soils with those of other tropical regions.Th eclassificatio n systemfollowe d istha t applied untilrecentl yi n the USA,a s described first by BALDWIN, KELLOGG and THORP (1938), and revised by THORP and SMITH(1949) .Th edetail s of theclassificatio n of the zonal tropical soilsagre ewit h the adaptations and elaborations made on the USA system by the national Brazilian Soils Commission (cf. II.2).Th e 'Guide to the classification of the Late Tertiary and Quaternary soils of the Lower Amazon valley' (DAY, 1961) is followed as closely as possible, the said guide also beingbase d on the USA system. However, DAY'Sclassi fying had to be done practically only on morphometrical field studies. Laboratory data afterwards becoming available, expansion of thefield studies , and more definite correlation withth efindings i nothe r regionso f Brazil,mad e itnecessar y to elaborate, and inpar t deviatefrom , thisGuide . The levels of classification are those of the 'Great Soil Group', its subgroups, and phases of these. For practical soil mapping, the soil texture is often included in the nomenclature. In the following, the soil units found to date in Amazonia are described systematic ally.A ful l description ofth econcep t and the range incharacteristic s of a unit isgive n only for those ones of which frequent occurrence was established, for instance those mapped in survey areas. Further details are giveni n direct survey reports (DAY, 1959; SOMBROEK, 1962a; SOMBROEK, 1962b; SOMBROEK and SAMPAIO, 1962).Th e listing of thesoil sdiscusse d isgive ni nTabl e9 .Fo r analytical data on most individual profiles, cf. Appendix 10.

(I) KAOLINITIC LATOSOLS AND KAOLINITIC LATOSOLIC SANDS (KL and KLS) A general discussion of the characteristics of this most important group of Amazon soils isgive n in 11.2.2.Thei r qualities are dealt with in V.3. da) Kaolinitic Yellow Latosol(KYL ) Generalconcept: This unitcomprise sdeepl yan d strongly weathered soils,well-drain edan d permeable,predominantl y withcolou ro fyellowis h hue,an d ofvaryin gtexture , but with at least 15%cla y in the Bhorizon . The soils are very strongly or extremely acid,havin ga bas esaturatio n below40% . The profiles are well developed, having ABC sequence of horizons; the boundaries between these horizons aregradua l or diffuse. The soilshav ea latosolic-Bhorizo n (cf. H.2.1); the clay-sized particles consist very predominantly of silicatecla ymineral so f

125 Table 9 Listing of the describedAmazon soils

Profile no. and Soil name Symbol detailed symbol nomede solo simbolo numero doperfil e simbolo detalhado

(I) KAOLINITIC LATOSOLS AND KAOLINITIC LATOSOLIC SANDS LatosolosCaoliniticos e AreiasLatosolicas Caoliniticas KL + KLS (la) Kaolinitic Yellow Latosol KYL LatosoloAmarelo Caolinitico (Ia.l) Kaolinitic Yellow Latosol (firtho) KYL 24 KYL„» LatosoloAmarelo Caolinitico (,Orto) 25 KYLm (Ia.2) KaoliniticYellow Latosol, Compact phase KYL.C 26 KYL, C„A Latosolo AmareloCaolinitico, fase Compacta i 27 KYL,c rA (Ia.3) Kaolinitic Yellow Latosol, Concretionary KYL.CR 28 KYL, CR phase Latosolo Amarelo Caolinitico, fase Con- creciondria (Ia.4) Kaolinitic Yellow Latosol, intergrade to KYL-RP 29 KYL-RPr» Yellow Podzolic soil Latosolo Amarelo Caolinitico, ''intergrade' para solo Podzdlico Vermelho-Amarelo (Ia.5) Kaolinitic Yellow Latosol, intergrade to KYL-DHL 30 KYL-DHL* Dark HorizonLatosol Latosolo Amarelo Caolinitico, 'intergrade'' para Latosolode Horizonte Escuro (Ia.6) Kaolinitic Yellow Latosol, intergrade to KYL-GL Ground Water Laterite soil LatosoloAmarelo Caolinitico, 'intergrade' para soloLaterita Hidromdrfica

(lb) Kaolinitic Red Latosol KRL 31 KRLm Latosolo Vermelho Caolinitico (Ic) Kaolinitic Latosolic Sand KLS 32 KLS, F Areia Latosolica Caolinitico 33 KLS, s (II) RED YELLOW LATOSOLS RL 34 RL? Latosolos Vermelho-Amarelo (III) DARK RED LATOSOLS DL 35 DL,s Latosolos Vermelho Escuro (IV) RED YELLOW PODZOLIC SOILS RP Solos Podzolicos Vermelho-Amarelo (IV.l) Red Yellow Podzolicsoil (firtho) RP 36 RPI6 Solo Podzolico Vermelho-Amarelo (,Orto) 37 RP,.** (IV.2) RedYellow Podzolic soil, intergrade to Kaolinitic RP-KYL 38 RP-KYL,* Yellow Latosol 39 RP-KYLrA Solo Podzdlico Vermelho-Amarelo, 'intergrade' para Latosolo AmareloCaolinitico (IV3) Red Yellow Podzolic soil, intergrade to Kaolinitic RP-KYL, CR 40 RP-KYL, CR Yellow Latosol,Concretionary phase Solo PodzdlicoVermelho-Amarelo, 'intergrade' para Latosolo Amarelo Caolinitico, fase Con- creciondria (V) REDYELLO W MEDITERRANEAN SOILS RM 41 RM Solos Mediterranicos Vermelho-Amarelo (VI) LITHOSOLS ^ L 42 L Litosolos

126 Table9 continued/ Tabela 9 conlinuado

Profile no.an d Soil name Symbol detailed symbol nome de solo simbolo mimero doperfil e simbolo delalhado (VII) GROUND WATER LATERITESOIL S GL 43 GL SolosLaterita Hidromorfica (VIII) HYDROMORPHIC GREY PODZOLICSOIL S HP 44 HP*»,O Solos'Hydromorphic Grey Podzolic' 45 HP*!,, D 46 HP*»,s (IX) GROUND WATER PODZOLS AND WHITESAN D REGOSOLS GP 47 GP PodzolsHidromdrficos e Regosolos de Areia Branca WSR 48 WSR (X) Low HUMIC GLEY AND HUMIC GLEY SOILS LHG 49 LHG SolosClei Humico eClei Pouco Humico HG 50 HG 51 HG.u (XI) SALINE AND ALKALI SOILS 52 Sol,c Solos Salinose Alcalinos (XII) TERRA PRETA' SOIL TP 53 TPj Terra Preta (XIII) 'TERRA ROXA LEGITIMA' AND 'TERRA ROXA ESTRUTURADA'; GRUMOSOL; NON CALCICBROWN LIKE SOIL, GRAVELLY PHASE(NB , G); ACID BROWN FOREST-LIKE SOIL. GRAVELLY PHASE(AF , G); ALLUVIAL SOIL; 'PARA' PODZOL; GREY HYDROMORPHICSOIL ; HALF BOG SOILAN D BOGSOI L TerraRoxa Legitimae Terra Roxa Estruturada; Crumosolo; Soloparecido com 'Non Calcic Brown', fase Cascalhenta; Solo parecido com 'Acid BrownForest', fase Cascalhenta; Solo Aluvial; Podzol de Para; Solo 'Grey Hydromorphic'; Solos Orgdnicos

Applied soil textural classes Applied basesaturatio n classes aplicadas classes texturais desolo classesde saturacao de bases clay(%) class symbol V(%) class symbol argila(% ) classe simbolo classe simbolo < 5% very light textured vl <50 low lb texluramuito leve baixa 15-15% light textured I 50-70 rather high rhb texturaleve meio-alta medium textured 15-35% m 70-100 high hb textura media alia 35-50% rather heavy textured rh texturameio pesada 50-70? heavy textured texturapesada >70% very heavy textured vh texturamuito pesada If the textural class is applied for the soil asa whole , the texture of the B, horizon istake n as the criterion Sea classe texturali aplicada aosolo como um todo,entao a texturado horizonte Bti a queprevalece In generalisations, the first three classes are grouped together as 'relatively light textured', and the latter three as 'relatively heavy textured'.Th e terms 'sandy'an d 'clayey' for these two groupsar e normally avoided Ao sefazerem generalizacoes, astris primeiras classes saoagrupadas como 'textura relativamente leve' e as ultimas tres como'textura relativamente pesada'. Os tirmos 'arenoso' e'argiloso', para estesdois grupos,sao geralmente evitados Tabela 9 Os solosamazonicos descritos

127 the 1:1 latticestructure ,bein gkaolinite .Th esoil shav ea rathe rwea k macro-structure, atleas ti nth esubsurfac e horizon.Whe ndr ythe yar eslightl yhard ,o rhard .Th eB hori zoni sgenerall ysomewha theavie rtha n theA horizon ,a fac twhic hi sals onoticeabl ei n thefield, excep twher ever yheav ytexture d profiles aremet . Within the Kaolinitic Yellow Latosol group, various subgroups are distinguished. For practical mapping purposes, these subgroups are, for the most part, further sub divided accordingt oth etextur eo fth e B horizon(cf. Table9) .

(Ia.l)Kaolinitic YellowLatosol (,Ortho) (KYL) The profiles of the Ortho type are characterized by the following features: 1. Thetransitions ,bot h asregard scolou r and texture,betwee nth ehorizon sar e diffu se,o r gradual, exceptth e transition from the At to the A3.Fro m top to bottom of the profile, the colour gradually obtains more value, more chroma, and a redder hue; mottlesar eabsent . 2. TheB horizo n isneithe rcompac t nor firm. 3. Thesoil sar edeepl yporou san d deeply rooted,t oth eC horizon . 4. Hard plinthitic gravel isabsen t or nearly absent. If occurring scattered in the pro file, it comprises maximally 5% of the soil mass; if concentrated in a layer, then this layeri sles stha n 30c mthic k above 1 mdepth . The range in characteristics of the very heavy texturedprofile s isgive n below:

The profiles have heavy clay texture in all horizons. The Ahorizon , consisting ofa n A, and an As,i s between 20 and 45 cm thick; the Bhorizon , consisting of a B2 and often a B3, is 80 to 150c m thick; the Chorizo n iso fvaryin gan d often difficult todetermine ,thickness . The Aisubhorizo nvarie s considerably in thickness (from 2t o 30cm) . The colour is predominantly dark yellowish brown (10Y R 4/4) or dark brown to brown (10Y R4/ 3o r 5/3).Th estructur e isusuall y moderate, medium tofine subangula r blocky,wit h often weak to moderate,ver yfine granule s present. The superficial part may becruste d andclodde d (cf.IV .1.1.2) . The A3 subhorizon is 15-40c m thick. The colour is predominantly yellowish brown (10 YR 5/6 or 5/4) or light yellowish brown (10 YR 6/4). It has usually a weak to moderate fine subangular blocky and moderate to weak, very fine granular structure. On a considerable part of the insect and root channels,coating s may bepresent . The subhorizons of the Bhorizo n are normally strong brown (7.5Y R 5/6), lessfrequentl y yellowish brown (10 YR 5/6) or reddish yellow (7.5 YR 6/6 or 6/8). The structure iswea k to moderate, mostly medium subangular blocky, with weak, very fine granular elements present in varying frequency. A portion ofth echannel sma yhav ecoatings . The C-horizon is reddish yellow (7.5 YR 6/6 or 6/8, 5 YR 6/8), or yellowish red (5 YR 5/8), and massiveo r with aweak ,mediu m subangular blocky structure. The characteristics of the heavy, rather heavy, and medium texturedprofile s have the following range: Theprofile s haveheav ysand yloam ,sand ycla yloam ,sand yclay ,o rcla ytextur ei nth eB horizon .Th e A horizon, consisting of an At and A,, is3 5t o 75c m thick.Th e Bhorizon , consisting ofa B, and B„ and often a Bt,i s 100t o25 0c m thick.Th e C horizon iso fvaryin gan d noteasil ydetermine d thickness. The Aj subhorizon ranges from 5t o 30c m in thickness. The texture varies from sand to sandy clay. Thecolou r hasvalue so f3 t o 5,a chrom a of3 o r4 ,an da hu eo f 10YR ,les sfrequentl y 7.5Y Ro r 5YR ; most common are brown (10 YR 4/3 or 5/3) and dark yellowish brown (10 YR 4/4). In the lightest textures the horizon may bestructureles s (singlegrains), bu t moreofte n there isa structure ofwea k to moderate,fine subangula r blocky and veryfine granules . Both under a cover of primeval forest, and

128 under cultivation, theuppe r section ofth e A, often consistso fa layer ofloose , bleached sand (pepper and salt colour), ranging inthicknes s from 0.1 to5 c m ('micro-podzol'). The A3 subhorizon isnormall y between 20an d5 0c m thick. The texture varies from sandy loamt o light clay. The colour value isusuall y 5o r6 ,th e chroma varies from 4t o8 ,an d the hue isgenerall y 10YR ; most common are yellowish brown (10 YR5/6 ) andligh t yellowish brown (10 YR6/4) . The structure is mostly weak to moderate, medium to fine subangular blocky, with very fine granules present invaryin g frequency. The B, subhorizon, when present, is5 0t o8 0c m thick. The texture isusuall y heavy sandy loamo r light sandycla yloam . Thecolou r isbrownis h yellow(1 0Y R 6/6)o ryello w(1 0Y R 7/6),les s frequent lyreddis h yellow (7.5Y R6/ 6o r7/6) .Th estructur e isnormall y thesam e astha t ofth e A,. TheB 2subhorizon , 70t o15 0c m thick, variesi n texturefro m heavysand yloa m toclay ; thetextur ei s always noticeably heavier than that ofth e Ahorizon . The colour isnormall y reddish yellow (7.5 YR 6/6o r6/8 )o rbrownis h yellow (10Y R 6/6o r6/8) .Th estructur e isnormall ya slightl ycoheren t porous soil mass, falling apart into weak, medium to fine subangular blocky elements and some very fine granules. The B, horizon is commonly slightly lighter textured than the Ba, with the other characteristics approximately identical. The Chorizon , in texture varying from sandy loam to sandy clay, isusuall y reddish yellow (5Y R 6/8) oryellowis h red (5Y R5/8) . The horizon ismassive ,bu tha s mostly a large number offine pores .

Profile 24. KAOLINITIC YELLOW LATOSOL (,ORTHO), very heavy textured (KYLt*). Field description 210(Sombroek) Location:Alon g highway BR-14,k m 247,S o fSa o Miguel do Guama. Reliefand drainage: Totall y flat topo frathe r small section of high terrace(planalto) .Alt .20 0m .Well - drained. Parentmaterial: Plio-Pleistocen e lacustrine sediments (Belterra clay). Vegetative cover: Primeval tropical forest. Rather low timber volume (100m*/h a ca.)\ rather few,but thick trees; dense undergrowth, ofthic k creepers and climbers.

Ot 8-5cm: Undecompose d plant residues. O, 5-0 cm: Partly decomposed plant residueswit h manyfine roots . A, 0-2cm: Dark yellowish brown (10Y R 4/4)heav yclay . Moderate, medium tofin e subangular blocky and weak, very fine granular structure. Moist, friable. Sticky and plastic whenwet . Locally thehorizo n iscrusty ,du et ointens eactivit y ofinsects ,especiall y termites. Manypores . Verymany ,mostl yfine , roots.Transitio n clear. A» 2-20cm: Yellowis h brown (10Y R 5/6)heav y clay. Moderate, fine subangular blockyan d weak, veryfine granula r structure. Moist, friable. Stickyan d plastic when wet. Soft when dry. Many pores.Ver yman y roots.Transitio n gradual.

B2 20-60cm: Stron g brown (7.5Y R 5/6) heavyclay .Wea k tomoderate ,mediu m tofine subangu lar blocky and weak, very fine granular structure. A few, faint clay skins. Moist, friable. Sticky and plastic when wet. Slightly hard when dry.Man y pores. Many roots. Transition diffuse. B, 60-150 cm: Strong brown (7.5 YR5/6 )heav y clay. Weak, medium sized subangular blocky and weak, very fine granular structure. A few,ver y faint clay skins. Moist, friable to firm. Sticky and plastic when wet. Slightly hard when dry. Common pores. Many roots. Transition diffuse. C 150-250cm: Yellowish red (5Y R5/8 ) heavyclay . Massive toweak , medium sized subangular blocky structure. Moist,firm. Stick yan d plasticwhe n wet. Few pores.Ver yfe w roots.

For full mineralogicalanalysi san d specific surface data ofth e A3an d B2horizon so f acomparabl e profile: seesampl e303- 2an d 303-4o fth eTable s7 an d 8,Fig . 14a-ean d Photo9 . ForpF-curve so fth eB 2o fcomparabl e profiles:se e samples219- 3an d 303-3 of Fig. 37.

129 Profile 25. KAOLINITIC YELLOW LATOSOL (,ORTHO), medium textured (KYLm). Field description 233(Sombroek-Sampaio ) Location:Alon ghighwa y BR-14,k m12. 7S o fSa o Migueld oGuama . Reliefand drainage: Flat top ofterrace ,abou t 55c m above nearby river.Well-drained . Alt.6 5 m. Parent material: Pleistocenefluviatile (? )sediments . Vegetative cover: Tropical forest with medium timber volume (150 m'/ha ca.).A tth e spot probably onceo rtwic eattacke d by fire. Ot 6-1 cm: Undecomposed plant residues. 0» 1-0 cm: Partlydecompose d plant residues,wit h veryfine roots . Au 0-2cm: Loose bleached sand. Aia 2-20 cm: Brown (10Y R 4/3)coars e sand. Singlegrain s and afew , weak, fine crumbs. Moist, loose. Notstick y andno tplasti c when wet. Soft when dry. Very many, mainly fine roots. Transition gradual. A, 20-70 cm: Yellowish brown (10 YR5/4 ) light coarse sandy loam. Little coherent massive, to weak, medium to fine subangular blocky structure. Moist, very friable. Notstick y and not plasticwhe nwet . Slightly hard when dry. Slightly more resistant to penetration than other horizons.Commo n pores. Many roots. Somepiece so fcharcoal . Transition diffuse. B, 70-140cm: Brownish yellow (10Y R 6/6) coarse sandy loam. Weak, medium tofine subangu lar blocky structure. Moist,ver yfriable . Not stickyan d not plasticwhe n wet. Slightly hard,t o hard when dry. Common pores. Many roots. Transition diffuse. B2I 140-220cm: Reddish yellow (7.5Y R6/6 ) heavy coarse sandy loam. Little coherent massive, to weak, medium to fine subangular blocky structure. Moist, very friable. Notstick yan d slightly plastic whenwet . Slightly hard, tohar d when dry.Common , tofe w poers. Few roots. Transition diffuse. B22 220-320cm: Reddish yellow (7.5Y R6/6 ) light coarse sandy clay loam. Little coherent massi ve. Moist, very friable. Notstick y andslightl y plastic when wet. Hard when dry. Few pores. Veryfe w roots.

For full mineralogical analysis and specific surface data of the A3an d B2horizon so f a comparableprofile : seesampl e300- 2an d300- 4o fTabl e7 an d 8. Forth epF-curv eo f the B2o fa comparabl e profile: seesampl e23 1o f Fig. 37.

(Ia.2) Kaolinitic YellowLatosol, Compactphase (KYL,c ) The profiles ofth e Compact phase differ from those ofth e Ortho typei nbein gcom pact andrathe r firm, especially inth e Bhorizon . This Bhorizo n isrelativel y slightly porous;th e root development inthi s horizon islimited . The hue ofth eprofile s isyel lower(ofte n 2.5Y) . Inth ever yheav ytexture d profiles,th eA xhorizo n isextremel ythin . Inth elighte rtexture d profiles thetextura ldifferenc e betweenth eA an dth e Bhorizon s isfairl y great.Som emottlin gma yals ob epresen t inth e Bhorizo n ofth elatte r profiles. The very heavy textured andth erathe r heavy textured profiles were frequently en countered. Arang e incharacteristic sca ntherefor e begiven .

Therang efo rth e veryheavy textured profile s isa sfollows :

The Ahorizon , consisting ofa n A,an da nAj , is30-5 0c m thick. The Bhorizon , consisting ofa Bt and aB» , isbetwee n 100 and 150c m thick.Th e Chorizo n iso fvaryin gthickness . The Ai subhorizon rarely comprises more than 2-3cm . Thecolou r value is5 o r6 ,th echrom a 3o r4 , the hue1 0Y Ro r 2.5 Y. Light yellowish brown (10 YR6/4 )i sth emos t common. Thestructur e is usually moderate, fine subangular blocky, composing weak, medium subangular blocky. Due to the thinnesso fth e Aj, and thethinnes s© rabsenc eo fa n C^an d Ot, theyellowis hcla yo f theA , can often be observed barea t thesurface . The A, subhorizon, varyingfro m 30t o5 0c m in thickness,ha s colour

130 values of 6 or 7, c'hromas of 4 or 6, hues of 2.5 Y or 10YR . Most common are very pale brown (10 YR 7/4)an d paleyello w(2. 5 Y7/4) .Th estructur e is a moderate,fine subangula r blockyan d weak very fine granular. The Bs horizon, varying in thickness from 50 to 100cm , has colour values of 6 or 7, chromas of 6o r 8,hue s of 10Y R or 7.5YR . Most common isyello w (10Y R 6/7). Mottling does not occur. The structure is commonly weak to moderate, fine subangular blocky and locally weak, very finegranular . The B3subhorizo n isabou t 70c m thick.Th e colour isgenerall y reddish yellow (7.5 YR 6/8). The structure is usually weak, medium subangular blocky but sometimes a weak, medium sized platy structure occurs.Th e C-horizon isgenerall y reddish yellow (5 YR 7/6).Th e horizon is normally massive.

For the ratherheavy texturedprofile s the range incharacteristic s isa s follows: The Ahorizon , consisting ofa n A, and an A„ varies between 20an d 60 cm in thickness.Th e Bhori zon, consisting of a B2an d a B3, is between 100an d 120c m thick. The C horizon iso f varying thick ness. The A horizon is often capped with a rather thick (up to 3 cm) layer of loose bleached sand ('micro-podzol'). TheA ! subhorizon variesbetwee n 5an d 20 cmi nthickness .Th etextur e is predominantly sandy loam. Thecolou r isgre y brown (10Y R 5/2), brown (10Y R 5/3) or pale brown (10Y R 6/3). The horizon is structureless, or has a weak, fine subangular blocky structure. The A„subhorizo n is between 20 and 40c m thick.Th e texture ismostl ysand y clay loam,sometime ssand y loamo r sandyclay .Th ecolou r is predominantly verypal e brown orpal eyello w(1 0YR , resp.2. 5Y 7/4) ,sometime spal e brown (10Y R 6/3). The structure isusuall y weak, medium subangular blocky. The Bs subhorizon isbetwee n 20 and 70c m thick. The texture ispredominantl y sandyclay , sometimessand y clay loam, orclay .Th e colour is commonly brownish yellow(1 0Y R 6/6), but alsocolour s of lesschrom a occur (10Y R resp.2. 5 YR 6/3-6/4). The latter is mostly the case if the horizon iso fcla y texture;the n also common and distinct mottleso fyellowis h red (5Y R 5/8),ma yb epresent .Th estructur e isnormall y weak,mediu m subangu- ar blocky but sometimesa stronger structure (weak to moderate, medium angular blocky) is present, accompanied bya few faint clayskins .Th e B3subhorizo n isbetwee n 50an d 100 cmthick . The texture is predominantly a light sandy clay. The colour is commonly reddish yellow (7.5 YR 6/6-6/8), but yellow (10 YR resp. 2.5 Y 7/6) may occur. The horizon is massive, or has a weak, medium sized subangular blocky structure. The C horizon has normally sandy clay or sandy clay loam textures. Its colour is predominantly reddish yellow (7.5 resp. 5 YR 6/8-7/8).

Profile 26. KAOLINITIC YELLOW LATOSOL, Compact phase,ver y heavy textured (K.YL, cv/,) Field description 230(Sombroek ) Location:Highwa y BR-14, km 69 of Sao Miguel do Guama, 6 km E. Reliefand drainage: Totally flat terrace. Well-drained. Alt. 70m . Parentmaterial: Plio-Pleistocen e lacustrine sediments, re-worked in early Pleistocene. Vegetativecover: Primeval tropical forest, withmediu m(14 0m s/haca.) timbervolume .Scattere d thick: trees,ofte n in small groups; dense undergrowth, largelyconsistin g of thick creepers and climbers. O, 4-1cm: Undecompose d plant residues. O, 1-0 cm: Partly decomposed plant residues with many very fine roots. Surface rough and rather hard, locally bare and then with green cover (Algae). A, 0-3 cm: Grey brown (2.5 Y-10 YR 5/2) heavy clay. Structure of moderate, fine subangular blocks,composin g weak, medium sized subangular blocks. Moist, friable to firm. Sticky and plastic when wet. Hard when dry. Many pores,man y large insect channels.Ver y many, main lyfine roots .Transitio n clear. A, 3-40 cm. Light yellowish brown (2.5Y 6/4 ) heavyclay . Moderate,fine subangula r blocky and weak, very fine granular structure. Moist, friable. Sticky and plastic when wet. Slightly hard when dry. Rather compact. Many pores. Many roots.Transitio n gradual. B, 40-90 cm: Brownish yellow (10 YR 6/6) heavy clay. Structure of weak to moderate, very fine- subangular blocks,composin g weak, medium sized subangular blocks. Tendency to platy. A few, faint clay skins. Moist, friable to firm. Sticky and plastic when wet. Slightly hard, to- hard when dry. Compact. Very resistant to penetration with hammer. Few pores. Few roots- Transition diffuse.

131 WithDutch auger: B» 90-130cm: Brownish yellow (10Y R -7.5 YR6/8 ) heavy clay. Moist, firm. Sticky and plastic whenwet . Verydifficul t topenetrat e withsoi lauger . Transition diffuse. C 130-180cm: Reddish yellow (7.5 YR7/8 )heav y clay. Moist, firm. Sticky andplasti c when wet.Ver ydifficul t topenetrat e with soilauger .

Profile 27. KAOLINITIC YELLOW LATOSOL, Compact phase, rather heavy textured (KYL, CM) Field description 194 (Sombroek) Location: Highway BR-14,k m 63S o fSa o Miguel doGuama , 7k mE . Reliefand drainage: Extensive,completel y flat terrace,abou t 60m abov e river level.Somewha t imper fect internal drainage.Alt .6 0m . Parent material: Early Pleistocene fluviatile (?)sediments . Vegetative cover:Primeva l tropical forest of low timber volume (75 m'/ha fa.). Dense undergrowth, predominantly of thincreeper san d climbers. C^ 9-2 cm: Undecomposed plant residues. O, 2-0 cm: Partly decomposed plant residueswit h manyfine roots . An 0-1 cm: Loosebleache d sand. An 1-30cm: Pale brown (10Y R 6/3) sandy loam. Weak,fin e subangular blocky structure. Moist, friable. Not sticky and not plastic when wet. Slightly hard when dry. Slightlycompact . Com mon tofe w pores. Common roots. Transition diffuse. A, 30-60: Light yellowish brown (2.5 Y 6/4) heavy sandy clay loam. Weak, medium subangular blocky structure. Moist, firm. Slightly sticky andslightl y plastic when wet. Hard whendry . Compact. Very resistant topenetratio n with hammer. Very few pores. Few roots. Transition diffuse. Ba 60-120cm: Yellow (2.5Y 7/6 ) light sandy clay. Weak tomoderate ,coars e subangular blocky structure Moist,firm. Slightl y stickyan d slightlyplasti c when wet. Hard when dry. Compact. Veryresistan t topenetratio n with hammer. Common pores.Fe wroots .Transitio n diffuse. B, 120-170cm: Yellow (2.5 Y7/6 )sand y clay loam, with few,mediu m sized, faint mottleso f yellow (10Y R 7/6). Massive to weak, medium sized subangular blocky structure. Moist, friable tofirm. Slightl ystick y and slightly plastic when wet. Hard when dry. Slightly compact. Lessresistan t topenetratio n with hammer than B,. Common pores.N oroots .

(Ia.3) Kaolinitic YellowLatosol, Concretionaryphase (KYL , CR) The profiles ofth eConcretionar y phasediffe r from thoseo fth eOrth o typeowin gt o the presence ofplinthiti c gravel. This includes more than 5% o fth e soil masswhe ni t occurs scattered inth eprofile , orconstitute s a layer inexces s of3 0c mi nthicknes s above a depth of 1 m. Soft plinthite isabsen t above a depht of 1 m.Th e colour is often, but not always,somewha t reddertha ntha t ofth eOrth otype . Fora genera ldis cussion on the geneticso fthi s soil,pleas e refer to II.3.2.2.

Profile 28. KAOLINITIC YELLOW LATOSOL,Concretionar yphas e(KYL , CR) Field description 114(Day ) Location:Nea r Belem,3 k m So fIA N Headquarters (Lat. 1°.27'S ;Lon g48°.26 'W) . Relief anddrainage: Undulating terrain (slope 2-5%). Alt. lesstha n 50m . Well-drained. Parentmaterial: Mio-Pliocenefluviatile (? )sediments . Vegetativecover: Cleared from brush. Formerly forest.

Ap 0-25cm: Dark reddish brown (5Y R 3/2) gravelly sandycla y loam. Weak,coars e subangular blocky structure, breaking into moderate, medium tofine subangula r blocky. Moist, friable. Gravels are coarse grained iron-quartz plinthite fragments, ofdiamete r 0.5-3 cm, occupying about 75 % ofth esoi l mass.p H (Soiltex)4 o rbelow .Transitio n clearan d smooth. A, 25-50 cm: Gravelly clay. Transitional horizon between AIP and B„. Transition gradualan d wavy.

132 B2, 50-100cm: Yellowish red (5Y R 5/6)gravell yclay .Weak ,coars e subangular blocky structure. Moist, friable. Gravels areplinthit e fragments, containing some medium or coarse grained quartz.The y areo fdiamete r 1-5 cman d occupyabou t 75 % ofth e soil mass.p H(Soiltex )4 o r below. Transition diffuse and wavy. B2, 100-170cm: Red(2. 5Y R 5/8) gravelly clay. Weak, very coarse angular blocky structure, breaking into moderate, fine angular blocky. Moist, friable. Gravels areplinthit e fragments, fine, and occupyingabou t 75 %o f thesoi lmass .p H (Soiltex)4 o r below.Transitio n diffuse and wavy. WithDutch auger: Q 170-230cm: Red (2.5Y R 5/8) clay,wit h common tomany , medium sized,prominen t mottles of yellow (10 YR8/8 ) anddar k red(1 0 R 3/6). Stilla few plinthite concretions, fine grained, rather soft. Ct 230-290+cm: Light reddish brown (2.5Y R 6/4)clay ,wit h many,coarse ,prominen t mottleso f pinkish white(7. 5Y R 8/2)an d duskyre d (10R 3/4) .

(Ia.4) Kaoliniiic YellowLatosol, intergrade toRed Yellow Podzolic soil (KYL-RP ) The profiles ofthi sgrou p differ from those of the Ortho type in having a distinct colour change from the A to the B horizon. Also, the textural difference be tweenth e Aan d the Bhorizon s isslightl ylarge rtha n that ofprofile s ofth eOrth otyp e with the samegenera ltexture .Th echang e intextur e is,however ,diffus e togradual , as inth eOrth otype . Many exampleswer estudie d of the ratherheavy texturedprofiles . Therefore, arang e incharacteristic sfo rthi stextura lclas sca nb egiven :

The Ahorizon ,consistin g ofa n A!an d an A„ isbetwee n 35an d 60 cm thick.Th eB-horizon , consist ing ofa B2an d a B„ is 150t o 180c m thick. The Chorizo n iso fvaryin g thickness.Th eA horizo n is generally capped with onlya thin layer( < 1mm )o floos e sand ('micro-podzol'). The A, subhorizon is between 5an d 10c m thick. Thetextur e ispredominantl y sandy clay loam, sometimes sandy loam, orsand y clay.Th e colour isgrayis h brown orbrow n (10Y R 5/2 resp. 10Y R 5/3).Th estructur e iswea ko rmoderate ,fine subangula r blocky.Th etransitio n toth eA , ischaracteris tically clear. TheA ,subhorizo n varies between 20an d 50c mi n thickness. Thetextur e is normally sandycla yloa m orsand yclay .Th ecolou rvalu ei s6 o r 7, thechrom a 4o r 6, thehu e 10YR , sometimes 2.5 Y.Mos t frequent isver y pale brown (10Y R7/4) . The structure iswea k tomoderate , medium to fine subangular blocky.Th e transition toth eB horizo n ischaracteristicall y gradual,no t diffuse. Some timesa transitio n zone( A + B)occurs ,wit hcolour s both ofA , and B,. The B,subhorizo n is6 0t o 80 cm thick. The texture iscommonl ya ligh t clay,bu t sandyclay sar e also rather frequent. The colouri s reddish yellow(7. 5 YR 7/6,sometime s7. 5Y R 7/8o r 6/6). Thestructur e isa weak ,mediu m subangular blocky.A few, faint clayskin s may occur. The transition toth e B,i sdiffuse . The B,subhorizo n varies between 50an d 100c m inthickness . The texture isa sandy clay or,les s frequently, a light clay. The colour isreddis h yellow (5 YR6/8 , 5Y R7/8) .Th ehorizo n has a weak, medium subangular blocky structure, if any. TheC-horizo n has commonly a sandy clay texture. Sandy clay loam or light clay texturesals o occur. Thecolou r isreddis h yellow( 5Y R 6/8)o rligh t red (2.5 YR 6/8), often with some mottling ofyello w (10 YR8/6) . Sometimes colours ofrathe r yellow hue andlittl e chroma (e.g. pink, 7.5Y R 7/4) predominate.

Profile 29. KAOLINITIC YELLOW LATOSOL, intergrade to RED YELLOW PODZOLIC soil, rather heavy textured (KYL-RP,./,); theactua l profile would come under the heavy texture class Field description 197(Sombroek) Location:Alon g highway BR-14,k m13 6S o fSa o Miguel doGuama . Relief anddrainage: Edge offlat terrace , 10m above level ofnearb y rivulet Well-drained. Alt. 80 m. Parentmaterial: Pleistocenefluviatile sediments . Vegetative cover: Primeval tropical forest ofrathe r high timber volume (200-250 m'/ha). Erect, open undergrowth, nearly devoid ofcreeper san d climbers.

133 Oi 4-0.5cm: Undecomposedplan t residues. Oa 0.5-0cm: Partly decomposed plant residues with veryfine roots . Surface somewhat irregular and hard, duet ointens e termite activity. Loose bleached sand grains form a very thin discon tinuouscover . An 0-5 cm: Dark brown (7.5 YR3/2 )heav y sandy clay loam. Moderate, very fine subangular blocky structure. Many very fine granular insect outcasts. Moist, friable. Sticky andslightl y plastic when wet. Slightly hard when dry. Very many pores andlarg e insect channels. Very many, mainlyfine, roots .Transitio n abrupt. A„ 5-20cm: Dar k grey brown (10 YR4/2 )sand y clay, with many, fine, faint mottles of pale brown (10 YR 6/3). Weak tomoderate ,fine subangula r blocky andweak , very fine granular structure. Moist,friable . Stickyan d plasticwhe n wet.Sof t when dry. Many pores. Many roots. Transition clear. A, 15-50cm: Very pale brown (10 YR 7/4) clay. Weak tomoderate , medium subangular blocky and weak, very fine granular structure. Moist, friable. Sticky andplasti c when wet. Slightly hard when dry. Many pores. Many roots.Transitio n gradual. A + B50-70 cm: Reddish yellow (5 YR 7/8) clay, with many,fine, fain t mottles ofver y pale brown (10Y R 7/4). Atransitio n zone between A,an d Ba. B2 70-140cm: Reddish yellow (5 YR 7/8) clay. Structure weak, coarse subangular blocky, fall ing apart into weak, fine subangular blocky. A few,ver y faint clay skins. Moist, friable. Sticky and plastic when wet. Hard when dry. Many pores. Common roots. Transition gra dual. B, 140-190cm: Reddishyello w( 5 YR 7/8)clay ,wit hman yfine, fain t mottleso fpin k (7.5 YR 7/4). Massive, to weak, coarse subangular blocky or medium prismatic structure. Moist, friable. Stickyan d plasticwhe n wet.Slightl y hard when dry.Commo n pores.Commo n roots.Transi tion gradual. C 190-250 cm: Pink (7.5 YR 8/4) heavyclay , with many, medium sized, faint todistinc t mottles of white (10 YR8/1) . Massive, to weak, medium prismatic structure. Moist, friable. Sticky and plastic when wet. Slightly hard when dry. Few, tocommo n pores. Few roots.

(Ia.5) Kaolinitic Yellow Latosol, intergrade toDark Horizon Latosol(K.YL-DHL ) Theprofile s ofthi sgrou pdiffe r from thoseo fth eOrth otyp eb yhavin gi nth e B hori zon alaye r ofslightl y darker colour (lower value) than the earth immediately above and below.Th eclassificatio n isver yprovisional ,awaitin gmor efield studies ,an d ana lyticaldat a (for Dark Horizon Latosolcf. RUHEan d CADY, 1954; SOIL SURVEY STAFF, 1960, p. 241).

Profile 30. KAOLINITIC YELLOW LATOSOL, intergrade to DARK HORIZON LATOSOL, heavy textured (K.YL-DHLA) Field description 319(Sombroek , Falesi) Location: 8k m Eo fPort o Velho; Experimental Station ofIAN , 2k m So fHeadquarters . Relief and drainage:Approximatel y flat terrain, 30m ca.abov e larger rivulets. Well-drained. Alt. 100mra . Parentmaterial: Pleistocene sediments. Vegetativecover: High secondary forest. O, 7-4 cm: Undecomposed plant residues. O, 4-0 cm: Partly decomposed plant residues,an d very many veryfine roots . A, 0-5cm: Dark yellowish brown (10 YR 4/4)fine sand y clay. Moderate,fine subangula r blocky structure. Moist,friable . Slightly sticky and slightly plastic,t o plasticjwhen wet. Many coarse and few fine pores. Very many roots. Afe w pieces ofcharcoal . Transition clear. A, 5^45cm: Yellowish brown (10 YR5/4 )fine sand y clay. Weak, fine subangular blocky and weak to moderate, very fine granular structure. Moist, friable. Slightly sticky and slightly plastic,t oplastic , when wet'. Few coarsean dcommo nfine pores .Man yroots .A few pieces of charcoal. Transition gradual.

134 B,» 45-85 cm: Dark yellowish brown (10Y R 3/4)fine sand yclay .Weak ,fine t omediu m subangular and weak, very fine granular structure. Moist, very friable. Slightly sticky and slightly plastic,t o plastic,whe nwet .Littl eresistanc e to penetration with hammer. Common to many fine pores. Many roots.Transitio n gradual. Ba 85-130 cm: Strong brown (7.5 YR 5/6) clay. Porous massive, to weak, medium subangular blocky and weak, very fine granular structure. Moist, very friable. Plastic and slightly sticky when wet. Little resistance to penetration with hammer. Many medium sized pores. Many roots.Transitio n diffuse. B, 130-180+ cm: Reddish yellow(7. 5Y R 6/8)clay . Porous massive, to weak, medium subangu lar blocky structure. Moist,ver yfriable . Plastican d slightly sticky when wet. Little resistance topenetratio n with hammer. Commonfine pores .Commo n roots.

(Ia.6) Kaolinitic Yellow Latosol, intergrade to Ground WaterLaterite soil (KYL-GL) The profiles of this group differ from those of the Ortho type in having in thelowe r parto f theprofil e somesof t plinthite-in-formation: The Ban d C horizonshav e a fair quantity ofreddis hmottle si na matri xo fles schrom atha nth eOrth otype .Th eB hori zon has certain characteristics of a textural-B (cf. II.2), viz.relativel y large textural difference B/A, subangular to angular blocky structure,a few clay skins,an d limited porosity.Th eA 3subhorizo ni s often slightlypale r(e.g. hue2. 5Y ,chrom a4 t o6 )tha n that ofth eprofile so fth eothe rtypes . The KaoliniticYello w Latosol,intergrad e to Ground Water Laterite soil seemst o beidentica lwit hwha ti scalle d'Planosoli c Latosol'b yth e U.S.Soi lConservatio nSer vice. This name refers tosoil s assumed to occupy about 40%o f theare a in western Amazonia and describedb ySTRIKE R as:'ver y highly acid, with greyish and yellowish brown silty to silty clay surface horizons over yellow and reddish yellow, mottled, moderately plastic or compacted, slowly drained clay subsoils' (KELLOGG, personal communication).Se eals oProfil e 12(II.3.2.1) .

(lb)Kaoliniti cRe d Latosol(KRL ) Generalconcept: Thisuni t includesdeepl yan d stronglyweathere d soils, well-drained and permeable,predominantl y of reddish hue and ofvaryin gtextur ebu t with at least 15% clayi nth e B horizon.Th esoil sar ever ystrongl yt o slightlyacid ,an d havea bas e saturation above4 0%. Theprofile sar ewel ldeveloped ,havin ga nAB Csequenc eo fhorizons ;th eboundarie s betweenthes ehorizon sar egradua lo rdiffuse . Thesoil shav ea latosolic- Bhorizo n (cf. II.2.1); the clay-sized particles consist very predominantly of silicatecla y mineralso f the 1:1 latticestructure ,bein gkaolinite . Thesoil shav ea rathe rwea kmacro-structure , at leasti nth esub-surfac e horizon.Whe ndr ythe yar eslightl yhard ,t ohard .Th eB ho rizon isgenerall yslightl yheavie rtha n theA horizon. The only frequently encountered profiles of this unit are of mediumtexture. The rangei ncharacteristic swithi nthi stextura lclas si sgive nbelow :

The A horizon, consisting of an A, and Aj, is about 70c m thick. The Bhorizon , consisting of a B, and a B„ isabou t 200cm . The C horizon is of varying thickness. The A horizon is normally capped witha thin( < 1 cm)laye ro f loose bleached sand ('micro-podzoO.

135 The Aj subhorizon is 5-20 cm thick. The texture varies between loamy sand and light sandy clay loam. Thecolou r value is4 to 5, thechrom a 4 to 6, the hue7. 5Y R or 5 YR. Brown (7.5 YR 4/4 or 5/4) ismos t frequent. The horizon isstructureles s (singlegrains )o r hasa weak,fine subangula r blocky structure. TheA ,subhorizo n is40-6 0c m thick.Th e texture isusuall y sandyloa m or sandycla y loam. Thecolou r ispredominantl y yellowish red (5 YR 4/6 or 4/8). Weak,fine subangula r blocky isth epre dominant structure.Th e B2subhorizo n is50-10 0c m thick.Th e texture iscommonl y sandy clay loam, sometimes sandy loam, or sandy clay. The predominant colours are yellowish red (5 YR 4/6 or 4/8) and red (2.5Y R 4/6o r 4/8). Thestructur e isa weak ,mediu mt ofine subangula r blocky.Th e B,subho rizon isabou t 100c m thick.Th e texturei sligh t sandycla yloa m orsand yloam . Thecolou r isth esam e as in the B, horizon.Th e horizon isstructureles s(littl e coherent porous massive), or has a weak, me dium subangular blocky structure. The Chorizo n ispredominentl y of sandy loam texture.Th e colour is generally red (2.5 YR 4/8). The horizon is massive. Rounded quartz pebbles or rounded plinthite concretionscommonl y occur, insmal l percentages.

Profile 31. KAOLINITIC RED LATOSOL, medium textured (KRLOT); the actual profile is transitive to the light textureclas s Field description 226(Sombroek , Sampaio) Location:Alon g highway BR-14, km 429 So f Sao Miguel do Guama. Relief and drainage:Sid e of one several small hills in generally flat terrain. Somewhat excessively drained.Alt . 140 m. Parentmaterial: I n Pleistocene reworked sediments of Cretaceous/Tertiary age. Vegetativecover: Tropical forest, with low timber volume (75m'/h a ca.). Low, thintrees ;man ypalm s of variousspecies .Rathe r denseundergrowth , inwhic h thincreeper san d vinesar e common. Oi 4-1cm: Undecomposed plant residues. Oj 1-0 cm: Partly decomposed plant residues. A,i '0-0.5 cm: Loose bleached sand with manyfine roots . A,, 0.5-20cm: Dark brown (7.5 YR 4/4) loamy sand. Single grains. Moist, loose. Not sticky and not plasticwhe nwet . Many roots.Transitio n clear to gradual. A, 20-70 cm: Yellowish red (5Y R 4/8)fine sand y loam. Weak,fine subangula r blocky structure. Moist, very friable. Not sticky and not plastic when wet. Soft when dry. Many pores. Many roots.Transitio n diffuse. B, 70-180cm: Red (2.5Y R 5/8)fine sand y loam.Weak , medium to fine subangular blockystruc ture. Moist, very friable. Not sticky and not plastic when wet. Slightly hard when dry. Many pores. Many roots.Transitio n diffuse. Ba 180-300 cm: Red (2.5 YR 5/8)fine sand y loam. Massive to weak, medium subangular blocky structure. Moist, very friable. Not sticky and not plastic when wet. Slightly hard when dry. Many pores. Many roots. In A, and B, some quartz pebbles and rounded, black,mediu mtexture dplinthit e concretions, both partiallya sstone-lines .

(Ic)Kaoliniti c LatosolicSan d(KLS ) Generalconcept: This unit comprises highly and very deeply weathered soils,wel l or somewhat excessively drained,wit hcolour so fyellowis h or,les scommonly ,o freddis h hue. Thesoil sar eligh ttextured ,th e percentage ofclay-size d particlesi nth e B horizon beingless than 15 %. Theyar ever yhighl yo rextremel yaci dan d havea bas e saturation oflessthan40%. Theprofile s arewel ldeveloped ,havin ga nAB Csequenc eo fhorizons ;th e boundaries between these horizons are gradual or diffuse. The solum isgenerall y very thick (2m or more). The few clay sized mineral particles consist very predominantly of silicate clay minerals of the 1:1 lattice structure, being kaolinite. The soils are structureless (singlegrains) ,o rhav e aver ywea k structure.Th e B horizon isslightl y lesssand y than theA horizon , whichi softe n alsonoticeabl e inth e field.

136 Apart from texture and associated characteristics, the soils are comparable toth e KaoliniticYello wLatoso l (,Ortho). Two main phases aredistinguished , namely a Forest phase anda Savannah phase. Many profiles ofth eForest phase (F)wer e studied.Therefore , arang ei n characteristics canb egiven .

TheA horizon ,consistin g ofa n A,an d anA, , isabou t 70 cm thick.Th e B horizon isbetwee n 150 and 200c m thick,an d consists normally of a B,an da B».Th e Chorizo n iso f varying thickness. The A!subhorizo n is10-4 0c m thick. Its texturevarie sfro m sand tosand y loam.Th ecolou r valuei s 4 or 5,th e chroma 2o r 3,th ehu e 10o r 7.5YR ;dar k brown to brown (10 YR4/3 ) is predominant. The structure isusuall y weak fine subangular blocky.Th e subhorizon isgenerall y capped witha thin layero floos ebleache d sand. The A,subhorizo n isbetwee n 30 and 60 cm thick.It stextur evarie s from sand tosand y loam. Thecolou r value is4 t o6 ,th e chroma 3t o6 ,th ehu e is 10o r 7.5YR , less com monly 5o r2. 5 YR;brow n (7.5Y R5/4 ) ismos t common. The structure iswea k medium tofin e subangular blocky,o rlittl ecoheren t porous massive.Th e Bi subhorizon is70-15 0c m thick.It stextur ei s a loamy sand or light sandy loam. Thecolou r hue is 10o r7. 5 YR, lesscommonl y (forinstanc ei n relatively dry regions) 5o r2. 5YR ; value and chroma are 6/6o r6/ 8 (brownish yellow,reddis h yellow, sometimesligh t red). Thestructur e isnormall y little coherent porous massive,t owea k medium subangular blocky. The B, subhorizon, 50 to 100c m thick, varies between sand and sandy loami n texture. The colours arepredominantl y thesam e as in theB „ sometimes ofslightl y redder hue. The structure isusuall y porous massive.Th e Chorizo n isgenerall y somewhat lighter textured than the Bt. It has thesam ecolou r ori so fslightl y redder hue; some mottling, for instance light grey ina reddish yellow matrix, mayoccur . Thehorizo n is massive. Plinthite concretions or rounded quartz pebbles may occura tvaryin gdept h inthi shorizon .

Only afe wprofile s were studied ofth e Savannahphase (s). Except forth e develop mento fth eA t horizonan da slightl y largerdifferenc e intextur e between A and Bhori zons,th ephas eseem st ob elargel ycomparabl et oth eFores tphase .

Profile 32. KAOLINITIC LATOSOLIC SAND, Forest phase(K.LS , F) Field description 296 (Sombroek) Location:Araguai a Mahogany area,Ri oAntonino ,transec t6 ,stak e2 3 (Lat. 6° 21'S ;Long .48 °16 'W ) Reliefand drainage: Extensiveflat terrace ,abou t 10 mabov eleve lo fnearb y rivulet. Somewhat excess ivelydrained . Alt.20 0m . ca. Parentmaterial: Pleistocenesediments . Vegetativecover: Primeva l forest with lowtimbe r volume. Closed canopy of rather thin trees,an d palms.Ope n undergrowth,wit ha fe w thinclimbers .Ope nfield laye r ofseedling san d scattered clumps of grass. O, 12-2cm: Undecomposedplan t residues. O, 2-0 cm: Partly decomposed plant residues,an dfin e roots.Surfac e smooth, with scattered out casts ofparaso l ants. Au 0-0.2 cm: Loosebleache d sand. A„ 0.2-10 cm:Dar k brown (10 YR4/3 )ligh t loamy fine sand. Single grains,t oweak fine, sub - angular blocky structure. Moist, very friable. Notstick y andno tplasti c when wet. Soft when dry. Many,mainl yfine roots . Many large,an d commonfine pores .Transitio n clear. A, 10-70cm: Yellowis h brown (10Y R5/4 ) loamy fine sand. Little coherent porous massive,t o weak, medium to fine subangular blocky structure. Moist, very friable. Not sticky andno t plastic when wet. Slightly hard, to hard andsomewha t brittle when dry. Many roots. Many largean dfine pores .Transitio n gradual. Bn 70-140cm: Reddish yellow (7.5 YR 6/6) loamy fine sand. Little coherent porous massive,t o weak medium subangular blocky structure. Moist, very friable. Not sticky andno t plastic when wet.Commo n roots.Commo nfine pores .Som ekrotovina s ofA, .Transitio n diffuse.

137 WithDutch auger: Bu 140-220cm: Reddish yellow (7.5-5 YR5/8 ) loamy fine sand. Moist, very friable. Notstick y and notplasti c when wet. Few roots. Transition diffuse. B2 220-290cm: Reddish yellow (5Y R 6/8) heavy loamyfine sand . Dry. Not sticky andno t plas ticwhe n wet.A fe wfine roots .Transitio n clear. B3(K) 290-320+cm: Reddish yellow(7. 5 YR6/8 ) lighfinet sand y loam,wit hcommon ,mediu m sized, distinct mottles ofver ypal e brown (10Y R 7/4). Moist,ver yfriable . Not sticky and not plastic when wet.

Profile 33.KAOLiNrn c LATOSOLIC SAND, Savannah phase (KLS, s); theactua l profile istransitiv e to themediu m textureclas s of the KAOLINITTC YELLOW LATOSOL Field description 298(Sombroek ) Location:Araguai a Mahogany area, Araguatins, 1.5k m Eo ftow n (Lat. 5°38'S ;Long .48 °06 'W ) Relief anddrainage: Approximately flat upland, about 15m abov e levelo frivulets .Excessivel y drain ed.Alt . 150mca. Parent material: Reworked sand-stones of Sambaiba member of Pastos Bons beds (Jura-Triassic). Vegetativecover: Savannah. Shrubs and rare lowtrees . Open field layer: scattered palmlets, some grasses under theshrubs . Formerly frequently (?) burned.

Ot 2-0cm: Locally.Undecompose d plant residues. Oj Absent: Most ofsurfac e isbar ean d white sandy. An 0-5 cm: Loosebleache d sand. Al4 5-30 cm: Dark grey brown (10Y R 4/2) coarsesand . Little coherent porous massive, towea k medium subangular blocky. Some white sand around the structure elements. Moist, very friable toloose .No tstick y andno tplasti c when wet. Hard when dry; the horizon standsou t as a bank in long exposed pits. Common roots. Fewlarge , andcommo n to few, fine pores. Transition gradual. A, 30-65 cm: Light yellowish brown (10Y R6/4 ) coarse sand, with krotovinas ofbrow n (10 YR 4/3). Verylittl ecoheren t porousmassive ,t oweak ,mediu m tofine subangula r blocky structure. Moist, very friable to loose. Not sticky and not plastic when wet.Slightl y hard whendry . Common roots.Commo n largean dfine pores .Transitio n gradual. Bt 65-130 cm: Reddish yellow (7.5 YR6/6 ) loamy coarse sand. Little coherent porous massive. Consistence asi nA, .Ver y fewlarge ,an dcommo n fine pores. Few roots. Scattered pieceso f charcoal.Transitio n diffuse. WithDutch auger: B21 130-220cm: Reddish yellow (7.5 YR6/ 6-8 )ligh t coarse sandy loam. Consistence asi n A3. Common fine roots. Scattered pieces ofcharcoal . Transition diffuse. B2, 220-320cm: Reddish yellow (7.5 YR6/ 6- 8)ligh t coarse sandy loam. Consistence as in A,. A fewfine roots .

(II) REDYELLO WLATOSOL S(RL ) General concept (cf.grou p Heo fTabl e 6):Thi s group includes deeply and strongly weatheredsoils ,well-draine dan dpermeable ,wit hbelo wth esurfac elaye ra colou rtha t iscommonl yo freddis h oryellowis hhue ,hig hvalu e and highchroma .Th esoil sar eo f varying texture, but with at least 15%cla y inth e Bhorizon . Theprofile s arewel l developed,havin g anAB Csequenc eo fhorizons .Th eboundarie sbetwee n thesehori zonsar egradua lo rdiffuse . The Bhorizo ni sa latosoIic- Bhorizo n (cf. II.2.1).Thecla y fraction iscompose d ofintermediat epercentag eo fsilicat ecla ymineral so f1: 1 lattice structure, i.e.kaolinite ,an drelativel y low percentageso fF e clay minerals.Th e soils have a moderately strong macro-structure. There may besom e difference intextur e betweenth eA an dth e Bhorizon. - Only a few profiles of this group were encountered. A range of characteristics

138 of the Amazon Red Yellow Latosols cannot therefore be given. The description of only one profile follows. This profile, moreover, does not seem to bea typicalone ,i f onlybecaus eo fit scolour .

Profile 34. RED YELLOW LATOSOL(? ) (RL?) Field description 308(Sombroek ) Location:Serr a deNavio ,Amap aTerritory ; topo f ICOMI manganesemine . Reliefand drainage: Slope (15 %) in hilly upland. Alt. 300m . ca., well-drained. Parentmaterial: Undefined rock ofcrystallin e basement. Vegetativecover: Primeval forest, of rather high timber volume. A, 0-40 cm: Dark brown to brown (7.5 YR 4/4) light clay. Moderate to strong, veryfine suban - gular blocky, or granular, structure. Moist, very friable. Sticky and plastic when wet. Soft whendry . Manypores . Many roots.Transitio n diffuse. A, 40-100cm: Reddish brown (5Y R 4/4)clay . Moderate,ver yfine subangula r blocky, or granu lar,structure .Consistenc e asi nA, . Manypores . Many roots.Transitio n diffuse. B2 100-200+ cm: Reddish brown (5 YR 4/4) clay. Moderate, very fine subangular blocky, or granular, structure. Consistence asi nA t. Many pores. Many roots. In all horizons small (diameter < 5cm) ,hard , round, black concretions of Mn (?),i n a quan tityo f 5-10%.

(Ill) DARK RED LATOSOLS (DL) General concept (cf. group lib of Table 6): This group includesdeepl yan dstrongl y weathered soils,well-draine dan dpermeable ,wit hbelo wth esurfac elaye ra colou rtha t is commonly of reddish hue, low value and high chroma. The soils are of varying texture, but with at least 15% clay inth e Bhorizon .Th e profiles are welldeveloped , having an ABC sequence of horizons. The boundaries between these horizons are gradualo rdiffuse . The B horizon isa latosolic- Bhorizo n (cf. II.2.1).Th ecla y fraction is composed of intermediate percentage of silicate clay minerals of the 1:1 lattice structure, beingkaolinite ,an d intermediatepercentage so f Fecla yminerals .Th esoil s have a moderate to strong macro-structure.Ther e is normally no textural difference betweenth eA andth e Bhorizon . Only a limited number of profiles with Dark Red Latosol characteristics were en countered. A description of only one profile is given. This profile is moreover not typicalfo r thegroup ,becaus e of its hsallowness,an d the considerable percentageso f siltpresent .

Profile 35. DARK RED LATOSOL,Shallo wphas e(DL ,s ) Field description 279(Sombroek , Sampaio) Location:Araguai a Mahogany area, Rio Corda, transect 8,stak e O(Lat . 6°23'S ; Long.48°20 'W) . Relief and drainage:Edg e of almost flat terrain, about 10-15 m above level of nearby river. Well drained. Alt.20 0m ca. Parentmaterial: Dark shaleso f Pimenteiras beds(Devonian) . Vegetativecover: Primeval forest. Open canopy ofa few thick trees,lade n withclimbers ,an d scattered palms.Dens e undergrowth, with manycreeper san d climbers.Dens efield layer ,consistin g largely of a fern species. O, 4-2 cm: Undecomposed plant residues. O, 2-0 cm: Partly decomposed plant residues, with many fine roots. At the surface some worm outcasts.

139 A, 0-15 cm: Dark reddish brown (2.5 YR 3/4) light clay loam. Structure of moderate, very fine granules, composing weak, fine subangular blocks. Moist, very friable. Slightly sticky and slightly plastic when wet. A few, small (diameter < 5 cm) plinthite concretions, similar to those inA, .Ver y many roots.Transitio n diffuse. A, 15-40cm: Dark reddish brown (2.5Y R3/4)cla y loam. Moderate,ver yfine granula r structure. Moist, very friable. Slightly sticky and slightly plastic when wet. About 20% small (diam. < 5 cm), to a degree platy, hard, very fine grained, plinthite concretions, brownish yellow (10 YR 6/8) in their outer parts, black (5 YR 2/1) in their centres. Many roots. Transition gradual. Ba 40-80 cm: Dark red (2.5 YR 3/6) clay loam. Structure of moderate to strong, very fine gran ules, composing weak,fine subangula r blocks. A few, faint clay skins, for a large proportion on the surface of the concretions. Moist, very friable. Plastic and slightly sticky when wet, About 70%plinthit e concretions, similar to those in A,, but slightly larger (diam. < 10cm) , lesshard , and with more brownish yellow. Many roots. Transition gradual, irregular. C 80-120+cm: Dark red(2. 5Y R3/6 )heav ycla yloam .Structur ean dconsistenc ea s in B,.Abou t 95% large (diam. 10-20 cm), platy, rather soft, fine grained, black and yellow concretions (weathering dark shale?).

For thep Fcurv eo fth e B2se eFig . 37.

(IV) REDYELLO WPODZOLI CSOIL S (RP) General concept: According to BALDWIN, KELLOGG and THORP (1938) and THORP and SMITH(1949) ,th e modal Red Yellow Podzolic soilsar e well-drained soils witha thinorgani cto pove ra wel l developedligh ttexture dsurfac ehorizo n( Ahorizon )an da heavytextured ,block yo rprismati csubsoi l( Bhorizon) ,wit h silicatecla yskin spresent . ThisB horizo nofte n hasa re do ryello wcolou ri nth euppe rpar tan da redde rcolou ri n the lower part, while at the transition zone with the C horizon a mottling is often present. Accordingt oth eelaboratio no fthi sconcep ta sapplie di nBrazi l(cf. LEMOS, BENNEMA, SANTOS et a!., 1960) the clay-sized minerals should consist of silicate clay minerals mainlyo f1: 1 latticestructure ,an diro n oxides.Th elatte rar egenerall ypresen ti nsmal ler quantity than in most Latosols, and there are few if any aluminum oxides. The B horizon should have the characteristics of a textural-B as described in II.2.1. In Brazil, the Red Yellow Podzolic soilsar e subdivided in those of low base status,an d thoseo fmediu man dhig hbas estatus .

(IV.l) RedYellow Podzolic soil (,Ortho) (RP) Underthi shea dar edescribe dthos eRe dYello wPodzoli csoil so fwhic hth eB horizon hasabou tal lth echaracteristic so fa textural-B . Only a few profiles were encountered of those with lowbase saturation. Therefore onlya profil edescriptio ni s given. Many profiles were encountered locally of those with rather highbase saturation. Thereforeth efollowin g generalisationca nb egiven :

The A horizon, consisting of an A, and an A„ isabou t 40c m thick. The B horizon, consisting of a B, and sometimes a B3, is4 0 to 80c m thick. The C horizon isabou t 30cm . Below this, unweathered rock (mica schists) occurs. The A, subhorizon is 5-10 cm thick. Its texture is usually a loam. The colour value isabou t 4, itschrom a 2 or 3,it s hue 10,7. 5 or 5 YR; most common is dark brown (7.5

140 YR 4/2). Thestructur e is amoderate ,mediu m tofin e subangular blockyan d finegranular . TheA ,sub - horizon is about 30cm . Its texture is predominantly clay loam. Its colour value is 4 to 6, its chroma also4 t o 6,it shu e7.5, 5o r 2.5YR ; most common isreddis h brown (5Y R 4/4). The structure is usual ly a moderate, medium subangular blocky (in part composed of moderate, very fine subangular to angular blocks). Clay skins are few and faint. The B, subhorizon is about4 0cm . Its texture isa clay, lessfrequentl y a clay loam.Th e colour value is4 or 5,th e chroma usually 6, the hue 5Y R or 2.5 YR; most common isre d(2. 5 YR4/6). Thestructur e isnormall y moderate, medium angular, to subangular blocky (in part composed of very fine angular blocks). Clay skins are common to continuous, and distinct. The B, subhorizon, where existent, has usually a clayo r clay loam texture. Its colour is approximately identical to that of the Ba. The structure is somewhat weaker. Clay skins are mostly common, and faint to distinct. The C horizon has a loam or clay loam texture. Its colours are varie gated red, black and reddish yellow.Th e horizon isnormall y massive. Fine shiny flakes of mica are found in all horizons. Angular quartz stones are often present.

Profile 36. RED YELLOW PODZOLIC SOIL(,Ortho) , with low base saturation (RPti>) Field description 320(Sombroek , Falesi) Location: Rio Branco do Acre, km 0.5 W of town Reliefand drainage: Low top ingentl e undulating terrain, Alt.les stha n 100 m. Well-drained. Parent material:Tertiar y (?)clay-stone . Vegetativecover: Secondar y shrub. A, 0-3 cm: Dark brown to brown (10Y R 4/3) loamwit hcommon ,fine, distinc t mottleso f strong brown (7.5 YR 5/6). Moderate, medium subangular blocky structure. Moist, friable. Slightly sticky and slightly plastic when wet, slightly hard when dry. Many large, and few fine pores. Many roots. Activity of earth worms.Transitio n clear and smooth. A, 3-26 cm:Stron gbrow n (7.5Y R 5/8) loam.Weak ,mediu msubangula r blockystructure . Moist, friable. Slightly sticky and slightly plastic when wet, hard when dry. Few largean dfine pores . Few roots.Activit y ofeart h worms.Transitio n clear and wavy. B, 26-60 cm: Yellowish red (5 YR 5/8) clay loam. Moderate, medium subangular to angular blocky structure. Clay skins continuous, faint. Moist, firm. Sticky and plastic when wet, hard when dry. Common largean dfine pores . Few roots.Transitio n diffuse and smooth. B, 60-85cm: Yellowish red (5 YR 5/8)clay . Moderate tostrong ,mediu man dfine angula r blocky structure. Clay skins continuous, faint to moderate. Moist, firm. Sticky and plastic when wet, hard when dry. Few large, and common fine pores. Few roots. About 2% small (diam. < 0.5cm) ,round , hard,fine graine d plinthiteconcretions .Transitio n gradual and smooth. Bit 85-130 cm: Light grey(2. 5Y 7/2 )heav yclay ,wit h many, medium sized, distinct to prominent mottles of dark red (10 R 3/6) and strong brown (7.5 YR 5/8). Weak to moderate, medium subangular, toangula r blocky structure. Clay skinscommon , very faint. Consistence as in B,. Fewpores .N o roots.Transitio n gradual. Cg 130-300+ cm: White (2.5 Y 8/2) heavy clay, with many, coarse, prominent mottles of red (2.5Y R 4/8)an dyellowis h brown(1 0Y R 5/8). Massive.Consistenc e as inB, .Materia l isuse d for brick making. No IQAanalysi swa sexecute d onthi sprofile , but the SoilsLaborator y of the Royal Tropical Institute inAmsterda m (Holland) kindly provided for analytical data.

Full mineralogical analysis was carried out on the B2 horizon, by the Netherlands SoilSurve y Institute,Wageninge n(Holland) .Th edat a onsampl e320- 4i nth eTable s7 and 8,an d theFigs . 14a-e showtha tth emineralogica l composition isindee d distinctly different from that of latosolic profiles. In the clay fraction even 33 % mica (illite) occurs, more than kaolinite (24%); iron oxides comprise 12%an d aluminum oxides 7%. The remainder iscomprise d of clay-sized quartz (24%). In the fraction 2-16 mi cron some felspar (8 %) and chlorite (3 %) are present. The coarser fractions contain onlyquart z(90% )an d hematite(10%) .

141 For both the A2an d the B2 horizons also the specific surface was determined (samp les320- 2an d 320-4o f Table 7). Aspecifi c surface of 100 m2,o n wholesoi lbasis ,give s about 18m.e .catio n exchange capacity, which isconsiderabl y more than in the Lato- solprofile s(cf. 11.2.2). For both theA 2an d B2horizon s ap F curvewa s determined on undisturbed samples. The A2 proved to have 13.5vol . perc. available moisture (50.1 %moistur e at pF 0.4; 32.4% atp F 2.0; 18.9%a tp F 4.2), the B2onl y8. 6 vol. perc.(50.4 %moistur e atp F 0.4 36.9%atpF2.0;28.3%atpF4.2);c/.V.3.1.1.

Profile 37. RED YELLOW PODZOLIC soil (.Ortho), with rather high base saturation (RPr *t>) Field description 282(Sombroek , Sampaio) Location:Araguai a Mahogany area, Santa Isabel,transec t 10, stake 11 (Lat.6"08 'S ; Long.48°22 'W) . Relief anddrainage: Gentle sloping side of hill in undulating terrain. Well-drained. Alt. 200m ca. Parentmaterial: Micaceous schists with veinso f quartz, of Pre-Cambrian (Arquean) age. Vegetative cover:Primeva l forest of very low timber volume. Open canopy, of only a few trees, and somepalms . High undergrowth, consisting largely of a dense network ofcreeper s and climbers. Very dense field layer,consistin g predominantly of a fern species. O, 4-2 cm: Undecomposed plant residues. O, 2-0 cm: Partly decomposed plant residues,wit h fine roots and fungi. About half of the surface is covered withver ydar k greywor m outcasts. A, 0-8 cm: Dark reddish brown (5 YR 3/3) light loam. Structure of weak to moderate, fine gra nules, composing moderate, medium to fine subangular blocks. Moist, very friable. Slightly sticky and slightly plastic when wet. Very many large and fine roots. Many pores. Common veryfine shin y flakes (mica).Transitio n gradual, smooth. A, 8-40 cm: Reddish brown (2.5 YR 4/4) clay loam. Structure of weak to moderate, very fine subangular to angular blocks,composin g moderate, medium tofine subangula r blocks. Com mon, faint clay skins. Moist, friable. Sticky and plastic when wet. Many large and fine roots. Common fine pores. Common veryfine shin yflakes (mica) .Transitio n gradual, smooth. B2 40-80 cm: Red to dark red (2.5 YR 4-3/6) clay. Structure of moderate, very fine angular blocks, composing weak to moderate, medium subangular to angular blocks. Continuous, distinct clay skins. Moist, friable to firm. Sticky and plastic when wet. Common roots. Com mon,fine pores .Common , veryfine, shin y flakes (mica). Transition gradual, smooth. From 30 to 70 cm many (75 % ca.), small to medium sized (2-20cm) ,ver y hard, angular, milky white stones of quartz, the surfaces of which arecoate d with clayskins . B, 80-110 cm: Red (2.5 YR 4/6) clay. Structure of moderate, very fine subangular to angular blocks,composin gwea kt omoderate ,mediu mt o coarse,subangula r to angular blocks. Conti nuous, faint to distinct clay skins. Moist, friable. Sticky and plastic when wet. Few roots. Few fine pores. Many fine shiny flakes (mica). A few quartz stones similar to those in B2. Transition gradual, irregular. C 110-140+ cm:Weatherin gmicaceou sschist :Dar k red(1 0R 3/6 )an d black( N3/0 )sof t rock of heavy loam texture,horizontally veryfinely stratified . Angular milky quartz stones embedded ina scattered way.

For the B2horizo n ap Fcurv ewa sdetermine d on aundisturbe d sample.Th e amount of available moisture proved to be9. 0 vol.perc . (51.1 %moistur e at pF 0.4; 42.1 %a t PF2.0;33.1%atpF4.2);

(IV.2) Red Yellow Podzolicsoil, intergrade toKaolinitic YellowLatosol (RP-KYL ) Underthi sheadin gar e described Red YellowPodzoli csoil swhic hhav ea B horizon whose characteristics mostly belong to those of the textural-B, while a minor part of

142 them are characteristics of the B horizon of the Kaolinitic Yellow Latosol. Those belongingt oth etextural- Bar eth e following: 1. agradua l toclea rboundar y between theA an d the B horizons, 2. adistinc t change incolou r from theA t o the Bhorizo n or the presence ofa transi tional horizontha t ismottled , 3. adistinc ttextura ldifferenc e betweenth eA an dth e Bhorizons , 4. compactnessan d ratherfirm, t ofirm consistenc eo fth eB horizon , 5. alimite dporosit y and rootingi nth e B horizon. Ast oth echaracteristic so fth e B horizontha t belongt othos eo fth elatosolic-B ,ther e isi n particular an absence of a welldevelope d angular blocky or piismatic structure. Clayskin sar ewea k and discontinuous,whe npresen t at all.Th emineralogica lcompo sition ofth ecla yfractio n islargel y comparable to that ofth e Kaolinitic Yellow Lato sol. Thebas esaturatio n ofth esoil si slo wthroughou t theprofile s( < 40%). Plinthitecon cretions areabsen t or occur onlyi nmino r quantities.Th egrou po fsoil si ssubdivided , for practical mapping purposes, according to the texture of the Bhorizon . Thequali tieso f thesoil sar edeal twit hi nV.3 . Manyexample swer eencountere d and studied ofbot h veryheavy texturedand rather heavy texturedprofiles . For both variants, therefore, a range in characteristics can be given.

The very heavytextured profile s have an A horizon of40-5 0c m thick,consistin g of an A, and an A, subhorizon. The Bhorizon , which consists generally of a B, a B„ varies between 100an d 150c m in thickness.Th e thickness of the Chorizo n is undetermined. The A horizon lacks a capping with loose bleached sand. The A subhorizon is 2-5 cm thick. The texture ispredominantl y lightcla yo r heavy sandy clay.Th ecolou r isusuall y dark grey brown (10Y R 4/2). A moderate,fine subangula r blocky is the common structure. The transition to the A, is gene rallyclear . The A, subhorizon isabou t 40-50c m thick. The texture isa clay or heavyclay .Th e colour hue is 10YR , the value varies between 5an d 7,th e chroma between 6an d 8. Most frequent is brown ish yellow (10Y R 6/6). The structure isusuall y moderate,fine subangula r blocky and weak, veryfine granular. The transition to the B, is gradual or clear. Several times a thin (10-20 cm) transition zone occurs,wit h colours both of theA t and the B,.Th e Bs subhorizon isbetwee n 50an d 70c m thick. The texturei sa heavyclay .Th emai ncolou r hasa hu eo f7. 5YR ,a chrom a of 8 and a value between 5 and 7. Reddish yellow (7.5 YR 6/8) is most frequent. The profiles have normally common, fine, distinct mottles of red (2.5 YR 5/8). The structure is normally moderate, medium to fine subangular blocky. The presence of a few, faint clay skins is general. The transition to the B, is gradual to diffuse. The B, subhorizon varies between 40an d 80c m in thickness.Th e texture isa heavyclay .Th e main colour isyellowis h red (5Y R 5/8, 5Y R 6/8). Secondary coloursar e light red (2.5Y R 6/8) and yellow(1 0 YR 8/6), occurring asfe w tocommon ,fine, distinc t mottles.Th e structure ismoderate , medium subangu lar to angular blocky.Th estructur eelement sar ecovere d withcommo n faint clayskins .Th e transition to the C is diffuse or gradual. The C horizon has predominantly a clay texture. The colour is red or light red (2.5 YR 5/6 or 5/8, resp. 2.5 YR 6/6 or 6/8). The structure ispredominantl y weak, medium angular to subangular blocky.Cla yskins ,whe n present, are fewan d very faint. Rather common is the presence, in small quantities (5% ca.), o f very small fine grained, red or dark red plinthite concretions. They are hard in the A horizon, half hard in the B horizon. A few larger concretions may occur. The ratherheavy textured profiles have an A horizon of 40 to 70c m thickness which consists of an A, and an A,. The Bhorizon ,consistin g of a B„ and a B„, isbetwee n 100an d 150c m thick. Between theA an d the Bhorizo n a transitional horizon AB.o f 20-50c m thickness,occurs .Th e Chorizo n iso f undetermined thickness.

143 The Aj subhorizon ismostl y 5-10 cm,sometime s 10-20c m thick. The texture ispredominantl y light sandy clay loam, butsand y loam andeve n loamy sand textures may be found. The colour hue is 10 YR, the value4 o r5 ,th e chroma 3,4 o r6 . Yellowish brown (10 YR5/4 ) ispredominant . The struc ture isa weak tomoderate ,fine subangula r blocky.Th e transition toth e A, isclear . The A, subhori zon varies between 20an d 60c m in thickness.Th e texture ispredominantl y aheav y loam clay orligh t sandy clay, butals o sandy loams dooccur . The colour huei s 10 YR, thevalu e 5o r6 ,rarel y 7, the chroma 4o r6 . Brownish yellow (10Y R 6/6) isth e most common. The structure iswea k tomoderate , medium tofine subangula r blocky. The transition to theA B isclear . The AB horizon, 20t o 50c m thick, haspredominantl y lightsand y claytexture ,bu t light clay orsand y clay loam texturesals o occur. The main colour is normally identical totha t ofth e A,, namely brownish yellow (10 YR 6/6). Secon dary colour is yellowish red( 5Y R 4/8, 5/8) orre d (2.5 YR 4/6 or5/6 , sometimes 2.5 YR 4/8 or 5/8), occurring ascommon , medium, distinct mottles. Inth e upper section ofth e horizon very small, hard, round plinthiteconcretion s(5-2 0% o f thesoi lmass )occu r rather often. The structure isusuall ya wea k to moderate, medium subangular blocky. Thetransitio n toth eB s, isgradua l toclear . TheB „ subhorizon varies between 40an d 80c m inthickness .Th e texture isa sandy clay or clay. The colour hue is5 YR , sometimes 2.5Y R or 7.5YR , thevalu e is5 o r6 ,th echrom a 8.Th ecolou r is therefore predo minantly yellowish redo rreddis h yellow. The structure iswea k tomoderate , medium subangular to angular blocky. Clay skins are few and faint, if present at all.Th e transition toth e B2t isdiffuse . The B2, subhorizon varies between 20an d7 0 cmi n thickness. Thetextur e isa clay or sandy clay.Th e colour isre d (2.5Y R 4/8 or 5/8). Thestructur e isa weak , medium subangular to angular blocky. Clay skins are normally absent. The transition to the Ci sdiffuse . The Chorizo n was examined at only a few profiles. There it showed a sandy clay texture, anda massive structure. The colour was red (2.5 YR 5/8) with common, medium sized, distinct mottles ofre d (10R 4/8 ) and yellow (10 YR 7/8). Included inth e unit are profiles with many, rather largeplinthit e concretions concentrated inth eA B horizon,bu t withotherwis e similarcharacteristic s as described above. The concretions are apparently colluvial, incontras t toth erathe r few, very fine concretions which occur commonly inth eA B and whichar e believed tohav ebee n formed insitu (cf. II.3.2.2).

Profile 38. RED YELLOW PODZOLIC SOIL, intergrade to KAOLINITIC YELLOW LATOSOL, very heavy text ured (RP-KYUA) Field description 236(Sombroek , Sampaio) Location:Alon ghighwa y BR-14,k m23 2S o fSa o Miguel do Guama. Reliefand drainage: Flat stretch ofgentl e undulating terrain, about 10m abov e level ofnearb y inter mittent rivulet. Moderately well-drained.Alt . 140 m. Parentmaterial: InPleistocen e redeposited material from Plio-Pleistocenelacustrin e origin (reworked Belterraclay) . Vegetative cover: Primeval forest, with medium timber volume (150 m*/ha ca.). Open undergrowth, nearlydevoi d ofcreeper san d climbers.

O, 12-2cm: Undecomposed plant residues. O, 2-0cm: Partly decomposed plant residues,wit h manyfine roots . Surface rather irregular and hard,slightl ycruste d due to intense termite activity. A, 0-2cm: Dark greybrow n (10Y R4/2 )clay .Moderate fine, subangula r blockystructure . Moist, friable tofirm. Stick yan d plastic when wet. Slightly hard when dry. Many pores,an d large in sectchannels . Verymany ,mainl yfine, roots .A few ,ver ysmall ,fine grained , dark red plinthite concretions.Transitio n abrupt. A, 2-40 cm:Yellowis h brown (10Y R5/6 )heav y clay. Moderate, fine subangular blockyan d weak to moderate, very fine granular structure. Moist, friable. Sticky andplasti c whenwet . Slightly hard when dry. Many pores. Many roots. A few plinthite concretions, similar to those inA, .Transitio n gradual. B2 40-100cm: Strong brown (7.5Y R5/8 ) heavyclay ,wit h common, fine, distinct mottleso fre d (2.5Y R5/8) . Weak tomoderate , medium tofine subangula r blocky and weak, fine granular structure. Afew , faint clay skins.Slightl y compact, tocompact . Moist, friable tofirm. Stick y and plastic when wet. Hard when dry. Common pores.Commo n roots.Transitio n gradual. B, 100-180cm: Yellowish red (5Y R 5/8) heavyclay .Wea k tomoderate ,coars et omedium ,sub - angular and angular blockystructure .Tendenc y toprismatic .Common , faint clayskins .Com -

144 pact. Moist,firm. Stick yan d plastic whenwet . Hard when dry. Fewpores . Few roots.Transi tion diffuse. C 180-230+cm: Red(2. 5 YR5/8 )heav yclay .Weak ,mediu msize dsubangula ran dangula r blocky structure. Slightly compact. Moist, friable tofirm. Stick y and plastic when wet. Slightly hard when dry. Common pores. Fewroots .

Profile 39. RED YELLOW PODZOLIC SOIL, intergrade to KAOLINITIC YELLOW LATOSOL, rather heavy textured(RP-KYLrt ) Field description 237 (Sombroek, Sampaio) Location: Along highway BR-14,k m 205.8S o f Sao Miguel do Guama. Relief anddrainage: Nearly flat remnant of terrace, about 30m abov e level ofnearb y stream. Well- drained. Alt. 101 m. Parentmaterial: Pleistocenefluviatile sediments . Vegetative cover: Primeval forest, ofmediu m timber volume (130m'/ha) . Rather dense undergrowth, with several thin creepers and climbers. Oi 6-1 cm: Undecomposed plant residues. O, 1-0 cm: Partly decomposed plant residues with fine roots, and some loose bleached sand. A, 0-10 cm: Dark brown (10 YR3/3 ) sandy loam. Weak to moderate, fine subangular blocky structure. Moist, very friable. Not plastic and slightly sticky when wet. Soft when dry. Many pores.Ver yman y roots.Transitio n clear. A, 10-60cm: Brownish yellow (2.5 Y-10 YR 6/6) sandy clay loam. Weak tomoderate , medium tofine, subangula r blocky structure. Moist, friable. Slightly sticky andslightl y plastic when wet. Slightly hard, to hard when dry.Scattere d (2%)ver y small,fin e textured, dusky red plinthiteconcretions . Many pores. Many roots.Transitio n gradual. AB 60-110cm: Yellow(1 0Y R 7/8)fine sand yclay .Wea kt omoderate ,mediu m tofine subangula r blocky and weak, very fine granular structure. Moist, friable. Slightly sticky and slightly plas tic when wet. Slightly hard, to hard when dry. About 20% plinthite concretions, similar to those in At. Manypores .Commo n roots.Transitio n clear. B2l 110-160cm: Reddish yellow (5 YR6/8 ) fine sandy clay, with fewt ocommon , fine, distinct mottles of yellow (10Y R 7/8). Weak to moderate, medium sized subangular and angular blocky and locally weak,ver y fine granular structure.A few , faint clayskins . Moist,friabl e to firm.Slightl y stickyan d slightly plasticwhe n wet. Hard when dry. Common pores. Fewroots . Less than 1 %plinthit e concretions, similar to those inA, . Transition gradual. Ba 160-220+cm: Light red (2.5 YR 6/8) heavy fine sandy clay.Weak , medium sized subangular and angular blocky structure. Moist, friable tofirm. Slightl y sticky and slightly plastic when wet. Hard whendry . Fewpores . No roots.

For the pF curveo fth e B2horizo n ofa similar profile: seesampl e 213-3o f Fig.37 .

(IV.3) RedYellow Podzolic soil,intergrade to Kaolinitic YellowLatosol, Concretio naryphase (RP-KYL ,CR ) Thissoi li ssimila r totha t of unit(IV.2) ,excep ttha tth eA horizo n contains consider able amounts ofgravell y hard plinthite concretions. Moreover, the Bhorizo n isgene rallyo f heavytexture ,an d hasabundan t prominent mottlingo f red, yellowan d white, although thesoi li swel lo r moderately well-drained.Th estructura ldevelopmen t ofth e B horizon is normally slightly nearer to that of a textural-Bhorizo n than that of the (IV.2) unit. The soil has a low base saturation percentage (< 40%). The plinthite concretions areo fvaryin g size,form , colour, grainage and arrangement. The mottling inth e Bhorizo n iso fvaryin gpattern .Include d inth esoi luni t areprofile s inwhic h the concretions start onlya tsom edept h below the surface (maximally 50c m arbitrarely). Thegenesi so fth e soil isdiscusse d inII.3.2.2 ,i nwhic h section additional profile des-

145 criptions are given. Reference may be made also to 1.4.5.Th e qualities of the soil are dealtwit h inV.3 . Manyprofile s ofth esoi lwer eencountere d and studied,enablin gth e following generalisationst o bemade :

The A horizon, consisting of an A, and an A2, comprises about 80 cm. The B horizon, consisting normally of a B„ aB ,an d aB „ varies in thickness between 200an d 250cm . The C horizon is often many meters thick. The A, subhorizon, 5-10 cm thick, has commonly asand y clay loam texture, but also sandy loam, sandy clay or clay textures occur. The colour is predominantly brown (10 YR 5/2). The structurei s usually awea k to moderate, fine subangular blocky. Loose, hard plinthite concretions occur in per centages of 25-90%o f the total soil mass.Th e transition to the A, isclea r or gradual.Th e A, subhori zon, 50-70c m thick,consist s mostly of a sandy clay,sometime s ofcla y or heavyclay . The colour hue is 10Y R or 7.5YR , thevalu e5 o r 6,th echrom a 6 or8 .Brownis h yellow(1 0Y R 6/8) is predominant. Thestructur e isa moderate ,fine subangula r blocky and weak,ver yfine granular . Loose,har d plinthite concretions comprise 50-80% of the soil mass. The transition to the underlying horizon is gradual. The B, subhorizon, if present, is about 50c m thick. Its texture isgenerall y acla y or heavy clay. The main colour ispredominantl y reddish yellow (7.5 YR 6/8). Secondary colours,occurring as common, medium sized, distinct mottles, are red (e.g.2. 5 YR 5/6) and yellow (e.g. 10Y R 7/8). The structure is predominantly moderate, mediumsubangula r to angular blocky. Faint clayskin sar ecommon , occur ring especially at the surfaces of the concretions. Half loose, hard plinthite concretions comprise 25-50% of the soil mass.Th e transition toth e B2 is gradual. The B, subhorizon varies in thickness between 50an d 100cm .Th e texture isa cla yo r heavyclay . The maincolou r isnormall y yellowish red (5 YR 5/8) or red (2.5Y R 5/8). Secondary colours,occurring as many,coarse ,prominen t mottles, are red(1 0R 5/8) ,dusk yre d(7. 5R 3/4) ,yello w(1 0Y R 7/8)an d white( N 8/2).Th estructur e isa moderate, medium tofine subangula r and angular blocky,th e latter sometimescomposin g a weak, medium pris matic. Clay skins are usually fewan d faint. The transition toth e B, is gradual. The B, subhorizon varies in thickness between 50 and 150cm .Th e texture is mostlya clay ,sometime s asand y clay. The colours are generally identical to those in theB2. The structure is usuallya weak, medium subangular toangula r blocky. Clay skinsar e normally absent. The transition to the C is gradual or diffuse. The C-horizon has mostlya sand y clay, sometimes asand y clay loam texture. The predominant colour is dark red (e.g. 10 R 3/6). Secondary colours are often present, as mottles of variable quantity, size and contrast. The horizon ismassive .

Profile 40. RED YELLOW PODZOLIC SOIL, intergrade to KAOLINITIC YELLOW LATOSOL, Concretionary phase(RP-KYL , CR); typeo fconcretions : Maed o Rio Field description 238(Sombroek , Sampaio) Location:Alon g highway BR-14, km 43.8 So f Sao Miguel do Guama. Relief and drainage:Sid e oflo w hill ingentl e undulating terrain. External drainage good; internal drainage slightlyimperfect . Alt. 50m . Parentmaterial: Fluviatil esediment s of Late Miocene(? )age . Vegetative cover: Primeval forest, of rather high timber volume(150-20 0m*/ha) . Rather dense under growth,consistin g partly ofcreeper san d climbers,bot h predominantly thin.

Oi 11-1cm: Undecomposed plant residues. 02 1-0 cm: Partly decomposed plant residues,wit h fungi andfine roots . Ai 0-10cm: Brown (10Y R 5/3) sandycla yloam . Weak,fine subangula r blocky structure. Moist, very friable. Slightly sticky and slightly plastic when wet. Soft when dry. About 70%loose , hard plinthiteconcretion so fvariou sforms ,size san d grainage1. Manypores .Ver yman y roots. Transition gradual. A2 10-80cm: Strong brown (7.5 YR 5/8) clay. Moderate,fine subangula r blocky and weak, very fine granular structure. Moist, friable. Plastic and sticky when wet. Soft when dry. About 75% loos e hard plinthite concretions of various forms, sizes and grainages (in lower part pre dominantlyfine blocky) .Th econcretion s are very mixed, without specific arrangement. Many pores. Many roots.Transitio n clear to gradual.

146 B, 80-160cm: Yellowis h red (5Y R 5/8)clay ,wit h many,coarse ,prominen t mottleso f dusky red (7.5 R 3/2), red (10 R 5/8), yellow (10Y R 7/8) and white (5 YR 7/1). Mottling partly in well defined horizontal stripes,partl y in poorly defined vertical pipes. Moderate, fine subangular and angular blockystructure . Tendency toprismatic . A few, faint clay skins. Moist,ver yfirm. Sticky and plastic when wet. Hard when dry, locally, namely the dusky red, very hard. Com pact. Much resistance to penetration with hammer. Very few pores. Few roots. Transition diffuse. Within this plinthitic soil mass a few sharp boundered pockets occur (diam. 20 cm ca.) of yellowish red (5 YR 5/8) clay with a moderate, fine subangular blocky and weak, very fine granular structure. Moist, friable. Slightly hard when dry. Many pores. Many roots. Probably the result of local activity of termites in the plinthite. B, 160-250 + cm: Reddishyello w(SY R 6/8)clay ,wit h many,coarse ,prominen t mottleso f dusky red (7.5 R 3/2),yello w (10 YR 7/8),whit e (5 YR 7/1) and red (10 R 5/8). Mottling poorly de finedlaminar . Massive to weak, medium to fine, subangular and angular blocky structure. Moist,ver yfirm. Stick y and plastic when wet. Hard when dry. Compact. Moderate resistance topenetratio n with hammer. Nopores .Ver y few roots.

For mineralogical analysis of the lower part of the C horizon of a similar profile: see sample 302-4o f the Tables 7an d 8, and Fig. 14a-e.

(V) REDYELLO WMEDITERRANEA NSOIL S (RM) General concept (reference be made to BARROS, DRUMOND, CAMARGOet ah, 1958; LEMOS, BENNEMA, SANTOSet a!., 1960):Th e Red Yellow Mediterranean soilsar erela tively shallow soils,well-draine d and moderately weathered.Th e profile has an ABC or ABR sequenceo fhorizons .Thes ehorizon sar edistinctl ycontrastin gan d thetrans itionbetwee nthe mi sclear .Th eA horizo n isthin ,consistin go fa nA lt andoccasional lyals o an A3.Th e horizon isdar k coloured and hasofte n a granular structure.Th eB horizon is a textural-B (cf. II.2.1), with well developed angular blocky or prismatic structure.I t isnormall y reddishcoloure d and clayey.It sbas e saturation ismediu m to high. The lower part of the B horizon and the C horizon contain still appreciable amounts of easily weatherable primary minerals.Th e silicatecla y minerals of thesoi l areo fth e2: 1lattic estructur efo r agoo dpart . Few of such profiles were encountered in Amazonia. None contained all the above characteristics.Th eA horizon wasofte n thick (moretha n 20cm) ,consistin geithe r of a thickA jonly ,o rhavin ga subhorizo nwit hth eappearanc eo fa nA 2,i nwhic hplinthit icgrave lwa sa rathe rcommo nfeature .Th etextur eo fth eB wa smor eofte n loamytha n

') Notepage 146 The following concretions are present: Large (>20 cm diam.), irregularly shaped, massive blocks; dusky red (10 R 3/4); coarse grained: iron cemented quartz grains. Medium sized (ca.2 0c m diam.), angular, massive blocks; dusky red (7.5 R 3/2), light red (10 R 6/8) and black (N 2/0); medium grained. Medium sized (ca. 10c m diam.),platy ;alternatel y dark red (2.5 YR 3/6) veryfine grained , and black (N 2/0) coarse grained. Medium sized (ca.2 0c m diam.), vesicular; very dusky red (7.5 R 2/2) and red (10 R 4/8); predominantly fine grained. Rather small (ca. 5 cm diam.), prismatic; dusky red (7.5 YR 3/3); predominantly fine grained. Very small (0.5-1 cm diam.),subangula r blocky; dark red; predominantly fine grained.

147 clayey, and the horizon sometimes showed considerable mottling. The amount of easilyweatherabl eprimar ymineral sseeme d often rathersmall . A portion ofth e profiles encountered are likely tobe ,i nfact , intergrading toeithe r Reddish Prairie soil ort o Red Yellow Podzolic soil and Ground Water Laterite soil respectively.

Profile 41. RED YELLOW MEDITERRANEAN-LIKE soil (RM) Field description 225 (Sombroek, Sampaio) Location:Alon g highway BR-14, km457. 5S o fSa o Miguel do Guama (km 9N o fImperalriz) . Relief anddrainage: Low hill topi n irregular, undulating terrain. About 10m above level of nearby rivulets.Alt . 120m . External drainage good. Internal drainage slightly imperfect. Parent material:Intimatel y interbedded shales and silt-stones, belonging to the Cod6 beds (Middle Cretaceous). Vegetativecover: Original vegetation largely destroyed. At present lowshrub s anda few scattered palms.

Oi-02 5-0cm: Undecomposed and partly decomposed plant residues,an d some loosefine sand . A, 0-15 cm: Dark brown (7.5 YR3/2) , indr y condition light reddish brown (5 YR6/3) , heavy veryfine sand y loam. Weak tomoderate ,fine subangula r blocky, tocoars e granular structure. Moist, friable. Notstick y andno tplasti c when wet. Soft when dry. Many pores. Very many roots.Transitio n gradual. A3(A2) 15-50cm: Reddish brown (5 YR5/3-4 ) heavy veryfine sand y loam. Weak tomoderate , me dium tofine subangula r blocky structure. Moist, friable. Not plastic and notstick y when wet. Slightly hard when dry. Many pores. Many roots. From 20t o4 5c mdept h common (20%), very small (0.5c m diam.), ratherfine grained , red plinthite concretions. Transition clear. Bj 50-100 cm: Red (2.5Y R5/8 ) heavy veryfine sand y clay loam. Structure moderate tostrong , coarse angular and subangular blocky, composing weak, coarse prismatic. Common, faint clay skins. Moist, firm. Sticky and plastic when wet. Hard, tover y hard when dry. Fewpores . No roots.Transitio n gradual. B, 100-160 cm: Reddish yellow (7.5 YR6/8 ) very fine sandy clay loam, with many, finet ome dium sized (in lower part medium sized tocoarse) ,distinc t mottleso f white(2. 5 Y8/2 ) andre d (2.5 YR 5/8).Structur e moderate tostrong ,coars e angular blocky,composin g moderate,ver y coarse prismatic. Afew , faint clay skins. Moist, firm. Slightly sticky and slightly plastic when wet. Hard, tover y hard when dry. No pores.N oroots . Transition clear toabrupt , wavy. ClB 160-230 cm:Weatherin g shale: white (5 Y 8/1) heavy clay with many, coarse, prominent mottles, inhorizonta l stripes,o fdar k red(1 0 R 3/6) andyello w (10 YR7/8) . Moderate, me diumsize d tocoarse ,plat ystructur eelements ,fallin g apart intoangula r blocky. Moist, friable. Upper part rather hard when dry, lower part soft when dry.Transitio n abrupt. IIC20 230-270cm: White( N 8/0) loam,wit h many,ver ycoarse ,prominen t mottles,i nvertica l pipes, of red(2. 5Y R4/8 )an d yellow(1 0Y R 7/8). Massive.Sof t when dry.

(VI) LlTHOSOLS(L) General concept: Lithosols areshallow , often stony soils without horizon develop ment, over consolidated rock. Other characteristics and the properties vary consider ablydependin go nth enatur eo fth erock .The y aretherefor e commonly subdivided in tophase saccordin gt oth egeolog yo fth e substratum. ASand-stone substratu mphas e(L ,ss) ,a Chertysubstratu m phase(L , CH),a Quartzite substratum phase (L,QU )an da Kaolinite-stone substratum phase (L, K or L)wer e encountered. Of the latter, oneprofil e will be given, which is actually less shallow than normal.

148 Profile 42. KAOLINITIC LITHOSOL(L) ;o r Lithosol, Kaolinite-stone substratum phase Field description 241 (Sombroek) Location: Along highway BR-I4, km66. 0 So f Sao Miguel do Guama. Reliefand drainage: Edge of terrace about 30m abov e nearby rivulet. Surface drainage good, internal drainageslightl y imperfect. Altitude 60m . Parent material: Consolidated kaolinitic sedimentso f Mioceneag e(?) . Vegetativecover: Recentlycleare d from primeval forest.

Kx 0-15 cm: Grey brown (2.5 Y 5/2) sandy clay loam. Weak, fine subangular blocky structure. Moist, very friable. Slightly sticky and slightly plastic when wet. Soft when dry. About 75% rather soft, somewhat rounded stones (diam. < 10 cm), white (10 YR 8/1), with very fine veinso fyello w(1 0Y R 7/8). Many pores.Ver yman y roots.Transitio n gradual. A, 15-60 cm: Pale olive (5 Y 6/3) sandy clay. Weak to moderate, fine subangular blocky and weak, veryfine granula r structure. Moist, friable. Slightly sticky and slightly plastic when wet. Slightly hard when dry. 5% stones similar to those in A,, but not rounded, and somewhat larger (diam. 20 cm ca.). Manypores . Many roots.Transitio n gradual. R,-B, 60-110cm: Paleyello w (2.5 Y7/4 )clay , with common,fine, distinc t mottles of reddish yellow (5 YR 6/8). Moderate, fine subangular blocky structure. Moist, friable. Sticky and plastic when wet. Slightly hard when dry. About 90% of the horizon consists of hard, sharp edged stones (diam. < 20 cm), white (5 YR 8/1) and with thin veins of pale red (10 R 6/4) in their centres,reddis h yellow( 5Y R 6/8) and yellow(1 0Y R 6/8)a t the break-lines.I n theeart h some roots. Porous.Transitio n gradual. Rt J10-200 cm: Horizontally layered, broken, hard stone, with colours identical to that of the stones of Rj-B,. Some krotovinas of very pale brown (10 YR 7/4) earth, with structure and texture of the earth of R^B,, and with much insect activity. Transition gradual. R, 200-270+ cm: White (5 YR 8/1) clay-stone, with many, medium sized, distinct mottles of pinkish white (7.5 YR 8/2), yellow (10 YR 8/8) and red (2.5 YR 5/8). Massive. Rather soft when dry. Nopores .N o roots.

(VII) GROUNDWATE R LATERITESOIL S (GL) General concept (cf. also II.3.2.1). Ground Water Laterite soils are intermittently imperfectly drained, highly weathered soils. They have a light coloured and usually light textured A2 subhorizon. The B horizon consists of largely soft plinthite; it is made upo f dense,mor eo r lessclaye ymaterial ,wit h prominent,coars ean d abundant mottles of red, and often also some yellow, in a white or light grey matrix. The base saturation islo wan dth esilicat ecla ymineral sconsis to f kaolinite.Th egenesi san d the variation incharacteristic so f the soilsar ediscusse d indetai li nII.3.2.1 ,whic hsectio n alsoinclude sman yshor t profile descriptions.Onl yon eful l profile descriptionfollows .

Profile 43. GROUND WATER LATERITESOI L(GL ) Field description 162/173 (Day,Sombroek ) Location:South-eastern part Marajo-island; About 25k mW o f Soure(Lat.0°.48'S; Long48°.40 'W) . Reliefand drainage: Slightly dipped part of extensive,flat, lo w upland, about 2-3 m above high water level. Imperfectly drained: ground water level at 2,5 mdept h during the dry season, at 1m dept h dur ingth erain y season. Parentmaterial: Late Pleistocene,o r Early Holocene,fluvio-marine (? )sediments . Vegetative cover:Grasses ; scattered low trees and shrubs. Vegetation is burned in most dry seasons. Formerly probably forest covered. Theprofil e islocate d near a patch with traces of former Indian occupation (Terra Preta). Ai 0-30 cm: Black (10 YR 2/1) sandy loam, with white points of bleached quartz grains. Single grains. Moist, loose. Not sticky and not plasticwhe n wet. Soft when dry. Many pores. Many roots.Transitio n gradual.

149 A, 30-90 cm: White (10 YR 8/2) light sandy clay loam. Very little coherent porous massive Moist, very friable. Not sticky and slightly plastic when wet. Slightly hard when dry. Many pores.Commo n roots.Transitio ngradual . Bif 90-120 cm: Yellow (10 YR 6/8) light sandy clay loam, with many, medium sized to coarse, distinct to prominent mottleso f red(2. 5 YR4/8 ) andwhit e(1 0 YR 8/1). Within the reda few small plinthiteconcretions . Structurewea k tomoderate , mediumsize d prismatic, fallingapar t into weak, mediumsize d subangular to angular blocks. Common, faint clay skins. B^ 120-260cm: White (N 8/0) sandy clay loam, with many medium sized to coarse, prominent mottles of red (10 R 4/6) and some brownish yellow (10 YR 6/8), in a reticulate-prismatic pattern. Centres of the red half hardened. Structure moderate, medium sized angular blocky, composingweak ,mediu msize dprismatic .Common ,fain t todistinc tcla yskins .Moist ,friable . Slightly sticky and slightly plastic when wet. Hard whendry . Very fewpores . No roots.Tran sition gradual. With Dutchauger: IIBjf 260-300 cm: White (N 8/0) light sandy loam, with many, coarse, prominent mottles of pale yellow (2.5 Y 8/4) and light red (2.5 YR 6/6). Scattered, large, loose plinthite concretions, (fossil),especiall y inth elowe rpart .

(VII) HYDROMORPHIC GREY PODZOLIC SOILS (HP) Generalconcept: Thisname, whic h isno t encountered inth e literature,i stentativel y given to a locally frequently encountered group of imperfectly drained, moderately weathered soilswit hth efollowin gcharacteristics : The A horizon is grey or grey brown and rather light textured. The Bhorizo n isa textural-B(cf. II.2.1) ,wit hmottle so fyellowis h hue ina greyishmatrix .Thi s B horizon is dense,heav y textured, has a well developed prismatic to columnar structure, and prominent signso fcla yilluviatio n(cla yskins) .Th esilicat ecla y mineralsar enon-kao - liniticfo r a good part.Th e base saturation ismediu m inth e upper part ofth e profile, high to complete in the lower part. The pH of the lowest section of the profile (the C horizon)i softe n relativelyhigh , namely6-8 .Th esurface s ofth estructur eelement s inth eB horizonar eofte n comparativelydar k coloured.Slickenside sar edistinc ti nth e lowerpar t ofth e B horizon.Th esoil sdiffe r from the Grey Hydromorphic soilsi nth e high base saturation and the composition of the clay fraction. The soilshave ,i n fact, several characteristics of Grumosols, of which they may constitute a kind of imper fectly drained phase. Morphometrically the soils are also similar to the Solodized Solonetz. Becauseo fth ebas esaturatio n status,th esoil sar efurthe r classified as Hydromorphic Grey Podzolicsoil ,wit hhigh basesaturation (HP/j/>). Three subtypes are distinguished, namely one that is (arbitrarily) called the Ortho {\\?hb, o), one called theShallow phase (H?hb, s),an d one the Darkphase (HP/,/, , D).

The Ortho sub-typeha sa nA horizon of4 0 to8 0c mthick ,whic hconsist so fa nA , and A„ and some timesa n A,. The Bhorizo n is8 0t o 150c m thick andconsist s normally of a B,an da B,. The C hori zon isabou t 50c m thick. The A, subhorizon, 10t o 30c m thick,consist spredominantl yo f light loam, orloam .It scolou rvalu ei s2 o r3 ,it schrom a6 o r7 ,it shu e 10YR ; light brownish grey (10 YR 6/2) is most common. A few, fine, faint mottles, usually of yellowish brown (10 YR 5/8), are often present. Thestructur ei sweak ,fine subangula r blocky. The transition to theA , isgradual . TheA , subhorizon, 20t o4 0c mthick ,i spredominantl y aloa mo rheav yloam .Th emai ncolou ri sligh tgre yo rligh tbrown ishgre y(1 0Y R 6/1 resp.6/2) ,wit h few tocommon ,fine, fain t mottleso fyellowis h browno rbrownis h

150 yellow (10 YR 5/8). The structure is massive or weak,fine t o mediumsubangula r blocky. The transi tion to the B, isclea ro rabrupt , andwavy .Th e B,subhorizon , 30t o 80c mthick , haspredominantly , clay loamtextures .Th emai ncolou r isligh t grey (10 YR 7/1, resp. 6/1). Mottles with value 5, 6 or7 , chroma 8an d hue 10 YR or 7.5 YR (predominantly reddish yellow: 7.5 YR 6/8) arecommon , mediumsize dan d distinct.Als omottle so fyellowis hre d(5Y R 5/8)orre d(2. 5 YR 5/8)ar e often present, but less conspicuous. The structure is usually moderate, medium prismatic, orcolumnar . Clay skins arecommo n and distinct. The transition to the B, is gradual, wavy. The B, subhorizon, 40 to 80c m thick, consists of clay loam orclay . The principal colour is light grey (10 YR 6/1, 10Y R 7/1, N 7/0), with many, medium, distinct to prominent mottles of yellowish hue, as in the B,. Mottles of red(2. 5 YR 4/8) are few and fine, if present. The structure is a weak to moderate prismatic. Clay skins are common, and distinct to faint. Slickensidesar e general on thesubhorizonta l surfaces. The Chorizo n consists mostly ofsill ycla y loam. Itsmai ncolou ri spredominantl y light olive grey( 5 Y6/2) . Mottles, usually of brownishyello w (10 YR 6/8), arepresen t invaryin gquantity , sizean ddistinctness . Colour lesscrystals ,o fgypsum ,ar eofte n found.

TheDark phase ha sa nA horizo no fonl y2 0t o4 0cm .Th e B horizon is3 0t o 80 cm thick and is rela tivelydar kcoloured .Th eC horizo n isabou t4 0c mthick . The A, subhorizon, about 10c m thick, is loamy and grey or dark grey (10 YR 5/1, resp. 4/1). The structure iswea k to moderate,fin e subangular blocky.Th etransitio n to theA , isgradua lan dsmooth . The A, subhorizon, 10-30c m thick, consists of a loam or light clay loam. Itscolou r is normallygre y (10Y R 5/1 or6/1) ,ofte n withsom eyellowis h mottling.Th estructur e isofte n weak,mediu msubangu larblocky .Th etransitio n toth eB i scommonl y gradualan d irregular.Th e B, subhorizon, 30t o 40 cm thick,consist s of a heavyclay .Th e maincolou r isdar k grey(1 0 YR 4/1), with many,fin e to medium, distinct mottles,usuall y of brownish yellow (10 YR 6/6). Thestructur e isstron gprismatic . Clay skins arecommo n tocontinuous ,an ddistinct .Slickenside sar efaint .Th eB ,subhorizon , 10t o4 0c mthic k if existent, iso fcla y orheav ycla y texture. Itsmai ncolou r isgre y(1 0Y R 5/1), with mottleso f yellowish hue.Th e structure ismoderat eprismatic . Clay skinsar ecommo n and faint; slickensides distinct. The C horizonconsist so fa clay loam,wit hcolour snormall ysimila rt othos eo f the B,.Colourles scrystals , ofgypsum ,ar egeneral .

The Shallowphase hasa nA horizon of only 20 to4 0cm ,a Bhorizon , being usually only a B„ of 20 to4 0c man da C horizo no f 10t o 30 cm. Itstota l solumcomprise s 50t o7 0 cm; below it, hard rock is found. The A, subhorizon, 5t o 15c m thick, is of sandy loam texture. Itscolou r isdar k grey brown orver y dark grey brown (10 YR 4/2 resp. 3/2), its structure weak to moderate,fine subangula r blocky. The transition is gradual, smooth. The A, subhorizon, 20 to 30c m thick, has the same texture as the A,. Itsmai ncolou ri spredominantl y light brownish grey(1 0 YR 5/2). Mottleso fstron gbrow no rreddis h yellow (7.5 YR 56resp .6/8 ) aremany ,fine an d faint. Thestructur ei smostl y weak, mediumsubangu larblocky .Th etransitio n is cleart oabrupt ,irregular .Th eB ,subhorizo n iso fcla ytexture .Th e colour of the matrixi sgre y(1 0Y R 5/1 or6/1) ,wit hmany fine,, distinc t mottleso freddis h yellowo ryellowis h red. The structure isstrong , coarse prismatic, partly columnar. Clay skins are common and distinct. Slickensides are invariably present, and distinct. The C horizon is silty, with a variety of colours, in which olivegre ypredominates . Crystalso f gypsum wereno t found.

Profile 44. HYDROMORPHICGRE Y PODZOLICsoil , withhig h basesaturation ,Orth o(HP/,/, ,o ) Fielddescriptio n 288(Sombroek ,Sampaio ) Location: Araguaia Mahogany area, Bloco Piranha, southern half, transect 7, stake 38(Lat . 6 01'S ; Long.4 8 10'W). Reliefand drainage: Fla t terrain, less than I m above levelo f nearby,narrow , intermittent rivulets. Very distinct micro relief of canaleies (kauwfoetoes).Imperfectl y drained: the canaletesar efilled with non-stagnant rain water during rainy season. Profile studied in dry season, when ground water kvel wasbelo w 150 cm. Alt.20 0m fa . arenr material: Gypsiferous andcalcareou s siltyclay-stone s of Motuca membero f Pastos Bonsbed s (Jura-Triassic). tgetative cover:Primeva l semi-deciduous forest. Nearly closed, rather high canopy, consisting pre- ominantlyo f ratherbi gmahogan y trees;als osom epalms . Ratherope n undergrowth,wit ha few thin keepersan dclimbers .Ope nfield laye ro fseedlings .

151 (>! 2-0 cm: Undecomposed plant residues. 08 2-0 cm: Partly decomposed plant residues, with fine roots. At the surface many outcasts of worms. Aj 0-JOcm: Greybrow n (10Y R 5/2)ligh tloam .Wea k tomoderate fine, subangula rblock ystruc ture. Moist, very friable. Not sticky and not plastic when wet. Rather hard when dry. Many roots. Many,larg epores .Transitio n gradual, and smooth. Aa 10-50cm: Dark grey brown (10 YR 4/2) light loam, with many,fine, fain t mottles of yellow ish red (5 YR 5/8). Porous massivet o weak,mediu m to coarse, subangular to angular blocky structure. Moist, friable. Not sticky and not plastic when wet. Very hard when dry. Many roots.Commo n largean d fine roots.Transitio n abrupt,wavy . Bit 50-80 cm: Light grey( N 7/0)cla y loam, with many, medium sized, distinct mottles of reddish yellow (7.5 YR 6/8) and some red (10 R 4/8). Structure weak, very fine prismatic, composing moderate, very coarse columnar. Common, distinct clay skins, somewhat darker coloured than rest of soil mass. Moist,ver yfirm. Ver y sticky and very plastic when wet. Few roots. No pores.Transitio n gradual,wavy . B3B 80-120 cm: Light grey (N 7/0) clay loam, with common to many, medium to coarse, distinct mottles of reddish yellow (7.5 YR 6/8). Weak to moderate, medium to fine prismatic struc ture. Common, distinct clay skins, somewhat darker coloured than rest of soil mass. Distinct slickensides on the (sub) horizontal surfaces. Moist, very firm. Very sticky and very plastic whenwet . No roots.N opores .Transitio n clear, irregular. Cg 120-150+ cm: Olivegre y( 5 YR 5-6/2)silt ycla y loam,wit h many,mediu m tocoarse , distinct mottles of brownish yellow (10 YR 6/0) and some red (2.5 YR 5/6). Massive. Moist, very firm.Sligh t sticky and slightly plastic when wet. Locally white (10 YR 8/1) spots, effervescing with HC1: carbonates. Many, mainly very small (1 mm),colourles s crystals: gypsum.

Fullmineralogica l analysiswa scarrie d out onth eA 2horizo n ofa comparabl e profile (cf. sample 290-2o f theTable s 7an d 8,an d the Figs. 14a-e).It s clay fraction contains 15% mica(illite) ,2 5 % 'intermediate', 30 % kaolinitean d 30 % clay-sized quartz. Inth e fraction 2-16 micron, mica (illite) comprises 3% and chlorite 5%; the rest is quartz. Thecoarse rfraction s consist ofquart z (95 %)only ,wit hsom ehematite .

Profile 45. HVDROMORPHIC GREY PODZOLIC soil, with high base saturation, Dark phase (HP/,*,, D) Field description 291(Sombroek , Sampaio) Location:Araguai a Mahogany area, Bloco Piranha, southern part, transect 7, stake 25 (Lat. 6 01'S ; Long.48°10'W). Relief anddrainage: Fla t terrain, less than 1m above level of nearby, narrow, intermittent rivulets. Vague micro relief of canaletes(kauwfoeioes). Imperfectly drained: the canaletesar efilled wit h non- stagnant rain water during rainy season. Profile studied in dry season, when ground water level was below 150 cm. Parentmaterial: Gypsiferou s and calcareous clay-stones of Motuca member of Pastos Bons beds (Jura-Triassic). Vegetative cover: Primeval deciduous forest. Open canopy, consisting mainly of low, thin, stunted mahogany trees,i n a dense pattern. Rather open undergrowth,largel y consisting of thin creepers and climbers. Rather openfield layer ,o f some seedlings,clump s of Cyperaceae, and Selaginellae. Oi 3-1 cm: Undecomposed plant residues. O, 1-0 cm: Partly decomposed plant residues, with fine roots. At the surface a few outcasts of worms. A, 0-7cm: Grey (10Y R 5/1),ligh t grey(1 0Y R 6/1)whe n dry,loam .Wea k tomoderate , medium tofine, subangula r blocky structure. Dry, slightly hard. Not sticky and not plastic when wet. Friable when moist. Many roots. Many large, and few fine pores. Transition gradual and smooth. A, 7-30cm: Light brownish grey(1 0Y R 6/2), light grey(1 0Y R 7/1)whe n dry, loam,wit h many, fine, faint mottles of yellowish brown (10 YR 5/6). Weak to moderate, medium to coarse, subangular to angular blocky structure. Dry, hard. Slightly sticky and slightly plastic when

152 wet. Friable to firm when moist. Common roots. Few large, and common fine pores. Transi tion gradual, irregular. By, 30-75 cm: Dark grey(1 0Y R 4-5/1) heavy clay,wit hcommon , medium sized, distinct mottles of reddish brown (2.5 YR 4-5/4). Structure moderate to strong, medium to coarse prismatic, with tendency to columnar. Many, distinct clay skins. Faint slickenslides on the (sub) hori zontal surfaces. Moist,firm t over yfirm. Plasti can d verystick ywhe n wet.Ver yhar d when dry. Fewroots .A few , largepores .Transitio n gradual,wavy . BJO 75-120cm: Grey (10Y R 5/0) clay,wit h common, medium sized, distinct mottles of yellowish brown (10 YR 5/6) and red (2.5 YR 6/6). Moderate, coarse prismatic structure. Common, faint clay skins. Distinct slickensides on the (sub)horizontal surfaces. Moist, firm, to very firm.Ver ystick yan d veryplasti cwhe n wet. Cg 120-140+ cm: Light grey (N 6/0) clay loam, with many, coarse faint mottles of brownish yellow (10 YR 6/6). Massive, Moist, firm. Very sticky and very plastic when wet. Scattered, fine, white( N 8/0)points . Veryman ycolourles scrystal s(gypsum) .

Profile 46. HYDROMORPHIC GREY PODZOLIC soil, with high base saturation, Shallow phase (HP^j, s). Field description 294 (Sombroek, Sampaio) Loan/on:Araguai a Mahoganyarea ,Ri oAntonino ,transec t la,stak e190(Lat.6'09'S ;Long .48°14'W) . Reliefand drainage: Flat terrain, about 1m abov e level of nearby, narrow, intermittent rivulet. Cons picuous micro relief of canaletes(kauwfoetoes). Imperfectly drained: thecanaleies ar efilled wit h non- stagnant rain water during the rainyseason . Profile studied indr yseason ,whe n ground water levelwa s below 100 cm. Parentmaterial: Calcareou s cherty silt-stones of Pedra de Fogo beds(Permean) . Vegetativecover: Deciduous hydromorphic shrub. Aver y few, thin,stunted , lowtree semerg eabov e a denselaye r ofshrub san d thin creepersan d climbers. Rather densefield laye r(Cramineae, Cyperaceae, Selagine/lae, Araceae). O, 5-0 cm: Undecomposed plant residues. O, Absent. Surface very irregular, due toabundanc e of outcastso f worms. A, 0-8 cm: Dark grey brown (10 YR 4/2) sandy loam, with few, fine, faint mottles of yellowish brown (10 YR 5/4). Moderate, fine subangular blocky structure. Dry, hard, slightly crusted. Slightly sticky and slightly plasticwhe n wet. Friable when moist. Many roots. Common large andfine pores .Transitio n gradual, smooth. A„ 8-25 cm: Grey brown (10Y R 5/2) sandy loam, with many,fine, fain t mottleso f strong brown (7.5 YR 5/6). Weak to moderate, medium to fine, subangular blocky structure. Dry, slightly hard. Not plastic and slightly sticky when wet. Friable when moist. Common roots. Many largean dfine pores .Transitio n gradual,wavy . A„ 25-37 cm: Grey (10 YR 6/1) light loam, with many, fine, faint to distinct mottles of grey brown (7.5 YR 4/4).Weak , coarse, subangular to angular blocky structure. Dry, hard. Not plastic and slightly sticky when wet. Friable when moist. Few roots. Few large,an d common fine pores.Abou t 25%ver y small (diam. < 5mm) ,fin e grained plinthite concretions. Transi tionclear , irregular. B»» 37-57 cm: Light grey (10 YR 6/1) clay, with many,fine, distinc t mottles of reddish brown (5 YR 4/4). Structure strong,ver y coarse prismatic, with tendency to columnar. Common to general, distinct clay skins.Distinc t slickensides on the (sub)horizontal surfaces. Dry, hard. Very stickyan d veryplasti cwhe n wet.Ver yfirm whe n moist. Very few roots.N o pores.Tran sition gradual,wavy . Bu» 57-77 cm: Grey (5 Y 5/1)clay ,wit h many,fine, fain t mottles ofyellowis h brown (10Y R 5/6). Strong, coarse prismatic structure. Common, faint clay skins. Strong slickensides on all (sub)horizontal surfaces. Dry, hard. Very sticky and very plastic when wet. Very firm when moist. Very few roots. No pores.Transitio n abrupt, irregular. Ciy 77-87 cm: Brownish yellow (10 YR 6/8) silt loam, with many, distinct horizontal stripes of white (10Y R 8/1)an d grey (10 YR 5/1). Moderate, medium platy structure. Moist, friable to firm.Sticky ,ver yplasti c when wet.A few,fine roots . No pores.Transitio n abrupt, irregular. Ctg 87-102 cm: Bluish olivegre y and purplish olivesilt yclay . Moderate, medium platy structure. Moist, very firm. Very sticky and very plastic when wet. With white spots, which give effer vescencewit hHC1 . R 102+ cm: Paleyello wcalcareou s silt-stone.

153 (IX) GROUND WATER PODZOLS AND WHITE SAND REGOSOLS(G Pan d WSR) General concept: Ground Water Podzols are imperfectly drained, highly weathered soils.The y are characterised bya ligh t to very light texture throughout the profile,a bleached A2 horizon of varyingthickness ,an da dar k coloured B horizon with acon centration of humic material. This Bhorizo n isusually , fora part, cementedb y sesquioxides,thu sformin ga hardpa n(Ortstein) . The thickness of the A2 horizon inth e Amazon Ground Water Podzols varies from 20c mt o20 0c man d more.Th ethicknes san d hardnesso fth eOrtstei nals ovar yconsi derably. There are indications that the cementing sesquioxides are aluminumoxides , not iron oxides.Thi sfurthe r classifies the soila s Ground Water Humus Podzol.

Under WhiteSand Regosol are classified thosesoil stha tconsis to fbleache dsan dt oa great depth. The soil occurs commonly ina well-drained position.

Profile 47. GROUND WATER HUMUS PODZOL(GP ) Field description 234(Sombroek ,Sampaio ) Location:Alon g highway BR-14,k m 38.8S o f Sao Miguel do Guama. Reliefand drainage: Approximately flat, narrow terrace, 3-4 mabov e levelo f nearby rivulet. Alt.2 5m ca.Imperfectl y drained. Parent material: Pleistocenefluviatile sediments . Vegetativecover: Secondary shrub.

At 0-10cm: Dark brown (7.5 YR 3/2) sand, spotty by presence of very many white sand grains. Single grains. Moist, loose. Not sticky and not plastic when wet. Soft when dry. Many roots. Manypores .Transitio n clearan d spotty. A, 10-45cm: Light grey(1 0Y R 6/1) sand. Singlegrains . Moist, loose.No t stickyan d not plastic when wet.Sof t when dry. Manypores . Many roots.Transitio n gradual and smooth. B,» 45-60 cm: Dark grey (10 YR 4/1) sand. Weak, medium to fine subangular blocky structure. Moist,friable . Not stickyan d not plasticwhe nwet .Slightl yhar d whendry . Manypores .Man y roots.Transitio n abrupt and irregular. Krotovinas of Bj*penetratin g into underlying horizon. B21Am 60-80 cm: Ortstein: Brown todar k brown (7.5 YR 4/2) loamy sand. Massive, indurated. Moist, extremely firm. Extremely hard when dry. Not sticky and not plastic when wet. No pores.N o roots.Transitio n abrupt and broken. B22/im 80-110 cm: Ortstein: Brown to dark brown (7.5 YR 4/4) loamy sand. Massive, strongly ce mented. Moist,ver yfirm. Ver y hard when dry.No t stickyan d not plasticwhe n wet.N o pores. No roots.Transitio n gradual and irregular. B3 110-150 cm: Light olive brown (2.5 Y 5/4) sand. Single grains to weak, medium tofine sub - angular blocky structure. Moist, very friable. Soft when dry. Not sticky and not plastic when wet.N o pores.N o roots.

Profile 48. WHITE SAND REGOSOL(WSR ) Field description 131 (Day, Sombroek) Location:Lowe rAmazo n region,abou t 10k mN ofOroximin a (Lat. lc42'S;Long. 55°49'W) Relief anddrainage: Extensivelyflat, t o gently undulating terrain. Atl. 50m ca. Excessivel y drained. Parentmaterial: Pleistocenefluviatile (? )sediments . Vegetativecover: Densean d lowforest ; burned inprecedin g dry season. Oj 3-0 cm: Reddish brown,partiall y decomposed organic material,wit h very many fine roots A, 0-20cm: Gre y (N 6/0) sand.Singl egrains . Manyroots .Transitio n gradual. A,? 20-480+ cm: White (N 8/0) sand. Singlegrains .I n upper parta fe wroots ,i nlowe rpar t none. Sand grainsar eangula r tosubangular . Littlevariatio n ingrai n sizedistributio n with depth.

154 (X) Low HUMIC GLEY AND HUMIC GLEY SOILS(LH G and HG) General concept: These are poorly drained soils from recent non-marine sediments, with littleprofil e development.The yhav ea nA xhorizo n ofvaryin gthicknes san dpro minence, overlying a mineral, gleyed sub-surface and subsoil. The Ax horizon of the Low Humic Gley soils isthi n and/or little humic.Th e At horizon of the Humic Gley soils is prominent and dark, without being, however, an organic horizon proper. A prominent At in this respect is arbitrarily defined as being 20 or more cm thick and havinga percentag e ofCarbo nwhic hsurpasse s 1.5 4-0.01 5 x %clay .

The gleyed sub-surface and subsoilhorizon s of the profiles ofbot h Low HumicGle y and Humic Gley soils encountered in Amazonia have a main colour of grey or light grey. Mottling is usually common,fine t o medium sized and distinct, and its colours are predominantly yellowishbrow n or strong brown (10Y R 5/6-5/8an d 7.5Y R 5/6- 5/8 respectively).Hue s redder than 5Y R arerare .Th e soilsar e usually heavy textured and often contain considerable percentages of silt.Th e structural development of the subsoili smostl y weak.Sign so fcla yilluviatio n(cla yskins )ar eofte n littleconspicuous , if present at all. Because of stratification of the original sediments, sudden changesi n texturewithi n aprofil e may occur.Th e presenceo f one or more humico r peaty layers in the subsoil isno t uncommon.Th e soilsvar yconsiderabl y inconsistenc e and degree of subsoilcompactness .Th e mineralogical composition of the soilsvarie s from site to site, and there are also considerable differences in the base saturation percentages. Amongst the Low Humic Gley soils,a Carbonatesubsoil phase (LHG, c) is separated. This phase ischaracterise d by being strongly alkaline in the subsoil, which also con tains carbonate concretions. A very compact and usually dark coloured horizon (laklaag, 'lacquer') isinvariabl y found just above this alkaline subsoil. Amongst the Humic Gley soils, an Upland phase (HG, u) is distinguished. It is characterised by being extremely acid, having low amounts of silt, and silicate clay minerals of 1:1lattic e structure (kaolinite) only. Profiles, intergrading toGroun d Water Lateritesoil ,ar e not uncommon, concerning bothth e LowHumi cGle yan d theHumi cGle ysoil s(cf. Profile 18). A few profile descriptions are given. More details about the soils, many profile des criptions,an danalytica ldat a arereporte db y SOMBROEK(1962b) .

Profile 49. Low HUMIC GLEY soil (LHG) Field description 265(Sombroek , Sampaio) Location: Lower Tocantins floodplain, about 20k m WNW of Igarape-mirim (Lat. 1°.58' S; Long. 49°.09'W) . Relief anddrainage: Flat lowland; floodplain of fresh water tidal creeks. Poorly drained: daily flooded by water with a considerable load of sediments. The difference between high tide and low tide level of nearby creek isabou t 3 m. Parent material: Recent, fresh water deltaic sediments on older deposits. Vegetativecover: Swamp forest with many palms. A, 0-7 cm: Very dark grey (10 YR 3/1) heavy clay loam. Moderate, medium to fine subangular blocky, tocrumbl y structure. Moist, very friable. Plastic and very sticky when wet, hard when dry. Veryman y roots. Transition clear.

155 WithDutch auger: Cig 7-I8cm: Lightgrey(10 YR 6/1-7/1)silt ycla yloam ,wit ha few,fine , distinct mottless of strong brown (7.5Y R 5/8).Wet ,ver ystick yan d veryplastic . Common roots.Transitio n gradual. Cig 18-50 cm: Light grey to grey (N 6/0) heavy clay, with many, fine to medium sized, distinct mottles of strong brown (7.5 YR 5/8). Wet, very sticky and very plastic. Common roots. Transition clear. C,p 50-110cm: Dark reddish brown (5 YR 2/2)peat y sandy clay loam.Transitio n clear. Clg 110-130cm: Grey (N 5/0) silt loam. Wet, very sticky and very plastic.N o roots. Transition gradual. Cw 130-240 cm: Light grey (N 7/0) silty clay, with few, coarse, distinct mottles of light olive brown (2.5 Y 5/6). Wet, very sticky and very plastic. No roots. Transition gradual. \\C,g 240-380 cm: Light grey (N 7/0) clay, with many, medium to coarse, prominent mottles of dusky red (10 R 3/4). Mottles slightly hardened in their centres. Wet, very sticky and very plastic.Compact . No roots.

Profile 50. HUMIC GLEV soil (HG) Field description 188(Sombroek ) Location: Lower Amazon floodplain, about 20 km So f Prainha (Lat. 1°59'S ; Long. 53°30'W) . Reliefand drainage: Flat lowland; backswamp offloodplain o frive r Amazon and rivulet Purus. Poor lydrained : during high water seasonflooded wit habou t 1.5 mwate r which hasa load of sediments. Parentmaterial: Recent fluviatile sediments. Vegetativecover: Grasses,an d other herbaceousplants . Ai 0-25 cm: Grey (10 YR 5/1) silty clay loam, with some strong brown (7.5 YR 5/8) along the grass roots. Structure medium sized subangular blocky, of moderate strength. Surface hard when dry, slightly cracked. Moist, friable. Sticky and plastic when wet. Transition clear. WithDutch auger: Cig 25-70 cm: Light grey (N 6/0) silty clay loam, with many, medium sized, distinct mottles of reddish yellow (7.5 R 6/8). Moist, friable to firm. Sticky and plastic when wet. Transition abrupt. C&.b 70-120cm: Very dark grey (10Y R 3/1)clay ,wit hcommon ,fine, fain t mottles of dark yellow ish brown (10 YR 4/4). Moist, firm. Sticky and very plastic when wet. Compact. Transition gradual. Cw 120-170cm: Grey (N 4/0-5/0) clay withmany ,mediu m sized, distinct mottles of reddish yel low(7. 5Y R 6/6). Moist,firm. Stick yan d veryplasti c whenwet .Transitio n abrupt. C4b 170-190 cm: Very dark grey (10 YR 3/1) clay. Moist, firm. Sticky and very plastic when wet. Transition gradual. Cw 190-210cm: Grey (N 5/1) clay with common, fine, distinct mottles of yellowish red (5 YR 5/8). Moist,firm. Stick yan d very plasticwhe n wet.Transitio n clear. C,6 210-220+ cm: Black(1 0Y R 2/1)clay ,wit h organicrelics . Moist,firm. Stick yan d very plastic when wet.

Aful l mineralogicalanalysi swa scarrie d out on a sample of the second horizon.Th e data concerning sample 188-2o f Tables 7an d 8 and Figs. 14a-e show that indeed the mineralogical composition isfavourable , for whichth e Kian d Kr data givealread ya n indication. In the clay fraction, only 22% kaolinite is present, while mica (illite) occupies 30% , 'intermediate' 10% , swellingillit e8 %, and chlorite7 % . Aconsiderabl e amount offelspa r (15% ) occursi nth e2-1 6micro n and 16-80micro nfractions . Photo 11 illustratesth ecompositio n ofth ecla y fraction. Sample 190-2ha s approximately the same mineralogical composition; it iso f a Low HumicGle ysoil ,Carbonat e subsoilphase ,o fth esam ephysiographi cuni t(th ehorizo n sampledoccur sabov eth elacque ran dth ealkalin ehorizon so fth eprofile) .

156 Profile 51. HUMIC GLEYsoil ,Uplan d phase(HG , u) Field description 119 (Day,Sombroek ) Location: Amapa Territory, catchment area of Igarap6 do Lago (Lat. 0°2T N; Long. 51°38'W) . Relief and drainage:Botto m land in undulating upland with savannah cover. Poorly drained. Alt. below 50m . Parent material:Colluvia l deposits from heavy textured sediments of surrounding upland. Vegetativecover: Tallgrasses .

Au 0-40 cm: Black (5Y R 2/1)clay . Strong, veryfine subangula r blocky structure. Moist, friable. Plastic and slightly sticky when wet. Apparently high in organic matter. Many roots. WithDutch auger: Am 40-60 cm: Olive brown (2.5 Y4/2 )clay , withcommon , medium sized, distinct mottles of dark grey( N4/0 )an d dark brown (7.5Y R 4/4). Wet,stick yan d plastic. Cig 60-100 cm: Yellowish brown (10 YR 5/4) clay, with many, medium sized, distinct to promi nent mottles of yellowish red (5 YR 4/6) and black (N 3/0).Wet ,stick y and plastic, to very plastic. Cw 100-120+ cm: Pale brown (10 YR 6/3) clay, with common, medium sized, faint to distinct mottleso fyellowis h red (5 YR 6/6).

(XI) SALINEAN DALKAL I SOILS Generalconcept: TheSalin ean dAlkal isoil sar echaracterise db yeithe rexcessiv e con centrationso fexchangeabl eNa + and Nig"1-1", orsolubl esalts ,o rboth . Theyar eterme dsaline, ifth econductivit yo fth esaturatio nextrac texceed s4 mmhos / cm at2 5°C , correspondingt oabou t0.1 5% saltsi nth edr ysoi l(SOI L SURVEY MANUAL, 1931,pag e 360). Assumingth esolubl esalt st o bepredominantl y NaCl,thi svalu ecor respondst o about2. 5m.e ./10 0g 'solubl esalts' ,a sthi sdatu m isprovide db yIQA . TheSolonetz is asoi lwit ha clodd y B horizonan da prominen tprismati co rcolumna r structure.Th eexchangeabl ecation so f this horizon are Na+ and/or Mg++fo r agoo d + part (Mg+++ Na+ > Ca+++ H ; cf. SOILSURVE Y STAFF, 1960,p .45) . It hasrecentl y been suggested that Mg++ may have a similar detrimental effect on the structure as + Na andthi sha sbee nconfirme d bylaborator ytrials ,fo r instancethos eo f SCHUYLEN- BORGHan d VEENENBOS (1951).Solonetze sar eusuall yfoun d indr yclimates .The yals o exist, however, along ocean coasts, independent of climatic conditions,wher e they havedevelope d on sedimentswit hhig hpercentage so f Mg++an d Na+, duet odeposi tion in marine-deltaic conditions. They are known as 'coastal Solonetzes' and are described for instance by EDELMAN (1950) for Holland, where they have the local namespik orknip clays . The Amazon soils denoted as Solonetz,Coastal phase (Sol , c) have usually only a moderateprismati cstructure .A numbe ro fthe mhav ea thic k humicto pwit ha crum bly structure and slickensides in the subsoil horizons. These soils are, in fact,intergrades to Grumosols. Others,wit h only a thin humic top and moderate to weak pris matic structure,ar e intergrades toLow Humic Gley soils. The soilsma y have a saline subsoil. Incoasta lAmazoni aals other ear esalin esoil stha thav en ostructure ,o ra weak , fine, granular one.Th ewate r movementthroug hth e profile israpid , contrary to the situa tion with the Solonetz, Coastal phase. These soils are classified as Saline soils.Mor e data aboutth eSalin ean dAlkal isoil so fAmazoni aar ereporte db ySOMBROE K(1962b) .

157 Profile 52. SOLONETZ, COASTALPHAS E(Sol , c) Field description 175(Day ,Sombroek ) Location:Easter npar t of Maraj6 island, about 40k m NW of Soure(Lat . 0°32'S ; Long.48°47 'W) . Relief anddrainage: Flat lowland. Poorly drained:durin g the rainy season submerged under about I m rain water. Parentmaterial: Subrecen t marine-deltaicsediments . Vegetativecover: Grasses,an d other herbaceousplants . A! 0-20 cm: Dark grey (10 YR 4/1) with many, fine, distinct mottles of strong brown (7.5 YR 5/8), mainly along the grass roots. Trampled by cattle. Structure very coarse prismatic, of moderate strength;breakin g into coarse angular blocks,generall y of weak, but in lower part, locally, of moderate strength. Dry, very hard; strongly cracked. Very sticky and very plastic when wet. The upper 10c m has many insect channels; the lower part is massive within the structureelements .Transitio n clearan d irregular. B2,0 20-40 cm: Very dark grey brown (10 YR 3/2) clay, with common, very fine, faint mottles of strong brown (7.5 YR 5/8). Structure moderate, coarse prismatic, locally columnar. Faint, locally distinct, clay skins on all vertical ped faces. Distinct slickensides on the sub-horizontal surfaces. Dry,ver y hard.Ver y stickyan d veryplasti cwhe n wet.Scattered , small(2- 5c m diam), round, half-hard, black concretions, effervescing with H2Ot: concretions of manganese. Transition clearan d wavy. B220 40-140 cm: Dark grey (5 Y4/1 ) clay, with many,fine, fain t mottles of light yellowish brown (2.5Y 6/4) .I n theuppe r part thestructur e ismediu m prismatic,o fwea k tomoderat e strength. Common, faint clay skins on vertical ped faces. Distinct slickensides on the sub-horizontal surfaces. Moist; lower part wet. Very sticky and very plastic. Concretions of manganese as above.N o effervescence with HC1.Transitio n gradual. WithDutch auger: Bae 140-210cm: Light grey (5 YR 6/1) clay, with many, coarse, prominent mottles of yellowish red (5 YR 5/8). In the yellowish red, half-hard plinthitic tubes. Wet, very sticky and very plastic.N oeffervescenc e with HC1.Transitio n clear. Cg 210-225+ cm: Grey (N 4/0) clay. Wet, very sticky and very plastic. No effervescence with HC1. Note: Sampling of the Ax at thebeginning of thedry season, of the Bng, the Bi2g andB 3g at theend of thedry season.

Fullmineralogica lanalysi swa scarrie d outo n asampl eo fth e B22ghorizon .Th edat a ofsampl e175- 3 ofTabl e7 an d8 showtha ti nth ecla yfractio n kaolinitecomprise sonl y 25%, while 2:1 lattice silicate clay minerals comprisetogethe r about 50%.O f these, 'swelling illite' is particularly important (24%). Felspar is present in considerable amounts(1 7an d 10% respectively)i nth e2-1 6an d 16-80micro n fractions. Tosom eexten t sample 154-2ha sth esam emineralogica lcompositio n(cf. theTable s 7an d 8an d the Figs. 14a-e),whic hi so fth e B2g horizon of another Solonetz,Coasta l phase profile (intergrade to Low Humic Gley soil).Phot o 12illustrate s thecomposi tiono fth ecla yfractio n ofth elatte rsample .

(XII) TERRAPRET A SOIL (TP) Generalconcept: Terra Preta soil isa well-drained soilcharacterize d byth e presence ofa thick black, or dark grey,topsoi lwhic hcontain spiece so f artefacts. Asdiscusse di nIII.3.4 ,th eTerr a Preta soili sa kin d ofkitchen-midden ,develope da t the dwelling sites of Pre-Columbian Indians. A general discussion of its qualities is given inV.3.1.2 .Here ,onl y oneful l profile description isgiven .

158 Profile 53. TERRA PRETA,ligh t textured(TP< ) Field description 39(Day ) Location:Santarem , at Silviculture plots of Sawmill training centre. Reliefand drainage: Gentle undulating terrain (slope 1-3 %). Well-drained. Alt.les stha n 50m . Parent material: Pleistocene sediments. Vegetativecover: Lowsecondar y forest.

Kp 0-35 cm: Black( N 2/0)loam y sand. Moderate,mediu mcrum b structure. Moist,friable . Many roots. Scattered pieces of old ceramics. Transition gradual, smooth. A, 35-70 cm: Very dark greyish brown (10 YR 3/1) loamy sand. Moderate to weak, medium crumb structure. Moist, friable. Many roots. Transition diffuse, smooth. B21 70-100 cm: Dark brown (10 YR 3/3) heavy loamy sand. Weak, coarse subangular blocky structure. Moist, friable. Slightly more compact than B2,. Transition diffuse, smooth. B2, 100-160+ cm: Dark brown,t o brown (10Y R 4/3)ligh tsand y loam.Weak , coarse subangular blocky structure. Moist, friable.

(XIII) OTHER SOILS A number of other soils were encountered in Amazonia. Since they have not been properly studied and/orar e of limitedoccurrence , theywil lb edescribe donl yi nshort . Onesuc ha soi li sa fertile ,red ,claye ysoil ,locall y knowna sTerra Roxa. Whethe ri ti s predominantly a Terra RoxaLegitima (a Latosol,cf. grou pHa o fTabl e 6),o r a Terra RoxaEstrulurada ( asoi lwit ha textural-B) , is notestablished . Oneprofil e studied1 (Alenquer) proved to bemor elik eth elatte r than like theformer : The A horizon of thisprofil e iso fcla y loam texture and the B horizon of light clay texture, giving a textural ratio B/A of 1.43. The structure of the B horizon is moderate to weak, very fine to medium subangular, with common and weak clay skins. Analysis revealed that the silt fraction (2-50 micron) is relatively high(20-2 5% , while 45 %cla y ispresent) . Ki data decrease from 2.0 in the topsoil to 1.5 in the central part of the Bhorizon . The AIjO,:Fe20, ratio increases inth esam estretc h from 1.6 to 1.9. TheT-valu e decreases from 18m.e./lO Og i n the topsoil to about 3m.e .i n the subsoil. The pH-H20 isbetwee n 6.0 and 6.5 throughout theprofil e down to 200c mdepth , and the basesaturatio n isbetwee n 70an d 90%.

Another fertile soilencountere d is a black, very sticky and very plastic clayey soil, denoted Grumosol. Oneprofil e studied1 (Monte Alegre)consist s to 40c m depth (A horizon) of black clay loam, of weak or moderate, medium to coarse,mainl y subangular blocky structure. Below this, olive grey clay or heavy clay isencountered , with white specks of carbonates and presence of slickensides. Analysis of this profile gave Ki data of 4.5 to 5.0, high amounts of natural clay, high T-value (40 m.e./100 g and more) and high to complete base saturation, with Ca++ as strongly predominant cation.

Awell-drained ,moderatel yweathere d soilwit hth efollowin gcharacteristic sha sbee n tentatively classified as a Non Calcic Brown-likesoil, Gravelly phase (NB, G): Reddish browncolou rthroughou tth e profile; presenceo fmuc hgrave lan d stoneso f laterized chert; loam to clay loam texture, without much textural difference between theA an d the B horizons.Th estructur e inbot h horizonsi smoderate ,mediu mt o fine subangularblocky . An analysed example of suchprofile s(Araguai a Mahoganyarea )show sa lo wbas e saturation in both theA an d theB horizons ,togethe r 50 cm thick, but high basesaturatio n inth eC horizo n (50c m thick) in which the Ca++ion is very predominant. In all three horizons the Ki value isabou t 2.9 and the percentage Carbon 1-2%. ') Excursion Eighth Brazilian Congress of Soil Science, 1961.

159 Awell-drained ,moderatel y weathered, shallow soilwit hth efollowin g characteristics hasbee ntentativel yclassifie d asa nAcid Brown Forest-like soil, Gravelly phase (AF , G): Presence of many fine gravelly plinthite concretions (of laterized silt-stone) in the lower part of the profile; yellowish brown to light yellowish brown colour below the surface layer;silt ycla yloa m texture. The Ahorizo n does not seem to be much lighter textured than the B horizon, and the structure is moderate subangular blocky and weaksubangula r blockyrespectively . Oneanalyze dprofilefAraguai a Mahoganyarea )show sabou t 35% cla y inth eA horizo n (40c mthick) , 45% clay in the Bhorizo n (40c mthick) , high percentages ofsil t (40-50%),an d much natural clay in the Bhorizon . Ki and Kr data of the solum decrease with depth from 2.4 to 2.2 and from 1.7 to 1.5 respectively. The cation exchange capacity (T-value) of the solum decreases with depth from 12m.e, / 100g t o 8m.e. ,an d thebas esaturatio n from 40t o2 0% ,a tpH-H 20 valueso f 5.5-5.0.

Under Alluvialsoil are classified wello r moderately well-drained soilso n recent sedi ments,wit hlittl eprofil e development. Theprofile s ofthi ssoi lencountere d (Araguaia Mahogany area)ar egenerall y light or mediumtextured , usually with colourso fyellowis h hue,an d withflakes o f micai nth e sand fraction. In one analysed profile, the distinguishable soil layers up to 350 cm depth, vary inthei r clayconten t between 10an d 20%an d inthei r siltconten t between 20an d 30%. Natural claycomprise sabou t half ofth etota lclay .Th estructur e is weak subangular blocky in the upper part, massive in the lower part of the profile. Ki data vary between3. 5an d 2.7an d Kr data between2. 8an d 2.1.Th eT-valu evarie sbetwee n 3.7 and 5.0 m.e./lOOg ,an d the base saturation decreases with depth from 60t o20% .

The namePard Podzol ha s been given to a highly weathered, well-drained, veryligh t textured soil, with a thin bleached A2 horizon and humus accumulation, without cementation, in the Bhorizon . A profile description isgive n by DAY (1961). NO ana lyticaldat a areavailabl eabou tthi ssoil .

Under GreyHydromorphic soil i sclassifie d a stronglyweathered ,imperfectl y to poor lydraine d soil,whic hha sa ligh ttextured ,bleache dA 2horizon ,tha t gradessharpl yin to a relatively heavy textured, mottled, dense Bhorizon , which is not plinthitic. The main colour of this horizon is grey or light grey, and mottles of reddish yellow or strong brown arepresen t invaryin g abundance and size.Th e silicatecla ymineral s are supposedly of 1:1 lattice structure (kaolinite),an d the basesaturatio n isthough t to be low,t over ylow .

Organic soils have also been encountered in Amazonia. They are waterlogged and very acid. A peaty horizon of varying thickness overlies mineral soil material that is commonly very light textured, and bleached. For these soils,th e names HalfBog soil or Bogsoil ar e applied, if the organic layer is between 30an d 60c m thick, and more than6 0c mthic k respectively.

160 III.3 The Occurrence of the Soils in Relation to theVariou s GeomorphologicUnit s

Inthi s subchapter, a picture isgive n of the distribution of the soilsa s described and classified inth eprecedin g pages.Th e geomorphologic units described in1. 4 are taken asth ebasis .Th esituatio ni sillustrated ,ver ydiagrammatically ,i nFig . 17.Fo rth esoil s

Fig. 17 Mapaesbocado dos solos amazonicos principals na suarelacao comas unidades geomorfologicas

70* W

Holocene terrains - Saline and Alkali Outcropping crystalline basement and soils outcropping Paleozoic, Mesozoic and TerrenosHolocenos - Solos Salinos e Early Tertiary deposits,a s well as Early Alcalinos Tertiary(?) peneplanation surface - Holocene terrains - Low Humic Gley Red Yellow Latosols, Dark Red and Humic Gley soils Latosols, Red Yellow Podzolic soils, Terrenos Holocenos- Solos CleiHumico Lithosols, Ground Water Laterite soils e CleiPouco Humico Embasamento cristalino aflorante e depositosaflorantes doPaleozoico, Ground Water Laterite soils Mesozoico ou do Tercidrio Inferior, bem m SolosLalerita Hidromorfica como superficie depeneplanacdo do Pleistocene terraces- KaoliniticYello w Tercidrio Inferior(?) - Latosolos Latosols (of varying texture) and Vermelho-Amarelo, Latosolos Vermelho Kaolinitic Latosolic Sands Escuro, solosPodzolicos Vermelho- Terracos doPleistoceno -Lalosolos Amarelo, Litosolos, solos Lalerita Amarelos Caoliniticos (de textura varia- Hidromorfica da) e Areias Latosdlicas Caoliniticos Cretaceous(?) peneplanation surface - Amazon planalto - Kaolinitic Yellow Areno-Latosols Latosols of very heavy texture Superficie depeneplanacdo Cretdcea(?) (4 + 5: Planicie) -Areno-Latosolos Planalto amazdnico - Lalosolos Amarelos Caoliniticos de textura muito pesada (4 + 5:Planicie) Fig.17 Sketch map of themain Amazon soilsin their relation to thegeomorphologic units of the Planicie, which have been comparatively thoroughly studied, the influence of several soilformin g factors isdiscusse d in some detail.Thes efactor s are Time (degree ofprofil e development, depending uponterrac e level),Man (Terr a Preta),an d Ground Water(Groun d Water Laterite soilversus Ground Water Podzol).

III.3.1 The Soils of the Undulating Terrains with Outcropping Crystalline Basement

Only a few data are available as to the soils in the regions where the Pre-Cambrian crystalline basement, ofth e Guiana and the Central Brazilian shields, outcrops. Inth e Araguia Mahogany area (cf.Appendice s 2an d 6),o n the outskirts of Amazonia, the crystalline, western section consists of mica schists and quartzites. The mica schists, whichoccup yundulatin gt ohill yterrain ,hav eresulte d inth eformatio n ofshallo wred dish soilswhic h have been classified as Red YellowPodzolic soil, with ratherhigh base saturation (RPrhbl Profile 37). The quartzites are present in mountainous terrain (Dome formations), and have resulted in Lithosol, Quartzite substratumphase (L, Q). On the crystalline rocks elsewhere in Amazonia (granites, gneisses, syenites, mica schists),th edeep ,friable , reddish or yellowish Red Yellow Latosols (RL) are probably themos tfrequent . Anexampl eo fsuc ha soi li sth eprofil e from Serrad eNavi oi nAma - pa Territory, which has probably developed on amphibole and mica schists (Profile 34).Example so fprofile s ongranites ,gneisse san d syenitesar eno t available.Th ecom paratively small areas with basic effusions (gabbro, diabase, dolerite) have probably Terra Roxa (Legitima or Estruturada) soils for the most part.

III.3.2 The Soils of the Undulating Terrains with Outcropping Paleozoic, Mesozoic or EarlyTertiar y Deposits

The soils on these terrains, where consolidated and slightly metamorphosed sedi ments of a whole series of geological periods are found within short distances (cf. Appendix 8an d the Figs. 5an d 6), are naturally very diverse. It has to be borne in mind that a large portion of the areas indicated on the geological maps as having these deposits, have in fact a thin cover of Late Tertiary or Quaternary sediments. Areas with soils derived from Paleozoic, Mesozoic or Early Tertiary deposits are therefore comparatively small. Lithosols, of hematized sand-stones and shales, are apparently rathercommo namon gthem . No data are available on the soils of the Silurian and Pre-Silurian deposits of the Amazonbasin . Alittl ei sknow nabou tth esoil sderive d from theDevonia ndeposit so fbot hth e Ama zon and the Maranhao basins.I nth e Araguaia Mahogany area (Appendices 2an d6) , the principal soil on the Pimenteiras beds is a Dark Red Latosol, Shallow phase (DL,s;Profile35). SAKAMOTO(1960)mention sa soi lo nth eMiacuri io fbedsth eAmazo n

162 basin, in the area of the lower Trombetas and the Erepecuru rivers. He describes it as'loos esand swit ha limite d amount ofclay ,ligh t orangean d brilliant reddish orange coloured over more than 1m \ It is possibly a Red Yellow Latosol. In describing a red shale, occurring at the lower Jari river and supposedly of Devonian age, the same author states that it weathers into yellow kaolinitic clay. The Carboniferous deposits of the Amazon basin are very interesting economically becauseo fth epresenc eo flime-stones ,evaporite san d diabase(th elatte r areJurassi ci n part).I nth e MonteAlegre-Alenque r area,a fe w studieso fth esoil sderive d from these deposits have been made.Th e lime-stones have given a Grumosol,whil e aTerra Roxa Estruturada has developed on the diabase (Excursion Eight of the National Brazilian SoilCongress , 1961; cf notes under(XIII ) of III.2).Th e evaporites are likely to have resulted inth eformatio n ofsoil ssimila r tothos epresen t upon the Pastos Bons-Motu- ca beds of the Maranhao basin (cf. below). No consistent data are available on the soils on the less rich deposits within the Carboniferous strata of the Amazon basin (cf., however, Profile 17o f II.3). The Carboniferous strata in the Maranhao basin are also very diversified. In the Araguaia Mahogany area (Appendices 2an d 6),Acid Brown Forest-like soil, Gravelly phase(AF , G)ha s been found to be the common soil on reddish brown silt-stones, which are probably of the lower section of the Piaui beds. The presence ofLithosol, Sand-stonesubstratum phase (L ,ss )ha sbee nestablishe d on coarsegraine d sand-stones which probably belong to the Poti beds. In the same area, Carboniferous lime-stones occur, associated with very resistant chert (chalcedony) in the superficial layers (Piaui beds, e.g. Pedra de Fogo beds and upper section of Piaui beds). The soil is therefore only a Lithosol,Cherty substratum phase(L , CH), or a very gravelly Non CalcicBrown-like soil (NB, G). On flat land surfaces, with the same substratum, Hydromorphic GreyPodzolic soil, with high basesaturation, Shallow phase (HPAS, s; Profile 46)ha s developed. Alsoa few data are available on the soilsderive d from the Lower Mesozoic deposits of the Maranhao basin. As to the Jurassic-Triassic deposits, it can be said that on approximately flat terrains in the Araguaia Mahogany area, with a substratum of silt-stones and clay-stones that are, in part, calcareous and gypsiferous (Pastos Bons-Motuca beds; Piaui beds of other geological mapping), various hydromor phic soils have developed (mapping unit H of Appendices 2 and 6; cf also Fig.23) . Among them are Hydromorphic Grey Podzolic soil,with high base saturation, Ortho and Darkphase (HP/,6 , o and HPhb, D; Profiles 44an d 45 respectively) and Ground Water Lateritesoil, intergrade to Hydromorphic Grey Podzolic soil (Deep phasean d Clay-stone substratumphase, GL-HP, D and GL-HP,c ; cf Profiles 14 and 15 respectivelyo fII.3) .Undulatin ggroun dcompose do fcoars egraine dsand-stone s(Pasto s Bons-Sambaiba beds)wa sfoun d to have KaoliniticLatosolic Sandprofiles , Transition phasean d Savannahphase (KLS ,T an d KLS,s ; cf Profile 33 for thelatter) . The survey of the Guama-Imperatriz area (Appendices 1an d 5) provides for some cluesa st o thesoil sderive d from the Cretaceous deposits ofth e Maranhao basin. Red Yellow Mediterranean-like soil(RM ; Profile 41)i sth ecommo n soilo fth e undulating

163 country, at thesouther n end ofth earea ,whic hi scompose d ofth erelativel yric h Codo beds.I n thesam estretch ,patche so f Ground WaterPodzol (GP )hav ebee nfound . This soil is likely to have developed on the sandy Corda beds which occur immediately belowth eCod 6beds . SAKAMOTO (1960) states that the sediments of the Itapecurii or Serra Negra beds (Late Cretaceous, or Tertiary) contain a small percentage of volcanic elements, con trary to the sediments of the Late Tertiary Barreiras beds. He thinks it likely that the soilsderive d from the former sediments have a higher content of bases. In the meagre amount ofgeologica lliteratur erelevan tt othi sarea ,i ti sassume dtha tth eItapecuri i or Serra Negra beds occurs in a stretch of the Guama-Imperatriz area north of the zone of outcropping of the Cod6 beds. In studying Appendices 4an d 5i t is,however ,evi dent that the land surface in this stretch was modelled during the Plio-Pleistocene epoch (planalto with Belterraclay )an dth ePleistocen e epoch (terracesa tlowe rlevel) . Thecharacteristicall yrathe rhig hbas esaturatio n ofth e Kaolinitic Red Latosol occur ring on the terraces inth e stretch (cf. III.3.4), may however be an indication that the modelling wasdon eexclusivel y withth e above-mentioned, volcanicelement s contain ingEarl yTertiar ysediments . The Cretaceous and Early Tertiary sediments of the Amazon basin are apparently nearly everywhere covered with athi n layer of Late Tertiary and/or Pleistocene mate rials (cf. IH.3.4). In contrast, the Cretaceous and Tertiary deposits of the Acre basin outcrop, according to indications, inth e eastern part of Acre State over considerable stretches.I nthi sarea ,th e presenceo f RedYellow Podzolic soil, with lowbase saturation (RPib; Profile 36)ha s beenestablished .Als oa dar k red,friable , clayeysoil ,provision ally classified asDark RedLatosol (DL ) occurs.I t wasno t possible to makea definite correlationo fthes esoil swit hknow n geologicalstrat a(cf. Appendix8) .

III.3.3 The Soils of the Cretaceous and/or Early Tertiary Peneplanation Surfaces

Nofield dat a exist on the soils of the presumably Cretaceous plateaux in the transi tion zone between Amazonia and Central Brazil. On the FAO Soil Map of South America, second draft, they are indicated to have Areno-Latosols or Rego-Latosols. The supposedly Early Tertiary peneplanation surface, occurring about halfway be tween these plateaux and the Amazon river, are likely to have a predominance of Hydromorphicsoils an dLithosols. It is believed that the samesoil spredominat e inth e savannahcovere darea so fth eboundar y regiono fPar a Statean dth eGuiana s(Cretac eous and/or Early Tertiary). According to indications, the terrains concerned are extensivelyflat, wit ha numbe ro fcrystallin eridge s(cf. discussion s inIV .1.2) .

164 III.3.4 TheSoil so fth ePlanici e

Asdiscusse d in1.4 , theupland si nth ebroa d axial part of Amazonia consisto f uncon solidated kaolinitic sedimentslargel y of thevaguel ydate d Alterd o Chao or Barreiras beds.The y areshape d intoa heav ycla ycovere d Plio-Pleistoceneplatea u (theAmazo n planalto with Belteira clay), and more or less sandy Pleistocene terraces at a lower level.Thes e uplands are collectively called Planicie.Th e soil data on this Planiciear e relatively numerous. A general discussion of the soil qualities in relation to the agricultural capabilitiesi sgive ni nChapte rV .

II1.3.4. 1 TheSoils of the Amazon Planalto The soils of the Belteira clay covered Amazon planalto have an uniform profile development over large areas. In the eastern part of the region, east of Manaus, the soils are very predominantly Kaolinitic Yellow Latosol(,Ortho), veryheavy textured (KYLVA; Profile 24). Its presence has been established in a considerable part of the Guama-Imperatriz area (cf. Appendix 1),nea r Tucuruf on the Tocantins river, along Curua-una river (cf.Fig . 11),a t Belterra Estatealon gth elowe r Tapajos river(cf. Fig. 20)an d in the central section of the Manaus-Itacoatiaia area. There isa gradation in the compactness of the subsoil,whic h hasapparentl y an influence upon the quality of thefores t cover(cf. IV.1.1.2) .Th ethicknes so f theA xsubhorizo n alsovarie sconsider ably. Terra Preta soil (TP„A) only occursver y locally onth e planalto(cf. Fig. 20). Asdiscusse d in1.4.2 ,th e Belterracla y covered plateau isrelativel y low-lying inwest ern Amazonia (west of Manaus), and the clay itself probably slightly less heavy. The plateau has been studied near Porto Velho,alon g a 100k m stretch ofth e Acre- Bra siliahighwa y(BR-29) .Th esoil si nthi sstretc h are,fo r apart , Kaolinitic YellowLatosol (,Ortho), heavy textured (KYLA) and Kaolinitic Yellow Latosol, intergrade toDark Horizon Latosol, heavy textured (KYL-DHLA). The central sections of planalto have, however,heav ytexture d Kaolinitic YellowLatosol, intergradeto Ground WaterLaterite soil(KYL-GLft j Profile 12). It is likely that about the same soil situation exists for plateau land inwester n Acre State.O nth e low piateau land between the lowei Purus and the Catua rivers, (very) heavy textured Kaolinitic Yellow Latosol, intergrade to Ground WaterLaterit esoi l(c.q. Planosolic Latosol; cf.III.2 )i sprobabl y predominant. The notes of MARBUT and MANIFOLD (1925, 1926) concerning the dissected plateau landi nal lth eregio n westo fth eCatu a river,poin tt ogoo ddrainage .The ysugges tthei r soil group 3:'cla y loams and clays with red or reddish friable clay subsoils',t o be the most common on these plateau parts (cf.Fig . 18).Sinc e thisgrou p comprisesth e soil onth e planalto south of Santarem, it mayb eassume d that theKaolinitic YellowLato sol (,Ortho), (very) heavy textured(KYL t-ft or KYL*) is also predominant on the planalto areasi nthi spar t ofAmazonia . Theonl y available soildat a ast oth e planalto stretches inth e Peruvian-Columbian part ofAmazoni a are alsothos e of MARBUT and MANIFOLD. They indicate a similarity to the adjoining Brazilian parts. In Bolivia, however, much of the plateau land is imperfectly drained (Madre de Dios area; cf. ARENS, 1963).

165 Fig.18 Mapa de MARBUTan d MANIFOLD (1926) dosgrupos de solosda Bacia Amazonica Interior Fig.18 The soilgroups of the InnerAmazon Basinaccording to MARBUTan d MANIFOLD'S map (1926)

III.3.4.2 The Soilsof the PleistoceneTerraces Both the uplands with sediments of undisputedly Pleistocene age (cf. 1.3)an d those with supposedly Late Tertiary sediments that were however re-modelled into terraces duringth ePleistocen e (cf. 1.4.3),hav esoil stha t arecomparativel y wellknow nbecaus e ofthei reas yaccess .

MAIN SOILS Inth eeaster n part ofAmazonia ,al lth eare a east of Manaus,th eprincipa l soilso nal l Pleistoceneterrace sar eyellowis han d deeply friable, and of varying texture. They are classified asKaolinitic YellowLatosol (,Ortho), medium, ratherheavy or heavy textured (e.g. Profile 25: KYLW).The yhav ebee nmappe d for largestretche so fth eGuama-Im -

166 --0"

400 KM -I

'Sandy loams with yellow friable subsoilsusuall yconsistin g of sandy clay'

JvXvX-l 'Fine sandy loamswit h predominantyello wt oreddis hyello w -—8° Iv.v.v . 1 friable clay tosand ycla y subsoils' \ S

pr-T-T-j 'Clayloam san dclay swit h redo rreddis hfriabl ecla ysubsoils '

I ••'•*•'• 'I 'Clays andcla y loams with heavy imperfectly oxidised clay ' !• • ' ' " subsoilsusuall ya ta dept ho ftw o feet orless '

|:|;|:|:|:|| 'Veryfine sand y loamswith reddisho r reddishyello wfriabl e rrl:l;hU claysubsoils ' 60°

'Alluvialsoils '

peratriz area (Appendix 1).I nth e Caete-Maracassume area (cf. Fig.19 )an dth e Bra- gantina area they are the predominant soils (mapped as 'Yellow Latosol' by DAY (1959) and FILHOet al. (1963 ) respectively; medium textures predominate).Th eligh t textured relativest othes e soils,classifie d asKaolinitic Latosolic Sand(KLS) ,ar enor mally present near the rivers andar eo f frequent occurrence in the Lower Amazon region.Th eupland s with Kaolinitic Yellow Latosols orKaoliniti c Latosolic Sandar e largely under forest. A number of the savannah terrains, however, also have these soils(cf. IV.1.2.1) .I ntha t instance, 'Savannah phase'i sapplied .

PROFILE DEVELOPMENT IN DEPENDENCE OF TEXTURE AND TERRACE LEVEL Many profiles of bothth e Kaolinitic Yellow Latosols and the Kaolinitic Latosolic

167 Fig.19 O levantamento expeditoda area Caete-Maracassumi por DAY (1959)

«°30'

1"30 S

___ Association Maracassume ('Yellow Latosol' + 'Ground Water Laterite' predominantly) ^^ Associacao de Maracassume" (predominantemente 'Latosolo Amarelo' + 'Laterita Hidromdr- fica') . , Association Gurupi ('Ground Water Laterite, low phase' predominantly) ' Associacao de Gurupi(predominantemente 'Laterita Hidromortica,fase Baixa') Association Pitoro ('Ground Water Laterite' + 'podzolised Ground Water Laterite' + UEk i Yellow Latosol' predominantly) Associacaode Pitoro(predominantemente 'Laterita Hidromorfica'+ 'LateritaHidromorfica podzolizada' + 'Latosolo Amarelo") 'Low Humic Gley soil' 'Solo Glei Pouco Humico' 'Saline soils' 'Solos Salinos" „_ 'Ground Water Laterite, low phase' b&vd 'Laterita Hidromorfica,fase baixa' . . 'Agricultural soils, undifferentiated' ' ' 'Solos agricolas, nao diferenciados'

Fig. 19 Thereconnaissance soilsurvey of the Caete"-Maracassume area by DAY (1959)

168 Sands have been studied. There is a general and clear tendency for the profiles of heavier texture to have shallower B2 horizons, a smaller textural ratio B/A,thinne r At subhorizons, and thinner superficialfilms o f loose bleached sand ('micro podzol'). Iti sinfe ire dtha tther ear eals osligh tdifference s inth emorphometr y and inth ephysic al and chemical properties of the profiles, which are related with the terrace level on which the profiles occur. Since higher terraces are of greater age, this may enable an evaluationt ob emad eo fth einfluenc e ofth esoi lformin g factor Timeo nthes e profiles Thisi sespeciall ypossible ,becaus eothe r soilformin g factors - Parent Material,Topo graphy, Vegetation and Soil Fauna, Ground Water, and Climate - are fairly equal throughout the Planicie of eastern Amazonia. The profiles on higher level,i.e. older, terracessho wth efollowin g trends:

The superficial film of bleached loose sand is slightly thicker; the Ax subhorizon is slightly thicker; the depth of the horizon of maximum clay accumulation and mini mum siltcontent , i.e. the B2subhorizon , islarger ;th edifferenc e intextur e betweenth e A and the B horizons, as expressed in the textural ratio B/A, is slightly larger; the silt/clay ratio isslightl y smaller;th eactivit y ofth eclay-size d particlesi sslightl ysmall er;th eC/ N valueo fth eorgani cmatte r issomewha tlarger .Th eP-fixatio n and theper centageo fexchangeabl e(Al) +ma ydepen dals oo nth eterrac elevel . Thesetrend s are separate from the above-mentioned much larger differences in mor phometry whichar ea consequenc e ofa differenc e inth eover-al ltextur eo feac hindivi dual profile. For studying of the time factor therefore, profiles of similar general tex tureshoul d becompared .A quantitativ e evaluation isno t possiblebecaus eno tenoug h precisedat a areavailabl e onprofile s withcomparabl etextur efro m completelyflat an d non-eroded partso fth eterrace sa tdifferen t levels.Moreover ,th edifference s duet oth e time factor are so smalltha t they cannot be measured with accuracy with the method of horizon sampling used and the present day level of refinement of laboratory ana lysis. Minutedifference s in parent material,i nth e forest cover,th esoi lfaun a or inth e climatema yals oobscur eon eo rmor etrends .Th ementione d aspectso fprofil e 'ageing' are therefore only exemplified by two profiles of Kaolinitic Yellow Latosol (,Ortho), mediumtexture d(KYL m).The yoccu r both onflat an dnon-erode d sectionso fterrace s in the northern part of the Guama-Imperatriz area, and are both under primeval forest cover(Tabl e 10,th enumber s23 3 and231) . Thedat a ofthre eprofile s ofKaoliniti c Latosolic Sand (KLS)ar eals ogiven ,althoug h the terrace levels of these latter profiles are not well established. Parent material and climate of Profile 296 are moreover probably slightly different from those of the numbers 45an d 169. That under longexistin gan d man-induced savannah growth(cf. IV.1.2.1) ,th etrend s are stronger, becomes apparent from the data on number 337.Th e much clearer in fluence of the over-all texture on the variations in morphometry can bejudge d from the data of two profiles of Kaolinitic Yellow Latosol (,Ortho), very heavy textured (KYLDA) of thenon-erode d planalto, also under primeval forest cover.

169 Table 10 Variations in a number of data on soil profiles from terraces at different levels, showing the influence of Time

Classifica No. tion Terrace field classifica- level descr. (00 nivelde decsr. de (cf.Tabl e9 ) Location Vegetativecove r terraco Age campo (cf. Tabeta9) localizacao cobertura vegetal (m) idade

42/112 KYL„„ Curua-una, km6. 5 primeval forest 180 Calabrian floresta primitiva Calabriano 210 KYLBA BR-14, km247. 0 primeval forest 200 Calabrian florestaprimitiva Calabriano

233 KYLm BR-14, km 12.7 primeval forest 65 Milazzian floresta primitiva Milaziano 231 KYLm BR-14,k m58. 0 primeval forest 3-4 Epi-Monastirian florestaprimitiva Epi-monastiriano 337 KYLA Amapa-Fazendinha, anthropogenicsavanna h 3-4 Epi-Monastirian km 8 savana antropogenica Epi-monastiriano 45 KLS Curua-una,k m5 ca. primeval forest 100 ca. Sicilian? floresta primitiva Sicilianot 169 KLS Curua-una, km2. 5ca. primeval forest 60 ca. Milazzian? florestaprimitiva Milaziano! 296 KLS Araguaia, Rio Corda primeval forest 10 ca. Late Monastirian floresta primitiva Monastirianosuperior

Constitution of Bt horizon / constituicao do horizonte B%

Base Moist. No. Si02: Clay Cationexch . cap. saturation equiv. field SiO,: (AI.O, + argila capacid. de troca saturacao equiv. descr. Al2Oa Fe8Os) «2(i) Carb. (NH4OAc-pH = 7) de bases de umidade descr. de (%) (°/o) (m. e./lOO g) (%) (g/100g) campo Ki Kr c. C T V M.E.

42/112 1.90 1.71 88.0 0.83 4.81 10.8 35.8 210 1.77 1.48 88.5 0.69 4.56 14.0 34.0 233 1.85 1.63 20.4 0.16 1.69 23.0 9.8 231 1.82 1.63 23.9 0.22 2.14 18.7 13.2 337 1.83 1.63 53.8 0.30 2.07 17.9 25 45 1.60 1.23 10.0 0.22 1.06 37.7 4.7 169 1.50 1.24 14.9 0.08 - - 6.7 296 1.77 1.53 10.9 0.09 1.61 25.5 6.0

') Silt-Int. fraction estimated,b ygraphica l interpolation on summation curve,wit h knownsilt-U.S . / Fracaosilte-Int. estimada, por interpolacaogrdfica em curvacumulative!, com silte-U.S. conhecido. *) Depth of central part of B, / Profundidade daparte central do B, %20z2u } %<2n * 1U0 > ') a Estimated, by graphical comparison of horizons of the individual profile that have approximately identical percentage of Carbon / Estimado, por comparacdo grdfica de horizontes do perfil individualQue apresentam percentagem de Carbono aproximadamente identica

170 Tabela 10 Variacoes em dados de per/is de terracos de diferentes niveis, moslrando a influenciadofalor Tempo P-fix Thick Bleached ness sand cover Depth Textural Int. silt fix. de (Al) + in espes- coberlura profun- ratio clay PnoBi B2 sura deareia didade relacao (%ofT) Classif. silie Int. C/N of Ido A! alvejada Bj" /P-total\ (AI) + {cf.Tabl e 9) (cm) (cm) (cm) BIA argila* A, B2 \P-Brayj no Bt (cf. Tabela 9)

30 0 65 1.12 4.9 13.7 13.8 - - KYLBft

2 0 40 1.12 5.4 10.9 8.6 150 23 KYL„ft

20 2.0 230 1.82 3.4 14.7 7.7 225 39 KYLm

20 0.5 100 1.15 16.7 9.7 7.3 75 25 KYLm

15 - > 90 1.60 7.41 15.4 15.0 130 16 KYU

35 7 >120 1.85 2.0 17.2 22.0 - - KLS 20 7 350 2.07 2.7 14.6 8.0 - - KLS

10 0.5 255 1.31 34.91 9.7 4.5 66 20 KLS

Cation exchange cap. of clay in B, Moisture equiv. of pure capacidade totalde troca daargila doB t Spec.surf . clav inB 5 NH OAc-pH = 7 of clay i n Bs equivalenli• de umidade 4 NaO Ac - daargila pura do B pure clay/ argila pura superf. espec. t pH = 8.2 daargila 1 Classif. (m.e. /100g) (m.e./lOOg ) clav/ areila* do B, (cf. Table 9) ! a* b» c* d4 (m.i e./lOO g) (m /100g) (cf. Tabela9)

39 46.7 0.5 1.8 4.8 54 KYLvh 43 41.0 1.3 2.1 4.8 81 KYL„»

45 43.8 1.5 5.2 6.5 73 KYLm 57 49.6 3.7 5.4 - - KYLm 46 43.6 2.7 1.7 - - KYLft 33 43.3 2.9 2.0 - - KLS 35 43.9 8.6 - - KLS 77 53.7 12.2 11.6 — - KLS

b Calculated, using the factor 1.7 for the activity of the organic matter (cf. V.3.1.1) / Calculado, usandoo fator 1.7 para a atividade da materia orgdnica(cf. V.3.1.1) c Estimated, by graphical comparison of horizons of the individual profile that have approximately identical percentage of clay / Eslimado,por comparacao grdfica de horizontesdo perfit individual que apresentam percentagem deargila aproxima- damenle identica d Calculated, using the factor 3.9 for the activity of the organic matter (cf. V.3.1.1) / Calculado. usandoo fator 3.9 para a atividade da materia orgdnica(cf. V.3.1.1) ' Organic matter not destroyed / Materia orgdnica naodestruida

171 OTHER WELL-DRAINED SOILS Reworked Belterra clay is present on parts of the higher Pleistocene terraces (cf. 1.4.2). Themai nsoi ldevelope do nthi scla yi smor ecompact ,an da tth e sametim e paler thanth esoi ldevelope d on Belterracla yo nth eplanalt oproper .I ti sclassifie d as Kaolin itic Yellow Latosol,Compact phase, very heavytextured (KYL , cVh', Profile 26). The presence of this soil has been established for a part of the Guama-Imperatriz area, where peculiar hummocks,called jabot is, ar ea n associated terrain feature (cf. Appen dix 1,k m 130-km 190;Appendi x 3 and Photo 20).I n the headwater region of Igarape do Lago in Amapa Territory, the soil is found also, largely under savannah cover. Also near both Manaus and Itacoatiara, where the reworked Belterra clay seems to havea sligh t admixturewit hcoarse r sediments,th esoi l hasa compac t appearance. Also relatively compact and pale profiles may be found which possess lighter textu res.The ysee mt o berestricte dt o the older Pleistocene terraces,fo r instanceth e oneo f about 60m +(Milazzian) .A nexampl ei sth eKaolinitic YellowLatosol, Compactphase, ratherheavy textured(KYL, crhl Profile 27) of the terrainseas t ofth e stretch km 50- km9 0o fth eGuama-Imperatri z area.

In the southern stretch of the Guama-Imperatriz area, which has a pronounced dry season although itha sa fores t cover,a soi lmuc hcomparabl e toth e Kaolinitic Yellow Latosol (,Ortho), medium textured has been found. It is, however, of redder hue and has a higher base saturation. This soil isclassifie d as Kaolinitic RedLatosol, medium textured(KRL m; Profile31) . Well-drained soilswer eals ofoun d tooccu rwhic hhav ea reddis h subsoil( Bhorizon ) below a yellowish topsoil (A horizon).Whe n thesesubsoil s arefriable , then such pro files are classified as Kaolinitic Yellow Latosol, intergrade to Red Yellow Podzolic soil. This a common soil in the stretch km 90-km 140o f the Guama-Imperatriz area (Ap pendix 1 and 3),wher eth esubsoi li slargel y rather heavy textured (KYL-RP^; Profile 29). In part, these reddish subsoils are firm, and then also the textural ratio B/A is rather large. Such profiles, which are classified as Red Yellow Podzolicsoil, intergrade toKaolinitic YellowLatosol, were frequently found to occur in the central part of the Guama-Imperatriz area. The texture of the subsoil varies there from rather heavy to very heavy (RP-KYLrA and RP-KYLVA; Profiles 39 and 38 respectively). Similar soilshav ebee n described for the Caete-Maracassume area, more especiallyth e south western part of the area of the 'Association Pitoro'. At that time, the name 'Red Yellow Podzolic' wasapplie d (DAY, 1959;cf. Fig. 19). Manypart so fth ePlanici ei neaster n Amazonia have fossil plinthite near the surface. The well or moderately welldraine d soils which have a considerable quantity of such fossil plinthite in the solum, are discussed extensively in II.3.2.2. They are classified eithera sKaolinitic YellowLatosol, Concretionaryphase (KYL , CR;Profil e 22,28) , ora s Red Yellow Podzolic soil, intergrade to Kaolinitic YellowLatosol, Concretionary phase (RP-KYL, CR;Profile s 19; 21,an d 20,40).Th eforme r soili sfoun d insizeabl eexpans e east of Belem (LatosolConcreciondrio of FILHO et al., 1963).Th e latter soil occurs frequently in thenorther n part of Guama-Imperatriz area.

172 IMPERFECTLY OR EXCESSIVELY DRAINED SOILS Imperfectly drained soils with plinthite-in-formation within the solum, i.e. Ground WaterLaterile soils (GL),occu r in anumbe r of areas ofeaster n Amazonia. In water shed parts of extensive flat terrains, where savannah or savannah-forest forms the vegetative cover, the profiles are often outstandingly hydromorphic, constituting the so-called Ground Water Lateritesoil ,Lo wphase .Man ydescription so fth eoccurrenc e of these hydromorphic soils are giveni n IV.1.2. Discussions on the variability in the profile characteristics, depending on the character of the fluctuation of the ground water level,th e texture ofth e parent material,an d the degreeo f pre-weathering of the parent material, are given in II.3.2.1. The Profiles 2, 4, 5, 7/43, 8, 9, and 10 are of Ground Water Laterite soilso n Pleistocene terrains in eastern Amazonia. Imperfectly to poorly drained soilswit h a bleached surface horizon and an Ortstein, i.e. Ground Water Podzols, have also been encountered, commonly in strip-like patches. The few analytical data suggest that GroundWater Humus Podzols(GP ; Profile 47) are involved.I t canb einferre d from thedetaile ddescription si nIV.1.2 ,tha t thesesoil softe n bearsavanna ho rsavannah-fores t ofa kind. It may be mentioned that Ground Water Laterite soils and Ground Water Podzols have a different position. The former predominate, as extensive units, on watershed areaswit hflat relief ,an d thelatte r on narrow stripso flo wuplan d alongth eriver san d on former river beds, when these terrains are sandy. This implies that on imperfectly drained siteswher e lateral ground water movement does not take place, Ground Wa ter Laterite soils develop, while on imperfectly drained sites where the ground water can move laterally, Ground Water Podzols develop. It may be concluded that on imperfectly drained watershed terrains there is merely a concentration, partly by downward movement (bleachedto po fGL ,Lo wphase) ,o fsesquioxide sint omottle si n the zone offluctuating phreati c level. On imperfectly drained sandy terrains along ri vers, on the other hand, only the Al component of downward moving sesquioxides concentrates,togethe rwit hhumus ,int oa homogeneou slayer ,whil eth eF e component iscarrie d away to the river (thelatte rconclusio n mayb edraw n onlywhe nal lGroun d Water Podzol profiles prove to have indeed only aluminum concentrated in their Ortsteins). An excessively drained soil which consists,t o a great depth, of bleached sand and is classified as White SandRegosol (WSR ; Profile 48),seem st o have a restricted, patch likeoccurrence ,commonl y onrelativel yhig hterrace swit hsand yparen t material.Thi s soilusuall ysupport ssavannah-fores t(cf. IV .1.2) ,bu t itma yb efoun d underhig h forest.

Shallow Podzols on sandy terrains in good drainage position were found to occur also, but very locally: Para Podzolo f DAY (1961). This Para Podzol occurs in close association with,an d often ina positio nbetwee nth e Kaolinitic LatosolicSan dan d the WhiteSan d Regosol.Th esoi l istherefor e probably anintermediat estag ei na weather ing process that may transform Kaolinitic Latosolic Sand into White Sand Regosol (cf. DAY, 1961). Because of the excessive thickness of the bleached sand layer of the White Sand

173 Foto 17 Algumas das cerdmicas das tribos indiaspri-colombianas que -com freqiiencia se acham nos solos de Terra Preta. Veem-seum sapo,a irompa de uma anta, as cabefas de irespdssaros e deurn macacoe uma estatueta deferiilidade

Bfe. * . s* ' « »>*"••" "9^jr^ M *

A* ' * v* . *• *'7%i' * t •. **

Wow /7 So/ne o//Ae ceramicsof Pre-ColumbianIndian tribes that are commonlyfound in the Terra Pretasoil. There can beseen a toad, the trunk of atapir, the heads of threebirds, a monkey and a fertility symbol

Regosol, it can normally not be established whether or not a humus accumulation occurs below the layer. The above mentioned physiographic position of the soil, however, implies that the White Sand Regosol islikel y to be actually very deeply and strongly wheatered in situ. Instead of Regosol, a very deep Podzol or Ground Water Podzol would be concerned. Even in the case that a layer of bleached sand was deposited as such, alluvially or colluvially, one may assume that the layer will have increased in thickness because of podzolisation, which seems to be the only possible soil forming process on such a site. In any case, therefore, a name as 'Giant Podzol' mayb emor eappropriat e for thesoi ltha nWhit eSan d Regosol.

TERRA PRETA SOIL Throughout the Planicie ineaster n Amazonia thereca n befoun d patcheso f so-called TerraPretaor Terrapreta do indio (TP ; Profile 53).Thi si sa n upland soilwit ha thick, black ordar k grey top,contain s pieceso fartefacts , and isfame d locally for its fertility. GOUROU(1949 )mention sa numbe ro fplace sw her eth esoi loccurs . The patches are invariably small, often not exceeding 1000 m2. They are usually

174 located at siteswher eth e upland isnea r thenavigabl ewaterways .The y areespeciall y frequent at outer bends of the rivers,wher e nofloodplains occur s between the water and the upland, and whereth e waterway can be scanned freely. Many of the TP pat ches occur at, or near, present-day villages or towns. Their location implies that the soil predominateso nth e Pleistoceneterraces ,no t onth e planalto. It isonl ywher eth e latterclosel yapproache sth e mainrivers ,bein gbande d byonl ynarro w stripso flowe r terraces, that patches of TP may befoun d on its edges.Th e latter isth ecas e in the Santarem-Belterra area, where the planalto edges provide for an excellent view over the area of theconfluenc e of the riversTapaj6 s and Amazonas.Th epatche so f TP on Belterra Estatehav ebee nmappe d(cf. Fig.20)1.O neaster n Maraj6 islandth eT Psoi li s found, likewise as small patches, on the ridgy elevations (tesos) within general low land ando nth ehighes tpart so fcontinuou slo wupland . Apparently, all TP profiles contain, in their dark toplayer and/or just below this, pieceso f artefacts, in varying frequency. Theyca n be identified with ceramics ofpre - Columbian Indiantribe s(cf. HILBERT, 1955). Thefantas y employedi nproducin gthes e ceramicsi sillustrate d inPhot o 17. It canb etake nfo r certain that thepatche swer eoccupie d byth eforme r Indianpopu lation of Amazonia. Those small groups of wild Indians still inexistenc etoday , have been driven far away from the means of communication and the population centres.

Fig. 20 Mapade solosdo Estahelecimentode Belterra no baixorio Tapajos {executado por Jost BENITO SAMPAIO, 1962)

54"55 W KLS = Kaolinilic Lato- solic Sand,'Areia Latoso- lica Caolinitica KYL« = Kaolinitic Yellow Latosol, medium textured La/oso/o Ama- relo Caoliniiico, de tex- luramedia K.YU* = Kaolinitic Yellow Latosol, very heavy textured,Lawsolo Amarelo Caoliniiico, de texturamuito pesada TP = Terra Preta TM = Terra Mulatta

Fig. 20 Soil mapof Belterra Estateon the lower Tapajos river{executed by J. B. SAMPAIO, 1962)

') The mapping unit Terra Mulatta (TM), which is characterised by a slightly less dark topsoil than the TP(TerraPreta )an d the absenceo fartefacts ,occur si n abroa dban daroun d several of the TP patches. It seems likely that this soil has obtained its specific properties from long-lasting cultivation. The gardens around the former Indian villages were pro bably situated here. No analytical data are however available concerning these TM soils, which have not been observed outside the Santarem - Belterra planalto.

175 At the time of the first European penetration, however, the then populous Indian tribes used to dwell in well developed communities on dry terrains along the water ways, which were the best sitesfo rfishing, huntin g and warfare strategy (cf. MILLER, 1954)v. There have been many discussions on the subject ofth e origin of the TP fertility (cf. GOUROU, 1949).Th efollowin g twohypothese spredominate : 1. The Indianschos ethes e settlements because of higher natural fertility ofth esoil . 2. The present day fertility issolel y due to prolonged Indian occupation. Previously, thesoil swer en obette rtha nth esurroundin gones .

Onlyth e second hypothesis can beth e right one,fo r the following reasons: a. Thetextur e ofth eT P profiles isvarying ,bu t alwayscomparabl e totha t ofimmedi ately surrounding soils(usuall y more or less sandy; sometimesconcretionary , or very clayeya so n Belterra Estate). b. The composition of the clay fraction of the TP profiles is identical to that of the surrounding soils. It consists very predominantly of kaolinite (Ki values between 1.7-2.0;K rvalue sabov e1.5). * c. The deeper subsoil (Chorizon ) and the substratum of the TP profiles are the same asthos eo fth esurroundin gsoils .

Apparently, the TP soil isa kind of 'kitchen-midden', which has acquired its specific fertility, notably much Calcium and Phosphorus, from dung, household garbage, and the refuse (bones) of hunting and fishing. For detailed discussion of the TP fertility please refer V.3.1.2.

THE PLEISTOCENE TERRACES OF WESTERN AMAZONIA Only a limited number of actual data is available on the soils of the terrains with Pleistocene influence inth e western part of Amazonia, the area west of Manaus.Indi cationsar etha t theprincipia l soilsar esimila rt othos eo feaster n Amazonia. However, thetexture ssee mt obe ,o nth ewhole ,heavie ran d thesan dfractio nfiner, a tleas ti nth e area between the rivers Solimoes and Madeira (cf. the textural data of MARBUT and MANIFOLD, 1926,an d Fig. 18fo r the area concerned).Onl y inth e area along the Rio Negroth etexture sar ebelieve dt ob erelativel y light.

Kaolinitic Yellow Latosol(,Ortho), rather heavy or heavytextured (KYLrA, KYL*) probably predominate. Rather heavy or heavy textured Kaolinitic Yellow Latosol, intergrade to Red YellowPodzolic soil (KYL-RP ) and Kaolinitic YellowLatosol, intergrade toGround WaterLaterite soil (KYL-GL, or PlanosolicLatosol; cf. III.2)- seem

') For southern Brasil such patches are not known. This may be because of nomadic living conditions of the former tribes in that region. The Amazon tribes, in contrast, are supposedt ohav ebee no fa mor esedentar y predisposition ort o havereturne d repeatedly to old dwelling sites. *)Othe r Ki and Kr data are to be expected for patches of TP in the non-Planicie part of Amazonia.

176 Folo18 Oterraco superiorda unidade deterra Candiru. Asuperficie de'plinthite''fossil 4 parcialmente enierradasob uma camadade argila de Bellerra remodelada, de uma espessura de 1 a2 melros. £notdvel o arranjovertical dos elementos concreciondrios (tipo Ipixuna)na parte superior da camadade 'plin thite'fossil, a altura domartelo. Tambe'm 4caraterislico ofim abrupto dacamada deargila de Bellerra. Adireila extrema vise aescarpa do terraco (BR-14, km 140m. ou m.)

^-v ** - £r'BllflHLiS l MMI

(* ' »\' . :- -V. .-.••• < \ J

/VHMO18 The upper terraceof theland-unit Candiru. The surface offossil plinthiteis for apart buried undera one to two metres thick layerof reworked Bellerra clay. Noteworthyis the vertical arrangement of theconcretionary elements (Ipixuna type) inthe uppermost part of thefossil plinthite layer, which is at the heightof the hammer. Also characteristic is the sudden end of the layer of Bellerra clay. The scarpof the terrace can beseen at thefar right (BR-14, km 140 ca.) tob eals orathe rfrequent ,th elatte rfo r instancea tTefe . It issuppose dtha t soilgrou p5 of MARBUTan d MANIFOLD(1926 )whic hconsist sof'ver yfine sand yloam swit hreddis h or reddish yellow friable clay subsoils'(cf Fig. 18), largely appliest o the three ment ioned soils.Als o welldevelope d Ground Water Laterite soil(GL ) isapparentl y rather frequent, especially betweenth eriver s Madeira and Puriis.I t is,fo i instance,th emai n soil of the extensive savannahs of Humaita - Labrea, as described by BRAUN and RAMOS(1959) .Ther eca n belittl edoub t that soil group 4o f MARBUTan d MANIFOLD, which consists of 'clays and clay loams with heavy, imperfectly oxidised clay soils usuallya t a deptho f twofee t or less' refers largelyt oGroun d Water Lateritesoils . Ground Water (Humus) Podzol(GP )i sals o present. It is,fo r instance,th esoi lo fth e caatinga amazonica ofth euppe r Rio Negro (VIEIRAan d FILHO 1961;cf. IV.1.2.2) . Concretionary soilssee mt o havea restricted occurrence,i ncontras t to thesituatio n ineaster nAmazoni a(cf MARBUTan d MANIFOLD,1926 ;p . 434).

177 III.3.5 TheSoil so fth eHolocen eterrain s

Ast oth e soilso fth e Holocene terrains in eastern Amazonia, many data are assem bledi nth estud yo f SUTMOLLER et ah (1964), and in the related report of SOMBROEK (1962b).Summarisin gth efollowin g canb esaid : On the varzeas, intermittendly poorly drained and non-peaty soils,classifie d as Low HumicGley soil(LHG ;Profil e 49)o r Humic Gleysoil (HG ; Profile 50) predominate. Also GroundWater Laterite-like soil is found. Well or moderately well-drained soil from recentalluvium ,Alluvial soil, seemst ob erestricte dt osmal larea sfa r upstream of themai nriver s(cf. aportio n ofmappin guni t Fo fAppendi x2) . Thesoil so fth eigapo sar epermanentl y poorlydrained ,ofte n peatyan d normallyver y acid.The y areBog soil, Half Bog soil an d Humic Gleysoil, intergrade toGround Water Podzol(cf. SOMBROEK,1962a) . The bottom lands within grass-covered upland, which physiographically are related to the igapos,wer efoun d to have HumicGley soil,Upland phase (HG, u; Profile 51). A sizeable area with this soil has been mapped near Braganga, as 'Glei Humico', by FILHOetal. (1963) . Alongth eOcea n coastan d ineaster n Marajo islandth elowlan d soilsar echaracteris edb ysalinit y and/or apredominanc e ofNa +o r Mg++a tth eadsorptio n complex.The y aregroupe d together asSaline and Alkali soils (fo r instance Solonetz, Coastal phase (Sol,c ; Profile52) . Themassap eterrain shav ewel l developed Ground WaterLaterite soil (GL ;Profil e 3). Also other Early Holocene terrains have such soil. In part, they intergrade to Low HumicGle yo r HumicGle ysoi l{cf. Profiles 11 and 13).

III.4 SoilAssociation s and Land-Units

The above described soils may occupy stretches large enough to be mapped on the scaleo f areconnaissanc e soilsurvey . Usually however, they occur together with other soils, with or without a specific pattern: soil associations and undifferentiated soil units. The soil association is a group of defined and named taxonomic soil units, regularly geographically associated ina define d proportional pattern.I fth e taxonomic units do not occur in a regular geographic association, or if they are not defined, the mappablegrou pi sa nundifferentiated soil unit (SOIL SURVEY MANUAL,1951) . A soil association or an undifferentiated unit may occur on one and the same topo graphicunit ,fo r instancea terrace .A nexampl ei sth eassociatio no f KaoliniticYello w Latosol (,Ortho), heavy textured (KYL^), Kaolinitic Yellow Latosol, intergrade to Dark Horizon Latosol, heavy textured (KYL-DHLA) and Kaolinitic Yellow Latosol, intergrade to Ground Water Laterite soil, heavy textured (KYL-GLA) whichoccur so nth eplanalt oeas to fPort oVelho . An example ofa nundifferentiate d unit on oneterrac e isth e occurrence of Kaolinitic YellowLatoso l (,Ortho),mediu mtexture d (KYLm),Kaoliniti c Yellow Latosol, Com-

178 Folo19 A basedo t err ago superior da unidade de terra Candiru. Ossediment os caoliniticos nao consoli- dadosque ocorrem por baixo da camada de 'plinthite' fossil, podem mostrar pinturas peculiares em cortes de estrada fundos. Isto 4 devidoaos diferentes coloridos das camadas sedimentdrias, provavelmente devidas a aguafredtica quepassa, rica em dxidosde/erro. Estaspartes vermelhasnao endurecem em poucosanos (BR-14, km 135 m. oum., ladeira dojacare")

jt.~

During the combined forest inventory soil survey of the Guama-Imperatriz area, it wasthough tdesirabl et ointroduc eth e'land-unit'. This is defined asa trac to fcountr y

179 Foto20 'Jabolis'. £stes monliculos sao carateristicos para muitos terrenos coberlos de argila de Belterrana area Guamd-Imperatriz. Osmonticulos saoprovavelmente osreslos de enormes termiteiros, construidos durantealguma e'poca do Pleistoceno. Morfomelricalmenteoperfil dosmonliculos nao difere notavelmentedo das partes planas dosterrenos (BR-I4, km 372m. ou m.)

-•? *»-.

Photo20 Jabotis. Thesehummocks are a characteristicfeature of many of the terrainswith Belterra clay cover in the Guamd-lmperatrizarea. The hummocks are presumably the relics of huge termite mounds,constructed during a timeof thePleistocene. Morphometrically, the soil profile of thehummocks isnot noticeably different from thatof the fiat parts of the terrains (BR-14, km 372ca.) within which the geologic, topographic and hydrographic elements, the climate, the pattern of vegetation, and the pattern of soils are approximately uniform. The des cription of such land-units can give, comprehensively, many data about an area hitherto completely unknown. An insight into a surveyed area, for the purpose of planning of forest management or agricultural settling, is likely to be obtained more readily from comprehensive descriptions of land-units than from the very technical data about soilsan dsoi lassociations .Thi si sparticularl y likelysince ,a tth ecommonl y applied levelo fclassification , thePlanici esoil sthemselve sar e similar inman yaspects . The differences in topographical characteristics, and others, are then very important. Future detailed soilsurvey s can also befacilitate d bytakin g the characteristics of the land-unit underconsideratio n intoaccount . A land-unit maycompris e only one soil.A nexampl e ofthi si sth e Kaolinitic Yellow Latosol (,Ortho), very heavy textured (KYL„ft), on non-attacked planalto terrain in the region south-east of Santarem. Usually however, a land-unit coincides with the area, or part of the area, of an association of soils. In that case the land-unit is not

180 Fig.21 Esboco dacomposicao daunidade de terraCandiru (paraa legenda, veja o Apendice 4 e Ta- bela9 dapdgina 126) cnlic RP-KYL.CR RP-KYL.CR SOUS:- RP-KYL.CR . KYL-RPr/p (B) KYl-RPrA RP-KYL^ . KYL,C„A

Fig.21 Sketch of theconstitution of the land-unit Candiru (see for legend Appendix4 and Table 9 of page126) identical with the association, but comprises many more characteristics of the area concerned. The three observed 'Associations' of the Caete-Marassume area (cf. Fig. 19 and DAY, 1959) are, in fact, also land-units1, as are the mapping units of the Araguaia Mahogany area (cf.Appendice s 2 and 6, and the detailed descriptions in SOMBROEK and SAMPAIO, 1962). In the Guama-Imperatriz area, eight land-units have been discerned: Santana, Medio Guamd, Candiru, Alto Guamd, Cunhanta, Gurupi-mirim, Planalto, Itingaan d Imperatriz(cf. Appendix4 ,an d SOMBROEK,1962a) . As to land-unit Candiru, see Fig.2 1an d thePhoto s 18, 19 and 20. A small portion ofth e areao fthi sland-uni ti sgive ni nAppendi x 3.

l) As real associations the Pitor6 and Maracassume ones are, as DAY himself states, practically identical. Their separation largely took place because of the difference in the composition of the vegetative cover.

181 IV TheSoil san dth e Vegetative Cover

INTRODUCTION Animportan t part ofth e Amazon soil studiesb y FAOtechnican sconcerne d therela tionships between the soilsan d the natural vegetative cover. In fact, the author's spe cific duty was to cooperate closely with the FAO/SPVEA Forest Inventory team. His task was to establish to what degree differences in vegetative cover, as encountered during the forest inventories, can be ascribed to differences in edaphic conditions. Such a study necessarily impliesth e recognition and evaluation of the influence of the non-edaphicfactor s climatean d man. The data collected onedaphi c and non-edaphic factors and those on vegetativecove r areincomplet e and onlycove ra smal lpercentag eo fth eenormou sare atha t constitutes Amazonia.Th edat a onth evegetativ ecove rar eals oo flimite dvalu e for thepurpos eo f the study of plant-soil relationships,sinc ei twa sno t possiblet o apply methods ofeco logicalresearc h proper. For a thorough study,a clear picture ofth e position of thevariou scomponent s (tree species, palms, climbers and creepers, epiphytes, shrubs) in the total structure of the tropical forest isa prerequisite .Th e relativecompetitio n strength ofthes e components directly determines the timber volume. Tree species themselves vary much in their reaction to conditions of micro climate,whethe r it be radiation, light, heat or humi dity. Some species need full sunlight, or associated conditions, during their entire life, someonl ywhe nthe yar eyoung .Other sd ono t need full sunlight,o rcanno t even stand it, during a part or the whole of their life. Commonly, two main groups of species are discerned: 'light-demanding' or 'intolerant' trees (upper-storey trees), and 'shade- bearing' or'tolerant'trees(under-store y trees).Th e average life cycle of species varies considerably.Ther ear esevera lnatura l causeso fdeca yan ddeath :ol dage ,smothering , direct or indirect windfall, sunburn, attack by (micro) flora or fauna. On the other hand, a number of species require a special soil micro flora (Mycorrhizae) for their development. The geographical distribution of a speciesdepend s not only on climate, man and soil conditions, but alsoo nth e length of timeelapse d since it first appeared somewhere in the hileia, and the rate of distribution. The latter is directly related to the mobility of the seeds. Possibly connected with this is the tendency of some species togro w inpatche so rcolonies ,o fvaryin gsize . Although HEINSDU^(1957, 1960) and PITT (1961) give many notes on the posi tion of tree species in the total structure of the Amazon forest units, systematic re searchi nthi s respect isstil llacking . Thefores t inventories weremean t onlyt o obtain, in ashor t time,genera l data on the

182 economic value of the primeval forest cover over large areas (cf. HEINSDIJK and MI RANDA BASTOS, 1963).It s sampling units therefore comprise only about 0.03% o f the forest covered land surface under consideration. Withina sample unit, only treeswer e considered, and onlythos ewit h a DBH (diametera t breast height)large rtha n 0.25 m. Only the merchantable bole, i.e.th e part of the stem below thefirst majo r branching, was considered in assessing the gross timber volume. Tree species with characteristic thin stemstherefore , andthos ewit hcharacteristi c lowbranching figure, wit h relatively low timber volume in the inventory tables. Stretches with swamp forest or secondary forest were not sampled. Areas with savannah were avoided. Likewise patches of savannah-forest and shrub-forest were not sampled, as it was assumed that such pat cheswoul d notcontai n treeswit h DBH largertha n 0.25m . Invie wo fth enea rabsenc eo fpublishe ddat a onplant-soi lrelationship si nAmazonia , and the windingu po fth e FAO/SPVEA forest inventory- soilsurve yprogramme ,th e conclusions of this programme are discussed in the present chapter, though inman y caseson ewoul d have preferred to await more definite and adequate data.

IV.1 TheUpland swit hFores tCove r

IV.1. 1 Evaluation of Non-Edaphic Factors

IV.1.1.1 TheInfluence of the Climate It islikel y that the difference in climate betweenth e northwestern part of Amazonia and the central and eastern parts (cf. Fig.2 )i sresponsibl e for some of the differences between the general vegetative covers of the uplands in those parts. This is suggested by the phytogeographic data of DUCKE and BLACK (1954) as related in 1.5. Thesedif ferences apply not only to the timber volume and the occurrence of tree species, but alsot o the composition of the undergrowth. For instance, it wasobserve d that within the 'pocket' area with Af climate around Belem(cf. Fig. 2),epiphyte s are rather fre quent on thetre e stemsi n the upland forests. In regionsoutsid e this pocket, with Am climate,epiphyte s are found in considerable quantities in the patches of forest on the permanentlywaterlogge d igapoparts ,bu tar epracticall y absent inth eforest so fth eup lands. It issuppose d that the absence ofepiphyte s in the latter isdu e to a drop in the relative humidity inth e tree layer during the dry season. The plantsca n only flourish >ith e igapo parts, where the waterlogging of the soil prevents such a drop. The ab sence or presence of epiphytes in upland forests may give an indication of the local climate in areaswher eweathe r recording stations are absent.

TIMBER VOLUME Higher annual amounts of rain and a more regular distribution of the rainfall over theyea r(Figs .3 an d 4)apparentl y doesnot impl y a higher mean grosstimbe r volume, edaphic conditions being constant. The FAO/SPVEA forest inventories apply to the Planicie for a large part, and the freely draining soilso f this section of Amazonia are

183 similar to a large degree (cf. III.3.4). As Fig. 22 shows, the highest timber volumes (inventory units Planalto I, Planalto II, Planalto II Baixo, Portel, and especially Caxiuana) arefoun d ina stretc hwit habou t 2000m mrainfal l and afairl y well defined dry season. In awesterl y directionth e mean grosstimbe r volume decreases, although the rainfall is higher and also more regularly distributed over the year. In the area where the Amazon enters Brazil,th e timber volume on the uplands isstil l lower(10 0 m3/ha ca.;unpublishe d data of FAO/SPVEA Mission about the Benjamin Constant area), although total rainfall is nearly 3000m m and a dry season is absent. The des cription by DUCKE and BLACK (1954) of their 'Norte' region (cf. Fig. 13), much of whichi slocate d ina sectiono fth ePlanfci e with Af climate, also pointst o a relatively lowtimbe r volume (for instance:'les shig htree stha n elsewherei nth ehileia') . Goingfro m thementione d stretchwit hmaxima l mean grosstimbe rvolum ei neaster ly and south-easterly directions, the timber wealth begins to decrease substantially only inth etransitio n zone towardsth esavannah so fNorth-Easter nBrazil .I nthi szon e the dry season becomes fairly pronounced (4 months with less than 50 mm rainfall each, although total annual rainfall isstill , probably, about 2000mm) .Thi s decrease in timber wealth was studied in the Guama-Imperatriz area. The fact that the timber volumei nth esouther n part ofthi sare a(inventor y unitAâilandia )i sver ylo wi ncom parison with that of areas north and northwest, must be largely due to a different distribution ofth erainfal l overth eyear .Thi sis ,becaus eth esoil sar esimila r oridenti cal, and planalto parts with Kaolinitic Yellow Latosol, very heavy textured (KYL„A) inth eare a ofth einventor y unitAâilandi ahav ea distinctl ylowe rtimbe rvolum etha n planalto parts with the same soil in the area of the inventory unit Ligaâo. That the timber volume in this region decreases with increasing seasonal drought is also illustrated by the fact that in the southern half of the Guama-Imperatriz area steep slopesfacin g theeast ,viz. thedirectio n ofth erains ,generall yhav ea bette rfores t cover than those facing thewest .Th e latter often bear a lowvegetatio n largely consistingo f shrubs. This phenomenon has not been observed in the areas of the inventory units nearert oth e Amazon. Alsoth e timber volume situation inth e inventory units Piria, Gurupi, and Maracas sume reflects the influence of an increasingly pronounced dry season. Although the areaso fth eunit sPiri aan d Maracassumehav eth esam emai nwel ldraine d soil,namel y the Kaolinitic Yellow Latosol,mediu m textured (KYLm) of the Associations Pitor6 and Maracassume respectively (cf. Fig. 19),ther e isa substantial difference in timber volumebetwee nbot hunits. 1 As will be discussed below (IV.1.2) , many differences in timber volume between the

x) DAY (1959) suggests that a larger extraction of timber throughout the years in the Maracassume unit may account for the difference. GLERUM (1960) however states that this extraction concernsonl yth e Cedro(Cedrela odorata) and theAndiroba (Carapdguianensis). The first species'contributes generally, and also in the Piria unit, hardly anything to the gross timber volume of the Amazon forests. The occurrence of the second species is only slightly higher in the PiriS unit than in the Maracassume unit (6.6 m'/ha and 4.3 m"/ha respectively).

184 Fig. 22 Invenidrios Florestais e Reservas Florestais. Extrato de MIRANDA BASTOS (1958), HEINSDIJK (1957,1958a, 1958b, 1958c), GLERUM (1960,1962), GLERUM and SMIT (1962a, 1962b) e HEINSDIJK and MIRANDA BASTOS (1963)

Forest inventory units, with mean gross timber volume in m'/ha Unidadesde inventdrio florestal, comvolume medio de madeira bruta em m'lha Established or proposed forest (production) reserves, andsilvicultur e centres Reservas florestais (de renda). e centros de si/vicultura

FOREST INVENTORY UNITS/Unidades de Inventdrio Florestal 0. Amapari-Matapi-Cupixi; 1.Capim ; 2.Acara ; 3.Belem-Sul ; 4.Cameta-oueste ; 5.Portel ; 6Caxi - uana; 7. Flanco II; 8. Planalto IIbaix o cipoal; 9.Planalt o IIbaixo ; 10. Planalto II; 11. FlancoI ; 12. Planalto I; 13. Arapiuns; 14. Maues; 15. Canhuma; 16. Piria; 17. Gurupi; 18.Maracassume ; 19. Santana; 20. Candini; 21. Medio Guama; 22. Alto Guama; 23. Ligagao; 24. A?ailandia; 25. To- cantins (Ucuuba); 26. Araguaia (Mogno)

Fig. 22 Forest Inventories andForest Reserves. Extracted from MIRANDA BASTOS (1958), HEINSDIJK (1957,1958a, 1958b, 1958c), GLERUM (1960,1962), GLERUM and SMIT (1962a, 1962b) and HEINSDIJK and MIRANDA BASTOS (1963) various inventory units may beascribe d todifference s insoi l qualities. Nevertheless, the above comparison with climatical conditions gives a tentative conclusion that highest timber volumes of the Planicie may befoun d inthos e areas where the total annual rainfall isno texcessivel y high anda shor t relatively dryseaso n exists. Only whenthi s dry season becomes pronounced (four or more months with lesstha n 50m m rainfall each) itma y have anegativ e effect on the growth ofth e forest.

OCCURRENCE OF INDIVIDUAL TREE SPECIES A number of the differences established inoccurrenc e oftre e species are also likelyt o bedu et o differences inclimate , inparticula r asregard s the amount and distribution of

185 rainfall. Comparison of the species composition of the unit Acailandia, which has a fairly pronounced dry season, with those of inventory units located north and north west of it confirms this impression (cf. GLERUM and SMIT, 1962a). Many species found inon eo r moreo fth e latter unitsar eabsent , or sparsely present, in theA

IV.1.1.2 The Influence ofMan Although Amazonia as a whole is very sparsely populated, it is possible to discern several traces of anthroprogenic influence on thecharacteristic s of its forests.

SECONDARY FORESTS The influence of man is of course very clearly visible in the areas with secondary forests (for their location cf. Fig. 12).Youn g secondary forest (capocira)ca n berecog nized easily on aerial photographs, as well as in the field. The first is often difficult with veryol d secondary forest (capoeirao), but for thetraine d eyeth edifference s from primeval forest are clearly visible in thefield. DUCKE and BLACK (1954) give data on the differences in composition of the secondary forests, whether regenerating after felling only,o rafte r felling and burning.

186 SELECTIVE CUTTING Also seemingly primeval forest may have undergone anthropogenic influence, for instance by selective cutting of certain very valuable tree species. This applies to the Mafaranduba (Manilkerahuberi), a tree often present in largequantities , which iscu t for its latex (HEINSDUK, 1958a). Forests are often penetrated very deeply for cutting the rarePau rosa (Aniba roseodora), which renders oil used in the perfume indus try (HEINSDUK, 1958C). The same applies to the rare Cedro (Cedrela odorata),whic h hasver yvaluabl e timber (HEINSDUK, 1957;GLERUM, 1960). HEINSDUK (1960) notes, in agreement with observations elsewhere (Surinam), that forests in regions which have long been under man's influence (occasional cutting of trees, large areas of secondary forest nearby) often have more trees of certain species than forests not influenced by man, but with otherwise similar growing conditions. Thisapplie st o thespecie s Cupiuba (Coupiaglabra) and the Quarubas (Vochysiaceae), and isth ecas ewit h the inventory units Belem-sul and Cameta-ouest, both 'anthropo- genefades *o fth e Pouteria association(

INDIANS Theorigina l inhabitants ofAmazonia , theAmer-Indians ,ar e likely tohav eexerte d a considerable influence upon the Amazonian forests. These people were, apparently, far from insignificant in number in pre-Columbian times. They used to dwell, in sizeablecommunities ,alon g very many of the navigable riversan d rivulets(cf. discus sionso n 'Terra Preta' in 1II.3.4).Indication so f their influence are present for instance in the area south of Santarem, where many Indian settlements are known to have existed (cf. Fig. 20). All the forest in this area was mapped, after aerial photograph analysis, as 'cipoalic forest east of Belterra' thus distinguishing it from the real (?) primeval forest east of the area (HEINSDUK, 1957). Thevisua l impression of the forest partso n the Belterraestat e itself isdistinctl y different from that ofth efores t at Curua- una centre, although both are growing on identical soil namely Kaolinitic Yellow Latosol, very heavytexture d (KYLrji). Pieceso fcharcoa l inth esoi l mightals o indicate former Indian influence. In thiscon nection it should be mentioned that in all of three profile pits dug under Planalto II forest at the Curua-una centre (cf. the numbers 112, 303 and 304 of Fig. 11) there ^ere found pieces of charcoal. These pieces mostly occurred at a depth of about 150 cm,an d weresurrounde d bybake d clay.On ema ydeduc e that extensiveburnin go fth e forest took place in former days. It is unlikely that lightning is the cause of this. It is true that burning of tap roots after a tree isstruc k by lightning sometimes occurs, but suchfire seat swoul d not spread overa nextensiv eare a(i ncontras t toth econdition so n the lower, sandy terrains, with Flanco I forest, intentionally started fires of artificial clearingso nthi s planalto stretch do not penetrate into adjacent forest parts). Human interference is therefore likely. This must have taken place long ago, because the Present day planalto forest has one of the highest mean gross timber volumes known for Amazonia and has hitherto been considered as real primeval. Large-scale burning bya former Indian population istherefor e likely to have occurred on the planalto of

187 Curua-una centre. It may benote d that a smallpatch of Terra Preta, which iscertai n indicator offorme r Indian settling, isfoun d at the riverside (point Ao fth e crosssect ioni n Fig. 11).

TABOCAL A peculiar example of human influence on the forest cover became apparent during thesurve y of the Guama-Imperatriz area.A sreporte d in1.5.1.2 ,a special forest type, called tabocal, isfoun d inth eheadwate r region ofth eGurupi ,wher ei toccupie s broad bands along the river and itstributarie s (cf. Fig. 23).Th e composition and the gross timber volume of the forests north, west and south of the tabocal-covered area, as wella so fth epatche so ffores t withinthi stabocal ,ar eapproximatel y uniform. Southo f theSerr a deGurup i morepalm sar efoun d than inth eothe rpart swit hforest , but allo f

Fig. 23 Vegetacao nas cabeceiras do rio Gurupi. De GLERUM and SMIT (1962a)

High forest Mata alia

Shrubby creeper forest with Guadua spp. Tabocal

Fig. 23 Vegetation in the headwater region of the Gurupi river. From GLERUM and SMIT (1962a)

188 it belongst o the inventory unit Acailandia.Th e soils ofth e tabocal-covered parts are neither physically nor chemically different from those under forest cover {cf. data of SOMBROEK,1963a) .Apparently ,th eGuadua specie so fth etaboca l dieafte r two or three years.Durin g the pronounced dry seasonth edea dstake sar ever ysusceptabl et o burn ing. It was observed that,durin gtha tseaso ni fburningi sinitiate da ton espot ,th e fire sweepsver yrapidl y over enormous areas of tabocal, also consuming considerable partso f adjoining, relativelypoo ran ddry ,forest . The Guadua species have an exten sivefire, resisten troo t stock,an d thetaboca l vegetation therefore canre-establis h itself fully after burning. The burned forest parts,however , do not regenerate asreadil y under the existingclimati cconditions .Th eresul t isa gradua lenlargemen t of theare a coveredwit htabocal . It is supposed that anthropogenic factors have stimulated, if not caused, the occur rence of tabocal vegetation in the region. The facts that have lead to this supposition are: the geographic distribution of tabocal; the uniformity of the adjoining forested parts; the equality of the soils both under tabocal and under forest; the behaviour of the vegetation under influence of burning, and the reported existence of considerable former, andlocall ypresent-day ,settlin go fIndian salon gth eheadwater so fth eGurupi . Indians have always used bamboo stakes, for arrows and other purposes. Probably the Guaduaspecie swer ealread ya constituen t ofth eprimeva l vegetation, but their pre sent occurrence asth e dominant speciesove r a large area isbelieve d to be due to the burningpractice so f Indians.Th etaboca ltherefor e isa 'fire-subclimax'.

IV.1. 2 Soilsan d Forest Characteristics

Nearly allfores t inventory areasar elocate d inth ePlanicie .I tha salread ybee nshow n in II.2.2 and III.3.4 that the freely draining soils of the Planicie are similar to a large degree, at the applied level of classification. Any differences in forest characteristics within an area of uniform climatic conditions and non-existent or uniform anthropogene influences, are therefore apt to show a correlation with lower category soil dif ferences. Among these are the moisture holding capacity, thetota l available amounts of the various plant nutrients, and the penetration possibilities for roots.Thes equali tiesdepen d on such factors asth e soiltexture ,th ecompactnes so f thesubsoil , and the presence of plinthitic materials. The coinciding of differences in forest characteristics with differences inthes e latterfactor s are discussed below.Th e exact causal factors of each ofth eestablishe dcoincidence sremain softe n uncertain. Among the differentiating forest characteristics, the gross timber volume and the occurrenceo findividua ltre especie sar econsidere d in particular.

IV.1.2. 1 Land-Units and Forest Inventory Units As related in 1.5.1.2,th e geographical boundaries of FAO/SPVEA forest inventory units wereno t solely established on the differentiating characteristics ofth e forests it self,bu t to adegre eals o on thetopograph y and the soils.O n the other hand, thegeo -

189 Table11 Forest inventory unitsand land-units inthe Cuamd-lmperatriz area

Forest Number Mean gross Salient tree Peculiari Land- Main soils inventory of trees timber volume : species ties unit (<•/.Tabl e9 ) unit perh a perh a unidadede numerode volume medio especies peculiari- unidade solosprincipals inventario drvores demadeira dedrvores dades deterra (cf.Tabela9) florestal porha brutaporha satienies

Santana 106 152.5 Acapu (Voua- Locally Santana KYLm capoua ameri- anthropo RP-KYL, CR cana) genic in (Mae do Rio) fluences influencias antropoge- nicasem al- gunslocal's

Medio Guama 108 161.2 Medio KYLm Guama KYL, Crft RP-KYL, CR (Ipixuna)

Candiru 124 191.6 Pau amarelo Candiru KYL-RPM (Euxylophora RP-KYL, CR paraensis) (Ipixuna) Pau roxo (Pellogyne lecointei) Quaruba (Vochysia maxima) graphical boundaries of land-units were established, in part, on the local pattern of vegetation(cf. definition ofland-uni ti nIII.4) .Therefore , acompariso n betweeninven tory unit and land-unit always gives a correspondence. For the Guama-Imperatriz area, the geographical boundaries of the inventory units coincide completely with those ofth eland-units .A compariso n between thefores t data and the soildat a of this areai s given,summarised ,i nTabl e 11. The land-units under consideration are, for the most part, concerned with associa tions of two or three main soils.I t can therefore be said that a degree of relationship between the forest growth and the soils has been established for the Guama-Impera triz area. This istrue ,eve nthoug h the over-all differences goingfro m North to South mayb eascribe d largelyt oa differenc e in climate,a spreviousl ydiscussed .Tha tfo r this area no relationship whatsoever has been established between a certain identified soil and oneo r morecharacteristic s ofth efores t growingo nit ,i swithou tdoub t duet o the systemo ffores t inventory.I fth etota lstructur eo fth evegetativ ecove rha d beenstudie d

190 Table 11 continued/Tabela 11 continuada

Forest Number Mean gross Salient tree Peculiari Land- Main soils inventory of trees timber volume species ties unit (cf. Table9 ) unit perh a per ha unidadede numerode volumemedio especies peculiari- unidade solosprincipal's inventdrio drvores demadeira dedrvores dades deterra (cf.Tabela9) florestal porha brutaperha salientes

Alto Guama 94 121.1 Cedro Many Alto K.YL, cVh (Cedrela 'cipoalic' Guama RP-KYL,* odorata) parts RP-KYL, CR Breu preto muitaspar (Paragominas) (Protium tes cipod opacum) licas

Ligacao 101 138.0 Angelim Pedra Locally Cunhanta KYLm (Hymeno- 'cipoalic' RP-KYLrft lobium parts RP-KYL, CR excelsum) partes (Campinho) cipodlicas Planalto KYL * em alguns t locais Gurupi- RP-KYLr» mirim

Agailandia 51 61.6 Cedro Locally Itinga KYL» (Cedrela 'cipoalic' KYLt* odorata) parts KRLm Jutai-agu partesci- (Hymenaea podlicasem , trf2 ; iCRLm courbaril) algunslocals RM Itauba 'Tabocal' (Mezilaurus parts fre itauba) quently tabocais frequentes

Tabela 11 Unidades de inventdrioflorestal e unidades de terrana area Guamd-Imperatriz

(apractica l impossibility), and the data from individual, or parts ofindividual ,inven tory sampling units had been compared with the soil in loco,the n certainly several relationshipscoul dhav ebee nshown .

IV.1.2. 2 Soilsand Timber Volume The gross timber volume of a forest willbecom e of special interest once it isecono mically possiblefo r the pulp and paper industry to utiliseal l or a great number ofth e tree species from, the highly mixed tropical forest. When forested areas are being considered for agricultural settlement, thestandin gtota l timber isals o of importance; it normally determines a good deal of the amount of organic matter that becomes in-

191 corporated into the soil at the time of clearing. The timber volume may also be handled as a guide for the productivity of a soil. For these reasons it is not only of purely scientific, but also of practical value,t o consider the relationship between soil andgros stimbe rvolume ,i fany .

SOIL TEXTURE Within the northern part of the Guama-Imperatriz area, the soil with the highest mean grosstimbe r volume isth e deeply friable, porous and well-rooted, rather heavy texturedsoi lwhic hi sclassifie d asKaoliniti cYello wLatosol ,intergrad et o RedYello w Podzolicsoi l(KYL-RP rft).Thi si sa mai nsoi lo fth eland-uni tCandiru ,whic hcoincide s with the forest inventory unit Candiru. The grosstimbe r volume of the forest on this soili sestimate d tob e20 0t o25 0m 3/ha,agains t 100t o20 0m 3/hafo r the other main soil occurringi nth eare ao fth elan d unit: RP-KYL, CR(Ipixuna) . The variation of timber volume with the texture of soils otherwise approxima tely similar was studied fairly closely by HEINSDIJK (1957) in the western half of the Tapajos-Xingu inventoryarea .Th econdition so fclimat ean danthropogeni c influence are approximately uniform in this area. The differences in soil texture are very out standing and consistent. The soils are namely either Kaolinitic Yellow Latosol, very heavytexture d (KYL„A), or Kaolinitic LatosolicSan d(KLS) .Th eforme r isth esoi lo f the planalto parts and the latter the soil of practically all the terrain below this level (flanco parts). The boundaries of the forest inventory units in the area coincide with theboundarie s ofth etw omentione d soilunits .O nth eplanalt o ofthi sarea ,th eoccur rence ofcipoa l (a distorting factor inth e comparison, seebelow )i slimited . HEINSDIJK (1957) established that the inventory units on the very heavy textured soil have adis tinctlyhighe rmea ngros stimbe rvolum etha nth euni to nth eligh ttexture dsoi l(Planalt o I unitwit h22 7m 3/ha.an d Planalto II unit with 233 m3/ha against the Flanco I unit with 132 m3/ha). 3 The very timber-rich Caxiuana unit (271 m /ha) occurs, according to HEINSDIJK (1958a), on a 'transition from a planalto to a plain', the highest parts of which are about 20m abov e local river level.Ther e are somepart s with savannah or savannah- forest coverage in the watershed region; these are most probably sandy (cf. IV.2.2). Theothe r parts,however ,especiall ythos ea tth e eastern side,accordin gt oindications , areal lsimila r and rather heavyt o heavytextured .Actua l soildat a havebee ngathere d from thenortheaster n part, whereth e terrains are practicallyflat an d 5t o 10 mabov e local river level. The soil in question is a heavy textured Kaolinitic Yellow Latosol (KYLA). In discussing all forest inventory units 1 to 15together , HEINSDIJK (1960) mentions as a general tendency, that heavier textured soils have a higher gross timber volume. The basis for this conclusion are his ownfield observation s concerning soil textures (the units 1, 4, 7, 11an d 13o f Fig.2 2 are believed to be located on 'sandy to pure sand soils', and the units 6, 8,9 , 10an d 12o n 'definitely heavy soils'). The above data enable us to draw the tentative conclusion that for the Planicie a relation exists between soil texture and gross timber volume. The rather heavy and

192 Foto21 O aspetoda flaresta alia nossolos da Planicie, defdcil penetracao das raizese relativamente boa armazenagem de umidade.Para o dossel sobem os troncoslisos de todas asespessuras efallam quasepor completo oscipds (BR-14, km 130 m. oum.)

Photo21 Theaspect of thehigh forest on thePlanicie soilswhich haveeasy penetrability for roots and comparativelygood moisture storage.Slender boles of allsizes tower towards the canopy, andcreepers andclimbers are practically absent (BR-14, km 130ca.)

193 heavy textured soils(35-70 % < 2micro n in the Bhorizon ) in particular, lend them selvest o support forest with a highgros stimbe r volume.A comparativel y high moist ure holdingcapacit y of thesesoil s isprobabl y a major reason(cf. V.3.1.1).

SUBSOILCOMPACTNES SAN DCIPOA L That the areas with very heavy textured freely draining soilsten d to have a lesshig h timber volumetha n areas with similar soilsbu t of slightly lessheav ytexture ,i srelate d to an often higher compactness ofth e subsoil of the Belterra clay soilsan d the occur rence of cipoal and 'cipoalic forest' (cf.th e descriptions in 1.5.1.2). Becauseo f there gularity in occurrence and the often large expanse of the vegetative types concerned, the possibility that a normal regeneration phase ofth e adjoining high forest isinvolv ed, can bediscarded . HEINSDIJK (1957)suppose sth e cipoalt o bea 'para-climax ' vege tation. The occurrence of these vegetation types was studied in some detail in the Guama- Imperatriz area. The completely flat central sections of planalto parts in the stretch km 190-km31 5(uni tLigagao ;cf. Phot o 22)ar ecovere dwit hcipoa lo r'cipoali c forest'. The very gently sloping edges, above the escarpment proper, bear however a heavy forest cover.Thi sconsist so fbi gtree san da nope nundergrowt h whichi snearl y devoid ofcreeper san d climbers.A simila r difference invegetativ e coverso f the edgesan d the central sections occurs on the southernmost planalto parts (stretch km 315-km 420, forest inventory unitA$ailandia )althoug hth etimbe rvolum ea sa whol ei smuc hlowe r there.Th esoi lo fbot hth eedge san dth ecentra l sectionsi sth esame ,namel y Kaolinitic Yellow Latosol, very heavytexture d (KYL„ft). Many planalto stretches were studied, and analyticaldat a ofsevera lprofile scompared . Phycicalan dchemica lqualitie so fth e profiles areabou t identicalfo r bothsites .On edetectabl e difference ishoweve rtha t the central sections may have a conspicuously dark (dark grey to black), crusted, often irregular and clodded, aquaphobe surface, with an intense activity of termites. The pH-H20 ofthi ssurfac e (field testing) isinvariabl y slightly higher (about 6.0instea d of 4.0-5.0) than that of anon-cruste d surface layer,a si scommo n onth e planaltoedges . Laboratory analysis of a few dark surface clods revealed a stronger predominance of Ca++ inth e adsorption complex than iscommo n for the Planicie soils.Anothe r small difference between the profiles on the edges and those on the central sections of the planalto is that, on the former, the At subhorizon is often not quite as thin (3-5 cm against 2cm) .Th ecentra l sections have also a slightly more compact subsoil (Bhori zon), which manifests itself ina friable tofirm consistenc e in contrast withth e friable to veryfriabl e consistence of the subsoil at the edges,an d a larger resistance to pene tration witha soi lhammer . 'Cipoalicforest ' iso fgenera l occurrence on theterrain swit h reworked Belterra clay inth estretc h km 150-km 195 ofth e Guama-Imperatrizarea .Th esoi lo fthes e terrains isclassifie d as Kaolinitic Yellow Latosol, Compact phase,ver yheav ytexture d (K.YL, Cp»).I tha scharacteristicall ya ver ythi nA t subhorizon (0.5-3 cm),an d iscompac t and ratherfirm, particularl y in the Bhorizon . Thefrequenc y offalle n trees on the terrains seemst ob elarge rtha nelsewhere .

194 Foto22 Floresta cipodlica numa parte centralinteiramente plana deurn trechode planalto. A ambosos ladosdesta estrada de floresta recentemente aberta, vise claramente afreqiienciade cipos quetrepam as copasdas drvores (BR-14, km 295m. ou m.)

Photo22 'Cipoalic forest' ona completely flat central section of a stretch of planalto. Onboth sides of this recently openedforest road, thefrequency of climbersup to the tree crowns is clearlyvisible (BR-14, km 295 ca.)

The cipoal on the planalto and on the terrains with re-worked Belterra clay in the regionbetwee n theriver sXingi ian dTapaj6 s(cf. Fig.24 )hav ebee nstudie d by HEINS- DUK (1957). He observed that under the cipoal the soil surface is much darker than underth e highforest . Theloca lfarmer s considerth e cipoal parts preferable for shift ingcultivation .Th eautho r wasunabl et o visitthes earea swit hcipoa l proper, but did studyth eloca lchange si nvegetativ ecove ran d soilqualitie so nth eplanalt oa t Curua- una centre (cf. Fig. 11). One examined soil profile is located on the edge of planalto (No. 112),wher ever ytimber-ric h forest isfound , witha n open undergrowth.A t8k m distancefro m thisedge ,anothe r profile (No.303 ) was examined.Th e forest there is slightly different in composition from that of the edge (cf. GLERUM and SMIT, 1960) andth etimbe rvolum ei sprobabl yslightl ylower .A thir dexamine dprofil e islocate da t 10k mfro m thementione dedg e(No .304) . Thisi sa nare awher eth etimbe rvolum ei s distinctly loweran d thecompositio no fth efores t quitedifferen t from that atth eothe r sites. Creepers and climbersar e rather frequent. Many trees with stilt-roots occur, for instanceth eAndiroba (Carapa guianensis) whichi softe n consideredt ob ea nigap o tree. Also,severa l tall herbs commonly found in igapopart sar e present, for instance Guarumd (Ischinosiphonaruma). Th efrequenc y offalle n treesi s remarkable.

195 Thesoi lprofil e atth eedg eha sa ver yloos eappearanc ei nit ssubsoi l (B horizon).Th e second profile is already somewhat dense, and the most interior one is considerably compact andresistan tt openetratio n with a soil hammer. While the colour of the Ax subhorizonsi sth esam e(brow nt odar k brown)i nal ltre eprofiles ,it sthicknes sdecreas esfro m 30c mt o2 0c man dt o 10c mrespectively .

From the above described comparative observations it istentativel y concluded that cipoal and 'cipoalic forest' grow on those very heavy textured soils that have a com pact subsoil, and athi n Axsubhorizon , in comparison to soilso fth eimmediatel y sur roundings. Owing to the compactness, the maximum depth of rooting issmaller . The soili stherefore , asregard sth erequirement so fth efores t vegetation, ashallo w one,o n whichhig htree scanno t compete with the apparently vigorously growingcreeper san d climbers.Probabl ybot ha lack of ground support for thetree s (fallen trees, stilt-roots) and a limited moisture supply during the dry season (cf. V.3.1.1)ar e the causes of the smallcompetitio n force ofth etrees .Th elocall yfoun d intensetermit eactivity , crustin ess, higher pH and dark colour ofth e superficial layer isbelieve d to bea consequence rather thana caus eo fth ecipoa lgrowth . Therear ea fe wfact stha t support theabov etentativ econclusio n ast oth ecaus eo fth e cipoal and the 'cipoalic forest' onth e Belterra clayterrains . In the Guama-Imperatriz area, othertype so fcreepe r forest (ingenera l much lessvigorou stha n the Belterracla y cipoal) and dense shrub vegetation are often associated with Ground Water Laterite soil (GL) and Kaolinitic Yellow Latosol, Compact phase, rather heavy textured (K.YL,c rft). Both soils have shallow root penetration, because they haveeithe r an in termittently shallowgroun d waterleve lo r acompac t subsoil.Fo r the Araguaia Maho gany area, which is outside the Planicie, it has been established that the soils with cipoal and 'cipoalic forest' are ofvaryin g natural fertility, but allo f them are shallow: shallowness of the bedrock or of the ground water level. The soils in this area which have acoverag e ofclosed-canopy-fores t are also of varying fertility (it is,i n fact, usu ally low), but are deep (cf. Appendix 6). Apparently, the occurrence of cipoal or 'cipoalic forest' as opposed to closed-canopy-forest in this area is lessdetermine d by the chemical than by the physical qualities of the soils. Considering the whole of Amazonia, it may be a general rule that creeper forests prevail on soils with shallow rooting,whethe r havinglo wo rhig hnatura l fertility. It iseviden t that it iso f much practical importance to study further the factors that determine the occurrence of forest types with very low timber volume such as the cipoal.A prerequisit e for further study isa sound plant-ecological classification ofth e various types of cipoal and 'cipoalic forest'. Of special interest isth e checking of the tentative conclusion as to the cause of the Belterra clay cipoal, since silvicultural me thods are being tried out predominantly on planalto terrains (FAO/SPVEA silvicul tureprogramm ea tCurua-un acentre) .

The foregoing discussions have concerned the timber volume on the Planicie soils that are freely draining. Generally speaking, the timber volume of the forests on the

196 Fig.24 Cipoal na regiao aoLesie do baixo rio Xingu. De HEINSDIJK (1957)

52°30' 52°15' W

High forest Escarpment Florestaalia EK Escarpa Creeper forest Rivulets and rivers Cipoal Igarape's e rios Fig.24 Creepervegetation inthe region east of thelower Xingu river. From HEINSDIJK (1957)

imperfectly drained Planicie soils(Groun d Water Laterite soil, Ground Water Pod- zol) is considerably lower. Large stretches with these soils even have a coverage of savannah-forest (with DBH of the trees supposedly all smallei than 0.25 m) or of savannah.I t isstipulate d thatth esituatio n isdifferen t for thefreel y drainingsoil swit h fossil plinthite,thoug h they aregeneticall y related toth e Ground Water Lateritesoils . The soils in question, classified as Kaolinitic Yellow Latosol, Concretionary phase (K.YL, CR) or Red Yellow Podzolic soil, intergrading to Kaolinitic Yellow Latosol, Concretionary phase (RP-KYL, CR), have on the average a notlowe r timber volume thanadjoinin g non-plinthiticfreel y drainingsoils ,excep ti nth erar ecase swhe npractic allyn oeart h is presentbetwee nth econcretions . Considerable differences in grosstimbe r volume may occur onth e various soils out side the Planicie part of Amazonia. This is already evident from qualitative observa tionsi nth eAraguai a Mahogany area(cf. GLERUMan d SMIT(1962b )an d Appendix6) .

197 Measurements of mean gross timber volume were however not executed for those areas.

IV.1.2. 3 Soilsand Occurrence of Individual Tree Species Severaltre e speciesar efoun d only ina part ofth e hileia-covered area. In theintro duction toth epresen t chapter itwa sstate d that thestud y ofth e causes ofestablishe d differences inth eoccurrenc e ofa certain tree species isa complex one. Some ofth e differences arebelieve d tob erelate d toth enon-edaphi c factors climate andma n (cf. IV.1.1) .T osa ytha t the othersar erelate d with soildifference s ishoweve r anoversim plification. Atypica l controversial species isth e Acapu (Vouacapoua americana). This tree,renderin gmuch-use d timber, isa conspicuou s under-storey tree.I toccur s usually incolonies ,o fvaryin g size.Average d overlarge r areasth e speciesca n occupy as much as 10%o fth e total number oftree s{cf. HEINSDIJK, 1957, 1960).O nth e south side of theAmazo n riverth etre ei sfoun d onlyeas t ofth etributar y Curua-una; itform s there one of the distinguishing features between the Planalto I and Planalto II inventory units.It seaster n boundary ofoccurrenc e isforme d byth emiddl ereache so fth eriver s Capiman d Guama,wher eth especie sform s oneo fth edistinguishin gfeature s between the inventory units Capim andAcara , andCandir u andSantan a respectively. At the North side of the Amazon river thetre e is found, according to DUCKE and BLACK (1954),onl yeas to fth elarg etributar yTrombetas .Th especie si sals ofoun d inSurina m andther e known asBruinhart. SCHULZ(1960 )studie d somesoil si nth e Mapane region of this country, where the tree is found in some sharply delimited stands and then occupies 1/6-1/3o fth e total canopy. Several analytical data ofthre e profiles takeni n such stands were, when averaged, slightly different from those ofprofile s in adjacent non-Bruinhartcontainin gfores t (the soil onbot h sitesi sre dcoloured , kaolinitic,an d predominantly mediumtextured) . SCHULZdeduce dfro mthes esoi ldat atha tth euneve n occurrence ofth especie si nth e Mapane region isdetermine d bysoi l factors. For the Curua-una - Xingu region of Amazonia, HEINSDIJK (1957) noted that the speciesi spresen t inlarg enumber so nth eplanalto ,o nth eclayis h parts ofth eslope so f the planalto, ando nclaye y terrains near the rivers. Kaolinitic Yellow Latosol, heavy or very heavytexture d (KYLA, KYL„A) are involved, which soilsar eals o found west of the Curua-una, on identical topographical units. HEINSDIJK states also that 'the colonieso fAcapi ineve rappea ro nsand ysoils' . Theautho rstudie dth egrowin gsit eo fth especie si nth eare ao fth eSantan a unito fth e Guama-Imperatriz area. Colonies of the tree were found to occur here, equally fre quently,o nal lth esoil so fth earea ,fo rinstanc eo nth ever ycommo n KaoliniticYello w Latosol, medium textured (KYLm), the few patches of Kaolinitic Latosolic Sand (KLS), the scattered Red Yellow Podzolic soil, intergrading to Kaolinitic Yellow Latosol, Concretionary phase (RP-KYL, CR) whether with thecommo n Maed oRi o typeo fplinthit econcretion s orwit hth e lessfrequen t Ipixuna type. The Amazon data therefore do not suggest a consistent coincidence with a certain soil.Ther ear en oindication stha tth eclimati ccondition so fth ePlanalt oI are aar e dif ferent from thoseo nth e Planalto IIarea . The samei smos t probably true ofth e areas

198 of the inventory units in the region of the eastern limit of occurrence of the species. One likely explanation for the absence of Acapti west of the Curua-una and east and south ofth e middlereache s ofth eGuam a and the Capim istha t the speciesstil lha sa limited historical dispersion (SCHULZ states that Bruinhart growsver yslowl y and that its heavy seeds need to be pressed into the ground to germinate). Some support for thissuppositio n isgive nb yth efac t that inth esouther n part ofth eare ao fth e Santana inventory unit, which is a boundary area for the species, the colonies of Acapii are infrequent, small,an d consisting of smalltrees .

SOIL TEXTURE Sincesoi l textural differences are rather easily noticed bynon-pedologists ,th e forest inventorists paid attention to a possiblecoincidenc e of soiltextur e differences anddif ferences inoccurrenc e oftre e species.A st o the Planicie(inventor y units 1-15,cf. Fig. 22), HEINSDHK (1957, 1960)report s that tree species found more abundantly on light textured soils are Axud (Sacoglottisguianensis), Angelim rajado (Pithecolobium race- mosum), Faveira bolacha (Vatairea cythrocarpa), Jutai-mirim (Hymenaeaparvifolia), Itauba (Mezilaurus itauba) and Macucu (Licaniamacrophylla). Speciesfoun d regular ly on heavy textured soils, and not or only in very small numbers on sandy soils are Ingd(Inga alba), Tauari (Curatarispp.), Castanheira (Bertholletia excelsa) and Ucuuba ( Virolacuspidata, Virolacebifera). It was not possible to check in the field if this is indeed a general rule, which also applies to the Guama-Imperatriz area. A comparison in the office of inventory data and soildat a ofthi sare adi dno t givean yclea rindications .

One of the speciestha t iseasil y noticed in thefield b y non-botanists, isth eAngelim pedra (Hymenolobium excelsum).Thi s is a conspicuous upper-storey tree, normally occupying lesstha n 0.5%, averaged over larger areas,o f the total number of trees. In the Guama-Imperatriz area its occurrence is restricted to the area of inventory unit Ligac,ao.I t wasfoun d that the tree isconcentrate d on the Kaolinitic Yellow Latosol, very heavy textured (KYLr») of the very gently sloping edges of planalto. Elsewhere the tree occurs practically only, and less frequently, on very heavy textured soils at a lower leveltha n the planalto.Thes e are Red Yellow Podzolic soil, intergrade to Kao linitic Yellow Latosol, very heavy textured (RP-KYLPA), and, locally, Kaolinitic •Yellow Latosol, very heavy textured (KYL»A). The species is not found on the flat centres of planalto (also with KYL„ft), nor on terrains with reworked Belterra clay north of km 195,whic h have Kaolinitic Yellow Latosol, Compact phase, very heavy textured (KYL,c Vh). These are siteso fcipoa l or 'cipoalic forest' asalread y described. The apparent concentration ofth e specieso n veryheav ytexture d soilsi nthos e parts where thefores t isno t 'cipoalic', nor where climaticcondition s are adverse (the latter in the area of inventory unit A^ailandia), ismor e or less in agreement with observa tions in the Tocantins-Xingii region. HEINSDUK (1957) states that 'Angelimpedra is... found in patcheso nth e planalto (not on planalto de Santarem) inth ehighes t partso f the forest. Found occasionally on the higher parts of the slopes, prefers clayey soil'.

199 It should beborn e in mind that an established predominance of a certain species on sandy soil as against on clayey soil, does not necessarily mean that the speciesfinds better absolute growingcondition so nth eformer . Itma ywel lb etha tth especie si sles s frequently found on clayey soil only because it falls short in competition force with other species there. HEINSDIJK (1960) even states 'all observations confirmed the fact that atre especie sgrowin go nsand ysoi lwas ,o n anaverage ,alway ssmalle r and gavea much poorer impression thanth esam especie sgrowin go n heaviersoils' .

PLINTHITE CONCRETIONS That the occurrence of a tree species may coincide with the presence of fossil hard plinthite concretions of a certain type, became apparent during the combined forest inventory-soi l surveyo fth eGuama-Imperatri z area. The valuable timber rendering Pau amarelo (Euxylophoraparaensis), which speciesonl y occurs in the Atlantic sector ofth ehileia ,abov e Maranhao (DUCKE and BLACK, 1954),seem st o have its maximum frequence in the middle reaches of the river Capim. It is a low branching tree and occupiesmaximall y 1% ,average d overlarge r areas,o fth etota l number oftree si nth e forest. Withinth eGuama-Imperatri z area,th especie swa sno t measured inth earea so f the Santana, Medio Guama and Agailandia inventory units.Th e speciesoccur s rela tively frequently in the Candirii unit (4.0 m3/ha timber volume), and sparsely in the Alto Guama unit (0.6 m3/ha). Duringth efield wor k inth e area ofth eCandiri i unit, it was observed that the tree isconcentrate d in the uppermost reaches of the rivulets. In viewo fthis , one smallpar t wasstudie d in detail. It islocate d within the area ofth e Candirii unit, but near the boundary with the Alto Guama unit, namely at km 130. A semi-detailed soil survey was executed, while the forest inventorists made a 10% survey of the Pau amarelo. The results are shown on Appendix 3 (only thosePau amarelotree swer eenumerate dtha tar ewithi n 10 mdistanc eo fal llatitudina l traverses and the outer and innermost longitudinal traverses of each block, and within 5 m distance of the other longitudinal traverses). From this map it is apparent that the speciesi sconcentrate d onth e slopesan d theapproximatel yflat edge so fth e upper terrace, where plinthite concretions arefrequent . The soil on these sites isclassifie d as Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase(RP-KYL , CR).Th econcretion sai eal lo fth eIpixun atyp e(c/ .1.4.5) . Afe wtree s arefoun d on RedYello w Podzolic Soil,intergrad e to Kaolinitic Yellow Latosol,ver y heavy textured (RP-KYL„A), which occurs locally alongside the concretionary parts. Pauamarelo isabsent , or practically absent, on the Kaolinitic Yellow Latosol, Com pact phase,ver y heavytexture d (KYL,c Vh)tha t hasdevelope d on the reworked Bel- terra clay of thecentra l parts of the upper terrace,a s wella so n the Kaolinitic Yellow Latosol,intergrad e to Red Yellow Podzolic soil, rather heavy textured (KYL-RPr/j) of the second terrace (with the highest gross timber volume!),an d on the soils of the lowland stretches.

The coincidence isonl ytru eo fon etyp eo fconcretions . In theare a ofth e Santana in ventory unit, many parts have RP-KYL, CR soil, but usually with concretions of the

200 Maed oRi otype .Her en oPau amarelo tree swer eenumerate d byth efores t inventorists. In fact, somePau amarelo do occur in the southern part of this area, aswel l as ina fewpart so fth eare a of the MedioGuam d unit,bu t thetree s fell outside the inventory sampling units. It was observed that in these places they are growing invariably on RP-KYL, CRsoi l with Ipixuna type concretions, not on RP-KYL, CRwit h Mae do Riotyp econcretions .A fe wtree swer efoun d onRP-KY L„h soil . Also the few Pau amarelotree s in the area of the Alto Guama inventory unit were found to occur either on soils with Ipixuna type concretions, which in this area are only very locally present at the surface, or on RP-KYL„ft. The situation is somewhat different further southward in the Guamd-Imperatriz area, inth e area of the Liga^ao inventory unit. There the tree is found too, down to km 315, but in relatively small numbers. South of km 260th e Ipixuna type concretions givewa y to the Ligaclo type concretions, which have the same relative position and age as the Ipixuna ones (cf. 1.4.5).Th e speciesoccur s on the soilswit h either of these types.I n the northernmost part of the area with this unit, namely at km 195,anothe r small tract was studied in semi-detail.Th eterrai n isthi strac tconsist sessentiall yo f 1)a nuppe rterrac e(planalto) , whichca nb edivide d intototall yfla tcentra lsections ,an dver ygentl yslopin gedges ,an d 2) very sloping terrain, below the level of the planalto, usually with plinthite concre tions,o fth e Ipixuna type.Th ePau amarelo ofth esurveye d block isfoun d concentrat ed on the strongly sloping terrain, as well as on the gently sloping edges of planalto. The flat central sections of planalto have a 'cipoalic forest' (seebefore) ,withou t the species.Th e soil of the strongly sloping terrains is classified as Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase (RP-KYL, CR- Ipixuna).Tha t of the edges ofplanalt o isclassifie d as Kaolinitic Yellow Latosol,ver y heavy textured (KYL„A), which is also the soil of the central sections. Ipixuna type concretions are not found at the surface of the planalto edges,ye tPau amarelo ispre sent. Also elsewhere in the area of the Ligacao unit, the species is not restricted to the placeswit h Ipixuna typec.q. Ligacl otyp econcretions .Th etre eals o occurso n the very gently sloping concretionless terrains with KYLeft or RP-KYL„ A profile, even wherethes eterrain sar ea ta considerabl y lowerleve ltha n theplanalto . Itma yb enote d thatLigacl otyp econcretion san dgentl yelopin gterrain swit h KYL„ft occur also in the area of inventory unit A^ailandia, which lacksPau amarelo. In this stretch however,th eclimat ei sbelieve d to beto o dry for this and several other species (f/.lV.l.I).

Combining these observations, itca n besai d that inth e Guama-Imperatrizare a Pau amarelo occurs predominantly on soil with Ipixuna type c.q.Ligaga otyp e concretions (RP-KYL, CR-Ipixuna; RP-KYL, CR-Liga$ao)and , to a lesser extent, on very heavy textured soil (KYL„A and RP-KYL**) where this occurs on gently sloping terrain, i.e. when the forest is not 'cipoalic'. The optimal growing conditions for the species are apparently found on the concretionary soils classified as RP-KYL, CR-Ipixuna (respectively Ligacao),bu tth especie sca nsprea d toth ementione d veryheav ytexture d soils.Th elatte r are located alwaysnea r theconcretionar y soils.

201 Table12 Comparison of chemical data of soils withand without Pau amarelo (Euxylophora paraensis)

Cation

No. field Classification {cf. Table9 ) Location Clay exch. capac. pH-HtO Exchangeable metallic dcscr. classificacdo(cf. Tabela9) locacdo argiia capacidadt cat ton s descr, decarnfto BR-14 dt troca basestrocdvtis (%) (m.e./IOOg) (m.e./100g)

Ca* Mg*

IOILI WITH PAU AMARELO/JO/OJ com Pau amarelo 202 RP-KYL, CR (Ipixuna) km 109.6 54.3 59.9 13.07 2.91 4.5 5.2 0.66 0.18* 2.11 0.18* 216 KYLtiA km 291.3 69.8 90.8 13.20 3.88 5.0 5.2 3.08 0.33* 2.28 0.33* 236 RP-KYLtA km 232.0 63.0 83.5 11.26 3.31 4.2 4.5 0.41 0.24* 0.72 0.24* 204 RP-KYU.A km 172.0 26.7 55.9 10.83 3.28 4.0 4.8 1.08 0.35* 0.75 0.35*

SOILS WITHOUT PAU AMARELO/WfM ttffl PaU anUXTtlo 238 RP-KYL, CR (Mae do Rio) km 43.8 21.9 55.6 9.29 3.31 4.7 4.5 1.98 0.24* 0.84 0.24* 205 RP-KYL, CR (Paragominas) km 175.6 17.1 52.7 12.82 3.65 4.1 4.7 1.39 0.24* 1.00 0.24* 201 KYL-RPrA km 117.0 13.7 32.2 5.83 2.72 4.1 4.7 0.46 0.20* 0.36 0.20*

197 KYL-RPrA km 136.0 33.1 63.2 11.01 3.15 3.9 3.9 0.33 0.20* 0.10 0.20* 210** KYL»A km 247.0 74.6 88.5 14.86 4.56 4.0 4.7 0.87 0.25* 0.98 0.25* 230* • KYL, cvh km 178.8 60.3 84.5 9.99 4.25 4.4 4.8 1.11 0.26* 0.80 0.26*

208 •• KYL, ah km 69.0 77.2 90.3 19.46 2.92 4.0 4.7 2.52 0.25* 0.69 0.25* 194 KYL, crA km 63.0 14.9 35.3 4.16 2.94 4.3 5.0 0.27"' 0.48* 0.27* 0.48* 242 RP-KYLrA km 256.6 13.2 48.1 6.08 2.42 4.1 4.6 0.31"'0.16 * 0.31* 0.16*

1 237 RP-KYLrA km 205.8 14.6 42.9 5.51 2.15 4.3 4.5 0.21* 0.29 * 0.21* 0.29* 233 KYLm km 12.7 6.7 19.8 4.20 1.99 4.6 4.5 0.24* 0.14* 0.24* 0.14* 231 KYLm km 58.0 17.9 23.9 4.90 2.14 4.6 5.1 0.15* 0.15* 0.15* 0.15* 206 KYLm km 201.0 7.3 19.8 5.40 2.33 4.0 4.3 0.32* 0.32* 0.32* 0.32*

• i (Ca++ + Mg++) the half of jointly determinded bivalent cations a metade dot cations bivalentet determinadot em conjunto ** cover of cipoal or *cipoalic forest*Icoberlura dt cipoal oufloresta cipodlka

Any special soil qualities related with very heavy texture cannot beth e conditioning factors for the occurrence ofth e species because the RP-KYL, CR soil sometimes falls into other texture classes. Special rooting possibilities are unlikely to beinvolved , be cause both the RP-KYL, CR-Ipixuna and the RP-KYL„A have a rather firm and compact subsoil, whilst the KYLrft is deeply friable and porous. Probably, the species requires a high content of a certain (micro) nutrient, or a specific balance of nutrients.Compariso n ofanalytica l data doesno t givea congruen t difference between Pauamarelo bearing and non-Pau amarelo bearing soils (cf.Tabl e 12).I t is, however, quite possibletha t amor e refined and elaborate chemical analysis,and/o r mineralogi- calanalysis ,woul d reveal acorrelation . Theplinthit e concretionsthemselve sma ypro videth especifi cchemica lqualitie so fth esoils ,o rth efossi l soft plinthiteofte n associat ed. Asdiscusse d in 1.4.5.bot h are related to alan d surface of Late Pliocene age,com posed of the upper strata of the Barreiras or Alter do Chao beds. The Belterra clay, which forms the parent material of the non-concretionary very heavy textured soils, was deposited right on top of this, in a time immediately following (Plio-Pleisto-

202 Tabela 12 Comparacdo de dados quimicos de solos com e sem Pau amarelo(Euxylophor a paraensis)

Exchangeable metallic Active Available P.O. Org.C N C/N P-total P-total cations acidity P,C>, total total 100N basestrocdveis acidez ativa assimildvel (m.e./100g) (m.e./lOOg) (mg/lOOg)

K+ Na (Al)* P-Bray

0.250.0 8 0.110.0 4 1.200.2 4 1.50. 1 60 50 3.440.1 3 0.320.0 3 10.8 4.3 1.87 16.6 17.538 4 0.220.0 8 0.080.0 * 0.290.3 3 0.90. 1 60 50 3.600.5 0 0.340.0 7 10.6 7.1 1.76 7.0 16.710 0 0.21 0.10 0.100.0 3 1.740.6 6 1.90. 1 60 40 2.990.5 0 0.260.0 5 11.5 10.0 2.30 8.0 20.0 80 0.170.0 7 0.070.0 7 1.550.4 1 3.10. 2 40 40 2.780.3 7 0.230.0 5 12.1 7.4 1.74 8.0 14.410 8

0.250.1 1 0.080.0 8 0.53 1.59 3.40. 2 50 50 2.750.1 7 0.250.0 2 11.0 8.5 2.0025. 0 18.229 4 0.170.1 0 0.060.0 3 1.100.7 3 3.70. 2 50 80 3.630.2 9 0.250.0 4 14.9 7.3 2.0020. 0 13.827 6 0.110.0 6 0.050.0 3 0.890.4 9 2.70. 2 30 30 1.200.1 6 0.110.0 3 11.0 5.3 2.71 10.0 25.018 8

0.100.0 6 0.050.0 2 2.16 1.54 1.40. 2 30 30 2.450.2 9 0.220.0 6 11.1 4.8 1.36 5.0 12.310 3 0.260.1 2 0.070.0 3 2.16 1.07 1.20. 2 50 30 3.610.6 9 0.330.0 8 10.9 8.6 1.52 3.8 13.8 44 0.120.1 3 0.080.0 6 0.95 1.03 2.50. 2 80 60 2.590.6 3 0.300.0 8 8.6 7.9 2.76 7.5 30.8 95

0-330.1 1 0.100.0 5 2.370.6 2 2.50. 1 70 60 5.170.4 2 0.440.0 7 11.8 8.4 1.46 8.5 13.514 6 0.110.0 9 0.080.0 8 0.720.2 4 1.20. 2 30 30 0.650.2 5 0.060.0 3 10.8 8.3 5.00 10.0 46.212 0 0.060.0 7 0.040.0 4 0.890.5 2 2.20. 1 30 30 1.370.3 8 0.120.0 4 11.4 9.5 2.50 7.5 21.9 79

0.080.0 8 0.030.0 3 0.930.3 6 1.70. 1 30 30 1.390.2 0 0.110.0 2 12.6 10.0 2.72 15.0 21.315 0 0.060.0 7 0.030.0 3 0.720.6 4 1.10. 2 30 30 0.880.1 8 0.060.0 2 14.7 9.0 5.00 15.0 34.016 7 0-080.0 7 0.040.0 3 1.080.5 3 1.20. 4 30 30 0.870.2 2 0.090.0 3 9.7 7.3 3.32 10.0 34.413 6 0-100.0 8 0.030.0 2 0.81 0.56 1.70. 1 20 20 1.230.1 9 0.080.0 3 15.4 6.3 2.50 6.7 16.310 5 a ==topsoi l {At horizon, or An when the former is subdivided) solo superficial (horizonte Auou An quando oprimeiro estd subdividido) b == subsoil (Bt horizon, or Bn when the former is subdivided) subsolo (horizonte Blt ouB n quando oprimeiro estd subdividido)

cene).A peculiarit y inparen t material,relate d with these geologictimes ,i s probably responsible for the uneven occurrence of the Pauamarelo in the Guama-Imperatriz area.

THE ECO-SITEO F MAHOGANY It ist ob eexpecte d that detailed studiesoutsid eth ePlanici epar t ofAmazonia , where the soilsvar y considerably both chemically and physically,wil l result inth e establish mento fmor ean d moredefinit e coincidencesbetwee n soilan doccurrenc eo f individual tree species than in the Planicie itself. DUCKE and BLACK (1954), for instance, report that on patches with 'Terra Roxa-like' soil, developed from diabase, the forest cover hasa floristic compositio n distinctly different from that onadjacen t soils.O nthi srela tivelyric hsoil ,mor esof t woodedtre especie swoul d bepresent . Thegrowin gsit eo fon etre especie so fth e non-Planiciepar t of Amazonia, namelyth e very valuable Mahogany(Mogno: Swietenia macrophylla), was studied insom e detail, namelyi nth eso-calle d Araguaia Mahogany area(cf. Appendices2 an d 6).Thi sare ai s

203 located in thetransitio n zonebetwee n theAmazo n hileiaan d thesavannah s of North- Eastern Brazil. It has a tropical humid climate with a pronounced dry season (Awi n Koppen's classification, cf. Fig.2) .I n the western part of the area,th e Pre-Cambrian crystallinebasemen t outcrops;eas to f this,strip so f Palaeozoic and Mesozoicdeposit s occur (Devonian, Carboniferous, Permean, Jurassic, Triassic), which belong to the sedimentary basin of Maranhao (cf.Appendi x 8).Locally , Quaternary sediments are present. Thefollowin g summarizesth efield dat ao fth ecombine dfores t inventory- soilsurve y executed inthi sarea :

1. Mahogany is absent in the areas of the mapping units with savannah cover, being L,QU and KLS,s .Als oth e mapping unitsL ,s san d KLS,T ,covere d withlo wan d high shrubvegetatio nrespectively ,lac k mahogany. 2. The species isver y sparsely present in the area of mapping unit DL,s + AF,G ,a s welli ntha t ofmappin guni t L,CH . 3. Low,thi n and stunted mahogany trees arefoun d scattered inth eare a ofth e mapp ing unit with Hydromorphic Grey Podzolic soil, with high base saturation, Shallow phase (HPftb, s; Profile 46)whic h has a general vegetative cover of hydromorphic de ciduousshrub . 4. Smallquantitie so fmahogan y arepresen t inth eare ao fmappin guni t KLS, F,whic h is under forest cover. Here, the tree isfoun d in the transition strips between the dry land and the bottom lands, which usually have deep Ground Water Humus Podzol (GP).I ti sals ofoun d onth enarro w stripso flowlan d alongth erivulet sinclude d inth e mapping unit and havingLo w HumicGle ysoi l(LHG) .Th etre edoe s not occur regu larly onal lsite swit hthes esoils ,bu tonl ylocally . Theimpressio n istha t theselocalitie s have relatively rich ground water, namely that coming from nearby calcareousde posits(chert swit hlime-stone) . The mahogany trees found in the area of this mapping unit can become very big, namelyu pt odiamete rclas s 151. 5. The floodplain and lowland soils of considerable extent are mapped under unit F. Thosealon gth erive rAraguai aproper ,whic har eeithe rAlluvia lsoi l(A )o r LowHum ic Gley soil (LGH), have no mahogany in their vegetative cover, which is often a creeperforest .Thos ealon gth etributaries ,whic har eLo wHumi cGle ysoil(LHG) ,hav e mahogany only locally. Asi n the case of those mapped under KLS, F(se e4) ,th e im pression is that these localities have relatively rich ground water, inthi scas ecomin g from nearby located cherts with lime-stones, and calcareous and gypsiferous silt and clay-stones.

6. Mahogany isfoun d in rather small quantities in the area of mapping unit RPr»&. Within this area, the tree is growing predominantly in the narrow strips of lowland along the rivulets, with Ground Water Laterite soil, intergrade to Hydromorphic

*)Diamete r class 0 stands for 0-4 cm DBH (diameter at breast height), diameter class 1 stands for 5-14 cm DBH; diameter class 2 for 15-24 cm DBH, 3 for25-3 4c mDBH ,etc .

204 Fig.25 Tiposde floresta e solosna area Araguaianade Mogno; levantamento 100% do Mognonum bloco de200 ha (para a legenda vejapagina205 eseq.)

6°ors-

Grey Podzolic soil, Micaceous phase (GL-HP, M; Profile 16).Th e tree is found also where greenish-black, hard schists (epidote and biotite schists) outcrop or nearly out crop,whethe r onth e stripso flowlan d oro nth eundulatin g torollin g upland.Th esiz e of themahogan y treesi nthi s area can bever y large, up to diameter class 15. 7. The bulk of the mahogany of the survey area isfoun d in the area of mapping unit H, Hydromorphic soilsundifferentiated , where thelan d isextensivel y flat. The maho gany is not spread regularly over all the land concerned, but concentrated in large patcheswher eth etre ei softe n thepredominan t constituent ofth ecanopy :canteiros de mogno. Adetaile d surveyo f a20 0h a block withinthi sare a wasexecute d to determine whicho fth evariou shydromorphi c soilssuppor t mahogany{cf. Fig.25 ;fo r themaho gany inventory map of this block, cf. GLERUM and SMIT, 1961; or SOMBROEK and SAMPAIO, 1962). The results ofthi s survey are as follows:

A. Forest of small-sized mahogany. Low, rather open canopy, consisting almost ex clusively oflow ,thi n (largelydiamete rclasse s2, 3an d 4)an d stunted mahoganytrees ; high, fairly dense undergrowth, with many thin creepers or climbers; rather closed field layer.Thi s isfoun d on Hydromorphic Grey Podzolic soil,wit h highbas e satura tion,Dar k phase (HPA6, D; Profile 45,se eals oPhot o23) .

205 Foto23 O micro-relevo de canale/ese a zona herbdcea e subarbustiva dos solos em que ocorre fre- qiientementeo Mogno.Os terrenos relativos sao extensamente pianos e de drenagem imperfeita. Osolo $ 'Hydromorphic Grey Podzolic', comsaturacao de basesalia. A direitae ao cenirono fundo hd dois mognospequenos. A/e'm das palmeiras, estasespecies formam prdticamenteas unicas que constituem o dossel (area Araguaiana de Mogno, Bloco Piranha)

Photo23 Themicro relief of canaletes andthe field layer of thesoils on which mahogany isof frequent occurrence.The terrainsconcerned are extensively flat and imperfectly drained. The soil is a Hydro morphicGrey Podzolic soil, withhigh base saturation. On the right and at centre-backgroundare two stunted mahogany trees.Apart from palms, this speciesforms practically the only constituentof the canopy(Araguaia Mahoganyarea, Bloco Piranha)

B. Forestof normal-sized mahogany. High, almost closed canopy, consisting for a good part of medium-sized (diameter classes4 t o 7)mahogan ytrees ,i n a much more open pattern than in the forest type A; high, only fairly dense undergrowth, with ratherfe wcreeper so rclimbers ;ope nfield layer .Thi svegetatio ni sfoun d predominant ly on Hydromorphic Grey Podzolic soil,wit h high base saturation, Ortho (HPA&, O; Profile 44)* , but also on Ground Water Laterite soil, intergrade to Hydromorphic Grey Podzolic soil, Clay-stone substratum phase (GL-HP, c; Profile 15).Th e latter soilhoweve ri scommonl yfoun d onth eedge so fth eterrai nwit hthi sfores t type,o r on small, encircled patches with less mahogany (interrupted shading on the map; also applyingt oa centra lpatc hwithi nth eare ao ffores t typeA) .

*)A patch of this soil, at a corner of the surveyed block, has a coverage of dense shrub forest without mahogany (D). It is supposed that former fires and/or very imperfect drainage is the cause for this deviating vegetation type.

206 C. lCipoalic'forest. High,rathe r open canopy,consistin g of several tree species and comparativelyman ypalms ,bu trarel ywit hmahogany ;high ,fairl y dense undergrowth withcommo n and ratherthic k creeperso rclimbers ,an d afrequen t occurrence ofhig h Sororoca (Ranavalia guianensis);ope n field layer. This is found on Ground Water Laterite soil,intergrad et o HydromorphicGre y Podzolicsoil ,Dee pphas e(GL-HP , D; Profile 14), and also on Ground Water Laterite soil, intergrade to Hydromorphic Grey Podzolicsoil ,Clay-ston e substratum phase(GL-HP ,c ; Profile 15).

E. Cipoal (creeper forest). Little or no canopy, and without mahogany; dense under growth, largely consisting of thin creepers or climbers; openfield layer . This vegeta tion typei sfoun d on poorly drained patches,wit h Low HumicGle ysoi l(LHG) .

From these data it becomeseviden t that in the surveyed area, mahogany growsver y predominantly onterrain swit ha nimperfec t drainage,wit hwel ldevelope d hydromor phic soils. The occurrence of mahogany is not general on these terrains, but concen trated on afe w specific hydromorphic soils,o n whichth especie soccur seithe r scatter edo ri na dens epattern ,remain seithe rlow ,thi nan d stunted,o rattain s mediuman di n somepart slarg ediameters . The tree occurs scattered in the strips of lowland, with GL-HP, M soil, of the area where the Pre-Cambrian crystalline basement (serie Araxa) outcrops. Thetre e occurs in a dense pattern on extensively flat terrains with HPA&, Oan d HPA&, D soils.Thes e soils have developed from gypsiferous and calcareous silt and clay-stones which be long most probably to the Motuca member of the Pastos Bons beds, of Jurassic- Triassic age (Piaui beds of the Upper Carboniferous according to other geolo gists).Wher eth etre eoccur soutsid ethes eterrain s iti susuall y whereth egroun d water isric h (high pH),o r whereric h rock (epidote and biotite schists)outcrops . Thefrequente d hydromorphic soils(Profil e 16,an d especially the profiles 44an d 45) are all relatively rich.The y have silicate clay minerals of the 2:1 lattice structure; the baseexchang ecapacit y isrelativel yhig h(25-3 5m.e./lO Og clay ,afte r correctionfo r the organicmatte rcontent) ;th ebas esaturatio n ismedium ,an d inth e subsoil high; Ca++ is the predominant exchangeable cation, but also exchangeable Mg++ is present in considerablequantities ,especiall yi nth eGL-HP , M. The HPftb, o and HPA&, D soils have a high pH in their deeper subsoil (6-8). In this subsoil, free anions, namely sulphates, carbonates, and/or chlorides are present. It is quite possibletha t these anions,especiall y thesulphates ,ar eth econditiona l factor for thepeculia r abundancy ofmahogan y on thesoils .Bu tth efield an d laboratory data on mahogany and non-mahogany soilso f the area are too few to allowdefinit e conclusi ons. Thedifference s in sizean d density ofth especies ,particularl y ofth e mahogany onth e HPAJ,, O and HP/n,, D soils respectively, warrants discussion. A likely explanation is that onth e HPA&, Dsoil ,wit h itsric h subsoil near thesurface ,th econdition s for rege neration are better, but full development of the trees is hampered by the very poor physicalqualitie so fth esoil ,whic h restrict root penetration. Support for this supposi-

207 tion isfoun d in the scattered occurrence of small sized mahogany on the Hydromor- phic Grey Podzolic soil,wit h high base saturation, Shallow phase (HPM,,). The rich subsoil of this is also rather shallow, and it possesses extremely poor physical qualities, resulting in a general cover of hydromorphic deciduous shrub. That the scattered mahogany in the areas of mapping units RPrh& and KLS, F can attain a bigger size than on the HP^, o is probably explained by taking into account that in thesearea sth ephysica lqualitie so fth esoil sconcerne dar egenerall y better.

PALMS The occurrence of palms in theAmazo n forests seemst o be,i n general, more related withth evariation s inclimati can dsoi lcondition stha n that of many dicotylenous tree species. AUBREVIIXE (1958) obtained the impression that every region, and every milieu of the hileia, has its characteristic large palms. Since most palm species are comparatively easily noticed and identified, they may therefore be suitable indicator plantsfo r vegetation typesan d soilunits .Quantitativ edat a onoccurrenc eo fpalm sar e not available.Th egrowin g siteso fsevera lspecie so fth e uplands,particularl y Tucumd, Babacu, and Bacaba have however been examined to a degree. The Tucumd (Astrocaryum vulgare) seemst o beconcentrate d on sandy, low uplands where a shallow ground water level occurs.Th e palm is,fo r instance,characteristi c of the sandy Ground Water Laterite soils of the eastern part of Maraj6-island, whether undersavanna h or forest. Thecommerciall y valuable1Babacu (Orbigniaspeciosa) isofte n amai nconstituen t of the palm forests in the transition zone between the hileia and the savannahs of North Eastern Brazil. The southern most part of the Guama-Imperatriz area contains an extremity of this zona decocais. The Babacu there was found to beconcentrated , ina dense pattern, on comparatively fertile soil. This is the Red Yellow Mediterran ean-like soil, (RM; cf.Appendi x 1 and Profile 41)whic h has developed on silt-stones and shales of the Cretaceous Codo beds. Also in the Araguaia Mahogany area the Babacu isencountered . The palm occurs there, fairly frequently, on Red Yellow Podzolic soil,wit h rather high base saturation (mapping unit RPr*»; cf. Appendix2) . It is also frequent on the highest parts of the floodplain of the Araguaia river proper (mapping unit F),whic hhav eAlluvia lsoi l(A )consistin go fver y micaceoussilt ysands . The palm hasa scattered occurrence in the area of the association of Dark Red Lato- sol, Shallow phase and Acid Brown Forest-like soil,Gravell y phase (DL,s 4- AF, G). On Lithosol,Quartzit e substratum phase(L ,QU )an d Lithosol,Sand-ston e substratum phase (L, ss) the species only occurs locally. The species is absent in the areas of the other mapping units, where the soils are well-drained but poor(Kaolinitic Latosolic Sand, Forest phase;KLS , F),o r imperfectly drained and relatively rich (Hydromorph icsoils ,undifferentiated ; H).Th e palm therefore isconfine d to the westernpar t ofth e survey area. It occurs on soils developed on the Pre-Cambrian crystalline basement

') Stands of Babacu are often denser where under human influence, owing to repeated burning of thevegetatio n between the palms, which themselves arefire-resistant, an d to some enrichment planting.

208 (mica schists,quartzites ) or on Devonian-Carboniferous deposits (shales,sand-stones , silt-stones), as well as on young, moderately well-drained soils from recent sediments (micaceoussilt ysands) . The above data suggest that the Babagu predominates on relatively fertile soilswit h free or almost free drainage, which may vary in depth and structure. It is, however, quite possibletha t inth eare ao f maximal occurrence ofth especies ,whic hi sth e north ern part of the State Maranhao, the palm isles sexactin g ast o soilconditions . Contrary toth e Babafii, the Bacaba palm (Oenocarpusbacaba) seemst o growpredo minantly on well-drained, poor sandy soil, in particular on Kaolinitic Latosolic Sand. In the Araguaia Mahogany area at least, this palm isfull y confined to the Kaolinitic Latosolic Sand, Forest phase(KLS , F; cf.Appendi x 2). Inth e axial part of Amazonia also, this palm species is found particularly on the same type of soil. A palm species not selective as regards soil conditions seems to be the Inajd (Afaximiliana regia). In the Araguaia Mahogany area the species isfoun d on practically all encountered soils, exceptthos eo fth esavannah s proper.

IV.2 The Uplands with Savannah or Savannah-Forest Cover

A number of places on the uplands havea coverage of savannah or savannah-forest, instead offorest . Thesevegetatio n typesar edefine d and described in1.5.2 . Aspossibl e causes of the occurrence of the savannahs and savannah-forests the following should beconsidered : 1. Local unfavourable climatic conditions, especially as regards relatively low total annual rainfall and/or arelativel y longan d pronounced dry season. 2. Local marked influence of man, especially as regards repeated felling, repeated burning,an d animal husbandry. 3. Local unfavourable edaphic conditions. This applies especially to the effective soil moisture reserve for the period of the year that the evapo-transpiration from a forest coveragewoul d exceed the precipitation (dry season or lessrain yperiod) . Effective soil moisture reserveca n be insufficient for various reasons: 1. Therema y bea ver ylo wmoistur estorag ecapacit y peruni t ofsoi lmaterial .Fo r this reason verysand ysoils ,eve nw he nthe yca n bedeepl ypenetrate d byroots ,ca n beshor t ofa sufficien t soilmoistur e reservefo r forestgrowth. 2. Therootin gspac ema yb esmall ,becaus eo fa smal lamoun t ofsoi lmaterial .O nver y stony soils, and on very shallow soils upon impermeable bedrock, the total stored moisture may be too small to carry a forest coverage through the dry period, even whenth estorag ecapacit y peruni t ofeart h ishigh . 3. Therema yb eintermittentl y imperfect drainage.A shallo wgroun dwate rleve ldurin g the rainy season greatly restrictsth edept h of penetration ofth e roots.I fo n such asit e the ground water leveldrop s considerably during the dry season, then the short roots loosecontac t with it,an d the moisture reserve in the rooted layeralon e isto o small to sustain forest growth.

209 Table13 Thesavannahs andsavannah-forests of Amazonia, intheir dependence upon edaphic andnon- edaphicfactors

Land form Drainage condition Soil constiluicao da terra condicao de drenagem solo

Parts of Early Pleistocene terraces excessively drained White Sand Regosol (Giant Podzol) partesde terracos doPleistoceno drenagem excessiva Regosolode Areia Branca Inferior (Podzol Cigante)

Undulating or mountanous terrains well-drained Lithosol outside Planicie terrenosondulados ou montanhosos boa drenagem Litosolo fora da Planicie

Pleistocene terraces well-drained Kaolinitic Yellow Latosol Kaolinitic Latosolic Sand terracos do Pleistoceno boa drenagem LatosoloAmarelo Caolinitico AreiaLatosolica Caolinitica

Flat watershed parts of Pleistocene intermittently imper Ground Water Laterite soil or Early Holocene terraces; Cre fectly drained taceous or Early Tertiary peneplanation levels partes planas, nos divisores de dgua, drenagemintermitente- soloLaterita Hidromdrfica de terracos do Pleistoceno ou do mente imperfeita Holoceno Inferior;superficies pene- planadas do Cretdcio ou Tercidrio Inferior

Elongated patches, often along imperfectly drained Ground Water Podzol rivers, of Late Pleistocene sandy terraces faixas, muitas vezes ao longo dos drenagem imperfeita Podzol Hidromorfico rios,de terracos arenosos do Pleisto ceno Superior idem imperfectly drained Ground Water Podzol drenagem imperfeita Podzol Hidromorfico

Floodplains and lowlands, of poorly drained Low Humic Gley soil Holocene age Humic Gley soil Saline or Alkali soil vdrzeas, de idade Holocena ma drenagem soloClei Pouco Humico solo CleiHumico soloSalino ou Alea lino

210 Tabela 13 As savanas efloresta-savanas da Amazonia, em suadependincia de fatoreseddficos enao eddficos

Vegetation type, in dependence upon the degree of human influences tipo de vegelacao, emdependincia dograu de influencias anlropogenicas

PATCHY SAVANNAH-FOREST CAMP1NA-RANA (edaphicleddfico)

SAVANNAH CAMPO (edaphic/eddfico) felling, felHng SECOND«mNn . ' burning, SAVANNAH HIGH FOREST burni burning "g t war' , SHRUB Srazing_^ CAMPO Mala alia denubacao, Capoeira derrubacao, Arbus'° derrubacao, fnon-edaphic/ incendimento incendimento pastacdo e mo edal,c0> incendimenlo

EXTENSIVE EXTENSIVE SAVANNAH SAVANNAH-FOREST burnme CAMPO CAMPINA-RANA incendimento(maske d edaphic/ (edaphicleddfico) eddfico encoberlo)

PATCHY SAVANNAH-FOREST CAATINGA AMAZONICA (edaphicleddfico)

PATCHY SAVANNAH CAMPIN A (edaphic/eddfico)

LOWLAND SAVANNAH CAMPO DE VARZEA (edaphic/eddfico)

211 4. Impermeable layers may occur inth e subsoil (Ortstein; densesof t plinthite). These layers restrict root development and thusth e layer ofeffectiv e soilmoistur e reserve. Actually, theabov esoi lcondition s often occurtogethe r and are interdependent. For instance,a nimpermeabl esubsoi llaye rca n givea shallow , perched ground water level during the rainy season. On the other hand, intermittent imperfect drainage results often in aver y sandy topsoil and a dense clayey, or cemented, subsoil. Other soil fac tors which might be considered as hampering forest growth are: a restricted ground support, and alo w natural fertility. In Amazonia, however, these can berule d out as causes of savannah or savannah-forest. Asregard sa restricte d ground support therei s the example of the igapo. Although it has a very shallow rooting,nevertheles s a high forest coverage occurs.A sregard sa lo wnatura l fertility there isth efac t that practical ly all forested Planicie soils have a very low base saturation. And even soils with an extremely lowcatio n exchange capacity, also in their topsoil, often have forest cover age(Kaoliniti c Latosolic Sand, White Sand Regosol).Th e closed-nutrient cycleo fth e tropicalfores t coverage,onc eestablished , doesapparentl y not depend upon fertility of thesoil . With the above considerations in mind, the origin of the various savannahs and sa vannah-forests ofAmazoni a willb ediscussed , asfa r asthi s ispossibl ewit hth e limited amount ofavailabl edata .Thei rorigi n isschematised ,ver ytentatively , inTabl e 13.

IV.2. 1 Primarily Non-Edaphic Upland Savannahs

In this subchapter the upland savannahs within the hileia are discussed, which are believed to have originated primarily by local adverse climatic conditions or human influence. Many notes on the upland savannahs of eastern Amapa, of southeastern Marajo, and ofth enorthban k ofth eLowe rAmazo n riverar ecompile d inth estud yo f SUTMOLLERet al. (1963).Th e savannahs ofth e northern part of Rio Branco Territory areno t discussed becauseth e areaconcerne d falls outside the hileia.

IV.2.1. 1 TheUpland Savannahs of Eastern Amapa Territory The upland savannahs (campos)o f eastern Amapa Territory are located on flat or gentlyundulatin gterrain swhic har ebetwee n 5 and 50m aboveloca lrive rlevel .Excep t for narrow stripso f bottom land, the soils of these savannahs are well-drained.The y are commonly of rather heavy texture, but very heavy textures and very light textures respectively areals o found. Anumbe r ofthe m contain plinthite concretions.Th esoil s with relatively heavytextur e have often compact subsoils.I nthei r general characteris ticsth e soilsar e identical tofores t supporting soilsnearby , for instancethos e of Porto Platon, and elsewhere inth ePlanicie .The yar eclassifie d as KaoliniticYello w Latosol, of various texture (KYL),Kaoliniti c Yellow Latosol,Compac t phase of various tex ture (KYL, c), Kaolinitic Yellow Latosol, Concretionary phase (KYL, CR) and Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase (RP-KYL, CR)respectively .Detaile danalytica ldat a aregive ni nChapte rV .

212 For southeastern Amapa Territory, therei sa greater amount of climatical data than for many other parts of Amazonia. Weather recording stations are installed at Maca- pa, Santana, Porto Platon and Serra deNavio .Th efirst thre ear elocate d inth esavan nah area.The y each haveo n an average 3month spe ryea rwit h each lesstha n 50mm , and 5month s with lesstha n 100m m rainfall. Of Serra de Navio,wel l in the forested part, only oneyear' s recordings are available. During that year, only one month had lesstha n 50m mrainfall , andtw omonth sles stha n 100m meach . A distinct difference in climate is likely to exist between the savannah area and the forested area. But the dry season of the savannah area is not more pronounced than that of many other parts of eastern Amazonia which have a forest coverage (cf. the Figs. 3an d 4, and Fig. 12).Advers e climatic conditions can therefore not be a main cause of the savannahs. But even when the present day climate in the savannah area would have been unfavourable for forest growth, a climatic origin of the savannahs could not have been taken as proven. A distinct dry season may be the result rather than a cause of savannah coverage.O n extensive openterrains , the air becomes more heated at daytime ,resultin g in lessrainfal l from air with the same absolute humidity as that carried to forested areas (dry season showers in Amazonia usually fall in the afternoon). The Amapa savannahs reportedly already existed in 1600. Patches of Terra Preta, sure indicators for former Indian settlements, are frequent. At present, the Amapa savannahs are grazed fairly intensively, and burned repeatedly. The boundaries with the forested part are often verywel l defined, for instance at Porto Platon. With every newfire, an d helped by shifting cultivation and grazing, the savannah area increases slightly.Thi s process wasobserve d in the headwater region of Igarape do Lago. In summarizing these data, it is concluded that the upland savannahs of eastern Amapa aredu eprimaril y tolong-lastin g human influence.

IV.2.1.2 The UplandSavannahs of South-Eastern Marajo Island The upland savannahs of south-eastern Marajo are located on flat terrains, and on smallridge s(tesos) withi nth ecentra llowlands ,bot h 1 to 3 mabov eth eleve lo f flood ing or submergence. The soils of these terrains are often slightly imperfectly drained, constitutingsand y Ground Water Lateritesoi l(GL) .Th ehighes tpart shav eKaoliniti c Latosolic Sand (KLS). Only a few ofth e savannah terrains,fo r instance those westo f Mariahi at themediu mcours eo fth erive rAfua , havea thic k surface layero f bleached sand, which belongst o the so-called Ground Water Laterite soil, Low phase. Portions of the described lowuplands , with the same soils, have a forest coverage, though ofrathe r poorquality .A nexampl ei sth esurroundin g ofSoure .Th edr yseaso n is distinct in eastern Marajo, although the annual rainfall is high (Soure has a total rainfall of290 0mm ,3 month swit hles stha n 50m meach ,4 month swit h lesstha n 100 mmeach).Thes eclimatica lcondition s apparently dono t preventfores t growth andar e anyhow not inferior tothos eo fman yothe r partso feaster n Amazonia. It is known that in pre-Columbian times, relatively many Indians werelivin g on the island. Their settling sites were the above mentioned low uplands (patches of Terra

213 Preta, cf. III.3.4). At present, the savannahs are repeatedly burned and intensively grazed. The boundaries with the forested parts of the same uplands are often sharp. Locally, deforestation ispurposel y effected for converting the land into pasture.

It is concluded that the origin of the upland savannahs of south-eastern Maraj6 is largely anthropogenic, both from former Indians' and present farmers' practices. Edaphiccondition sar e marginal.

IV.2.1. 3 TheUpland Savannahs at theNorthbank ofthe Lower Amazon River The soilso fth e upland savannahs atth e Northbank ofth e Lower Amazon riverwer e studied in a number of places. The narrow strip of savannah on high upland (about 60 m above river level) right alongside the Amazon river, from Prainha to Monte Alegre,ha swel lt oexcessivel ydrained ,ver ysand y soil.Thi ssoi lis ,however ,ver ysimi lar to that of the 'flanco' areas southeast of Santarem that are largely under forest cover. Bothar eKaoliniti c Latosolic Sand(KLS) 1.I nth egras s covered, approximately flatcentr e of the Dome of Monte Alegre,th e soilsar e shallow, while drainage condi tions are often imperfect. Ground Water Laterite soil, intergrading to Lithosol (cf. Profile 17)i srathe r frequent. The savannahs at about 20k m N of Prainha (Desterro) where the country is partially very broken, are for a part located on shallow and im perfectly drained soil,classifie d asGroun d Water Lateritesoi l(cf. Profile 2). Kaolinitic Latosolic Sand (KLS) and Kaolinitic Yellow Latosol, medium textured (KYLm) are howeverals o found. Thedat a ofth eonl yweathe r recordingstatio n located inth eregio n (6bidos) point to a relatively dry climate. On climatical maps (cf. Fig. 2) often a connection is drawn with the northern part of Rio Branco Territory which has a pronounced dry season (Awtyp e of Koppen). DUCKE and BLACK(1954 )repor t the occurrence ofdr y and low forests in the region. Incompariso n with other parts of Amazonia, the relief isgreatl y varied (Dome of Monte Alegre;tabl e lands between Almeirim and Prainha). Already BOUILLENE (1926) observed that the savannahs in the district between Almeirim and 6bidos are largely found at the western feet of the elevated parts. He supposed that, withth eprevailin geaster n winds,a smalle ramoun t ofannua l rainfall atth elay-sid eis . besidescondition s ofsoil ,on eo fth edeterminin g factors for thesavannahs . All these savannahs are more or lessregularl y burned. Cattle grazing during the dry seasoni so fimportanc ei nman yparts .Th epresenc eo fman yan d largepatche so fTerr a Preta indicatea comparativel y strongforme r Indian influence.

It isconclude d that acombinatio n ofadvers eclimati ccondition s and human influen ce islargel y responsible for the occurrence of upland savannahs at the Northbank of the Lower Amazon river. In several parts adverse soil conditions are a contributing factor.

') Directly around Santarem this soil is also covered with savannah.

214 IV.2.2 Upland Savannahs and Savannah-Forests of Edaphic Origin

IV.2.2.1 Savannahsand Savannah-Forests of thePlanicie

INTRODUCTION Theabov ediscusse d savannahs areal llocate d inth enorth-easter n part ofAmazonia . Aconsiderabl epercentag eo fth eothe ruplan dsavannahs ,an d ofth e savannah-forests, arelocate d inth ePlanici e(Fig . 12- onwhic honl y the extensive savannahs and savan nah-forests are indicated -, cf. Fig. 17).The yar eal lencircle db yhig hforest .Th erelie f isnormall yflat. Fro m theweathe r recording stations,althoug h theyar e few, itca n be deduced thatth eclimat ei sapproximatel y uniform overlarg eareas .I tca ntherefor e be saidtha t these savannahs and savannah-forests are not due to adverseclimati c condi tionsin loco 1. Incontras t toth emajorit y ofth esa\annah-forests(campina-ranas,caatingasamazdni- cas) and the savannahs of small extent (campinas), allstudie d savannahs of largeex tent (campos) showtrace so f burning,o fvaryin gfrequency . It ishoweve r supposed,i n agreementwit h DUCKEan d BLACK(1954) ,tha t burningi sno tth ecaus eo fthes esavan nahs. High forest adjacent to them does not get burned unlessfelled , and then secon dary forest emerges. Only under long-lasting and pronounced human influence this latter may degenerate into savannah. But thedwellin gsite s ofth e present population, aswel la sthos e ofth eforme r Indians,ar econcentrate d onth emai nrive rbanks ,whil e thecampo sar efoun d inwatershe darea sfo ra goo dpart .O fcourse ,som eslo wincreas e in the area ofa savannah, at theexpens e ofth eadjacen t forest, mayhav etake n place, but their centres must have had an original vegetative cover that isliabl e to burning. The present day vegetation of the campina-ranas, caatingas amazonicas, and campi nas may be stillth e original one, but that of the campos certainly has changed consi derably, under influence of burning. The original vegetation of the campos terrains wasprobabl y akin d ofsavannah-forest . At present, however,th espars ewood y plants are offire-resistant species ,wit h xerophytic leaves and thick bark (Curatella americana,an d others). Becausethe y easilyregenerat e afterfire, tufte d grassesar e frequent. Mosto fth ecampo sterrain s are not grazed bycattle ,o r very infrequently. Grazingi s absent inth e campinas,campina-rana s and caatingas amazonicas. It is concluded that neither adverse climatic conditions, nor anthropogenic factors arecause so fth eexistenc eo fth earea swit hvegetatio ntype sinferio r toth e surrounding highforest .Th epresent-da y character ofthes evegetatio ntype sma yb ea modifie d one, which isconditione d byfire. Bu t the origin of both the savannahs and the savannah- forests in the Planicie, except those discussed in IV.2.1, is to be found with adverse edaphic factors, notably imperfect drainage, impermeable subsoil layers, and/or very sandy and bleached topsoils. Inth e following, all known data about the extent, topo-

') This does however not contradict the fact that the present-day micro and soilclimat e of the savannah terrains may be distinctly different from that of adjacent forested terrains. Very interesting data in this respect are recorded by SCHULZ (1960) in thecoasta l region of Surinam, for sites with rain-forest, savannah-forest, and large clearings respectively.

215 graphy and soils of the savannahs and savannah-forests of the Planfcie that are of edaphic origin arebrough ttogether .Th e maincriteri a for theclassificatio n of thesoil s concerned are summarisedfirst (fo r full description of the soils cf.III.2) : Amor e or lesssandy ,non-bleache d topsoil (Ahorizon )ove ra subsoil (Bhorizon )o f soft plinthite {i.e.,dense ,normall y heavytexture d material withmany ,coars ean dpro minent mottleso fre d ina whit eo rligh t greymatrix )i sa Ground WaterLaterhe soil. A sandy, bleached topsoil over a subsoil of soft plinthite constitutes the so-called Ground WaterLaterite soil, Low phase. A sandy, bleached topsoil over a subsoil containing an Ortstein (i.e.a more or less indurated, homogenous dark brown layer)i sa Ground Water (Humus) Podzol. Aver ydeepl y bleached sandy soil isusuall y called White SandRegosol, although ac tually the very thick Ahorizo n ofa deepGroun d Water Podzol,o r of adee p Ground Water Laterite soil, Low phase, may be involved.

FIELD OBSERVATIONS The soilso fth e campinaso fsom eexten t inth e Bragantina area, for instance the one of Vigia, are mapped by FILHO et al. (1963) as 'Regosol-Ground Water Podzol'. DAY(1959 )give sdat a on the regionth elowe r Gurupi river,whic h ispar t ofth eCae - te-Maracassumeare a (cf.Fig . 19 and Photo 24).I nthi s region, savannahs and savan nah-forests are very frequent. The terrains concerned consist of relatively low upland, largely submerged withrai nwate r duringa par t ofth eyear . DAYfoun d GroundWate r Laterite soil (the 'Ortho' type) and especially Ground Water Laterite soil, Low phase to beth e common soils.1Groun d Water Podzol and White Sand Regosol were found also,bu t mostlyi nsmal lpatches .The ywer esee nfo r instancealon gth erive r Maracas- sumei nth e southern part ofth e survey area, wherefrequentl y they occur asa narrow band alongth eriver .I tma yb enote d thatth eprofil e studiesi nthi sare a normallywen t to 120c mdept h only. In the upper part ofth e Guama-Imperatriz area there occursa narrow terrace along the rivuletsan d at 1-3 mabov ethei r level(Epi-Monastiria n level).Th e originalvege tation on this terrace is largely destroyed at present, but indications are that it was savannah-forest. The soil isa Ground Water Humus Podzol (cf.Profil e 47). The savannahs and savannah-forests east of the lower Tocantins are indicated on Appendix 7. HEINSDIJK (1958b), who describes the region concerned as 'very slightly undulating,5 t o 10m abov eloca lrive rlevel' ,report stha tth epatche swit hsavanna h or savannah-forest directlyalongsid eth e Mojii riverhav ea whit esand ysoi lsurface .Wes t ofth e river Mojii, and to alesse rexten teas t of it,larg e areas inth ewatershe d regions arecovere d with 'grassand/o r shrubs'an d havea whit esand y soilsurface . Onesuc ha plain withgrasses ,som eshrub san d palmlets,locate d 2k m easto f Curucumbaba,wa s studied by a pedologist. It proved to be submerged by rain water during a part of the year, and to have a Ground Water Laterite soil, Low phase profile (SAMPAIO, field- notes).

*)Apar t from these, Grey Hydromorphic soil occurs on the terrains concerned (DAY and BENNEMA, 1958).

216 Foto24 Savanasabertas de relva(campos) que ocorrem enlre afloresta alta empedacos de extensao varidvel. Estasat ana i deorigem eddfica. Osolo i dedrenagem imperfeita, comuma cannula superior de areia alvejada e urn subsolo denso e argilosode mosqueadosvermelhos prominentesnuma matrix branca.Solo Later ila Hidromorfica,fase Baixa (area Caeli-Maracassume; l°40' S.,45° .52'0.fotografia TH. DAV)

. Sfv-V:, 4>.'** • ri"i

..«• .. r*V ;. '

/Vioto2- / Open ^raw savannah (campo), occurringin patches of varyingextent between high forest. Thissavannah is of edaphic origin. Thesoil is imperfectlydrained, witha bleached sandy topsoil, and a denseclayey subsoilwith prominent red mottles in a whitematrix: Ground WaterLaterite soil,Low phase (Caeti - Maracassumiarea; I°.40'S., 45°.52' W.photo TH. DAY,)

All savannahs and savannah-forests between the lower Tocantins and the Bahia de Pracui are mapped on Appendix 7. It can be seen that most of them are located in watershed parts, while small patches may occur alongside the rivers. HEINSDIJK (1958a) notes that between the Tocantins and the Camaraipi the land is 'flat, only slightlyabov erive rlevel' ,an d partially submerged duringa part ofth eyear .Betwee n theCamaraip ian dth eBahi ad ePracuf ,th elan d is'slightl yundulating ,maximall y5 t o 10 m above local river level'. The vegetation of the savannahs varies, according to HEINSDIJK,fro m nearlybar esan dwit hher ean dther esom egrass ,t oa mixtur eo fgrass , shruban dpatche so ftrees .Th esoi lsurfac e underbot hcoverage si softe n whitesandy . The area around Anauera river (2°30' S; 49°45' W) was visited by DAY (C/. notes in HEINSDIJK, 1958a). He reports that in this area practically all savannahs and thesur rounding bands of savannah-forest havesoil swit h a bleached,whit e sandytop .The y weremois t inthei r upper part at thetim e of examination,an d often wet or saturated within 1 mdepth . DAYcalle d thesoil spreliminar y'Whit eSands' .Althoug hhi s profile studieswen t to 1 mdept honly ,on eo fhi sfield profil e descriptionsshow sth epresenc e

217 of soft plinthiticmateria li nth esubsoil .Th esoi lconcerne d istherefor e believed to bea Ground Water Laterite soil, Low phase, the same as was found on the other side of the Tocantins(se eabove) .Thi si sth e more probable becauseo fth e presence of imper fectly drained soils,preliminar y classified as 'Gray Latosols', under nearby average to poor forest. For a strip of shrub savannah, probably a campina, alongside the river Tocantins itself(49°.35'W ;2°.25 'S) ,wher eth eterrai ni sa ta leve lo fabou t 10m abov e the river, SAMPAIO(fiel d notes)describe d aprofil e with moretha n 3 metreso fbleache d white sand: a White Sand Regosol. More upstream, namely near Tucurui, the author studied the soil of an interior large patch of shrubby savannah with palmlets, which occurs on flat terrain (6 km from the river; extension unknown; possibly identical with the 'campina de Breu Branco'of DUCKE and BLACK, 1954).Th e profile consisted of white dry sand to 2m depth, below which grey bands occurred in white wet sand. Probably adee pGroun d Water Podzoli sconcerned . The broad strip of savannahs and savannah-forests on the watershed between the Bahia de Pracui and the lower Xingii also have white sandy topsoils (HEINSDUK, 1958a). In contrast to the situation in the regions discussed above, savannahs or savannah- forests are sparse in the region between the lower Xingu and the lower Tapaj6s. HEINSDUK (1957)report sth eoccurrenc eo fonl ya fe wsmal lpatches ,o nth elowe rpart s ofth esand y'flanco 'terrains .Th esoi lo fthes epatche si s'covere d withpur ewhit esand' . In the western half of the region between the lower Tapajos and the lower Madeira, comprising the areas of forest inventory units Canhuma and Maues, the forest inven tory maps show the occurrence of patches of savannah-forest on the flat watershed terrains. These are 5-8 mabov e local low river water level inth e northern half, 20t o 30 mabov e low river water level in the southern half of the area (HEINSDUK, 1958C). This author's superficial description ofth e soilso fthes epatche ssuggest s that Ground Water Lateritesoi li sinvolved .I nth eregio n directly south-west of theinventor y area, namely around Manicore on the Madeira river,savannah s and savannah-forests are more frequent. Theforme r are mapped in Fig.26 .I t can besee ntha t they areconcen trated on watershed terrains. They partly form large irregular patches and partly elongated stretches.N ofield dat a areavailabl e onthei rvegetativ ecompositio n orthei r soils. A number of the elongated savannah areas have however on AAF (1942) maps the notes 'old streambeds with patches of white sand' or 'white sand with scattered trees'. The very extensive open savannahs (campos), and the surrounding broad fringes of savannah-forest, between the middle courseso fth e Madeira and the Purus (Humaita- Labrea-Porto Velhotriangle) , were studied by BRAUN and RAMOS(1959) .The y occur on flat watershed areas with imperfect drainage. The soil profile descriptions of BRAUN and RAMOSsho w that all subsoils, often already from about 20c m depth on wards, have strong, reddish mottling in a light grey or white matrix. There can be little doubt that thesesubsoils , and the upper section of the underlying materialwhic h isdescribe d asargilas mosqueadas da Forma^ao Barreiras, are soft plinthitic incharac ter.Groun d Water Laterite soils,develope d from relatively heavytexture d parent ma-

218 Fig.26 Savanasnaturals na regiao do baixo rioMadeira. Dos mapas bdsicos da AAF( 1942)

6ft0' W 61° 6

Lowland vegetation | j\ | Rivulet vegetacdo de baixada igarape Village * 1 Shifting cultivation povoacdo rofadas Houses casas

Fig.26 Natural savannahs in the regionof the lower Madeira. From AAF preliminary base maps (1942)

terial, must be involved. That this is the case, is already suggested by BRAUN and RAMOSthemselve si nwritin gabou t'lateritas hidromorficas\ The numerous small savannahs (campinas) on the uplands between the Trombetas and the Rio Negro haveall ,accordin gt o DUCKEan d BLACK (1954),a surfac e ofblac k humus and white sand. The author studied those near Itacoatiara (Photo 25)an des pecially those north of Manaus.The y occura s irregular stripsalon g the rivulets,a t a height of 1 to 5 mabov ethem .A tal lstudie d sitesth esoi l profileconstitute sa Groun d

219 Water Podzol with a well defined Ortstein. This Ortstein occurs sometimes at great depth, givingth e soilth eappearanc e ofa White Sand Regosol. As to the extensive natural clearings south of Barcelos on the Rio Negro, no field data exist.Thei r largeexten t and their location on supposedlyflat terrai n suggest that Ground Water Lateritesoi lpredominates . The soil of the elongated patches of savannah-forest, known ascaating a amazonica, ofth euppe r RioNegr oha sbee nstudie db y VIEIRA and FILHO(1961) .I t hasa bleache d sandy top and an Ortstein. The mentioned authors suggest that former riverbeds are involved, on the bottoms of which the Ortstein developed before the beds were filled with white sand. They therefore classify the soils as Regosol. In the author's opinion however, real Ground Water Podzol is involved. SIOLI and KLINGE (1961) collected 'Podzol' profiles near Sao Paulo de Oliven^a, on the Solimoes near Peru. These pro files are presumably located onsite swit h acove ro f caatinga amazonica, and the Pod zoli sprobabl y aGroun d WaterPodzol . Tiny patches of savannah-forest within normal high forest may occur on relatively high upland ina freel y drainingposition .O nthes esites ,th esoi lconsist softe n ofbleach ed sand deeper than augering reached (3m depth) .The y seemt o bewidel y distributed on the sandy parts of the Planicie, including the north-eastern section of it. Such pat cheswer e seen,fo r instance, in the Manaus-Itacoatiara area, at 10k m north of Oro- ximina{cf. Profile 48),an da tPort o Platon inAmap aTerritor y (for thelatte rcf. DAY'S description in PITT, 1961).

Itma yb ementione d inpassing ,tha t inth ebleache d A2horizo n ofth eGroun d Water Laterite soil, Low phase,th e formation of a secondary B horizon or even of an Ort stein,jus t above the soft plinthite, may take place. This phenomenon is reported by DAY (1959) for the Caete-Maracassume area. The author observed it on the low up landso fsouth-easter n Marajo island. Inth elatte r area,ther ear eindication s that such secondary profile development has taken place only after clearing of the naturalvege tative cover {cf Photo 14). Iti sstil lnote d thatth eedaphi ccharacterisatio n ofth ecampin a by DUCKEan d BLACK (1954),a shavin ga topsoi lo fwhit esan dan dblac khumus ,doe sno thol dgoo dgeneral ly.Thi s is because of the above mentioned fact that campos, through the presence of Ground Water Laterite soil, Low phase,ma yals o havea bleached sandy top.

CONCLUSIONS From the above observations it iseviden t that all savannahs and the majority of the savannah-forests of the Planicie, outside the north-eastern belt, are found on terrains of imperfect drainage, with hydromorphic soils. They are, for a part, Ground Water Laterite soil,an d thenofte n the so-called Lowphas eo fthi ssoil ,an d for theothe r part Ground Water (Humus) Podzol. Ground Water Laterite soil seems to be associated predominantly with extensive savannahs and surrounding savannah-forests on water shed areas with flat relief (the campos; a part of the campina-ranas). Ground Water Podzol, in contrast, seems to be predominantly associated with savannahs or savan nah-forests onstrip so fsandy ,relativel y lowuplan dalon gth eriver san d on sand-filled

220 Foto 25 Savana deareia branca comcoberta defetos e palmeiras pequenas (campina),que ocorre em faixas estreilasao longo dearroios em area que de resto 4florestada. Tambem eslasavana e" deorigem eddfica. Por baixo de utna camada de uma espessura de urn metrode areiaalvejada enconlra-se um subsolo duro de castanho escuro homogeneo (Ortstein):Podzol Hidromorfico. Neste lugar a areia alve jadafoi excavadapara fins de construcdo. O Ortstein constitui a base daexcavacao (AM-I, km 10m. oum. de Ilacoatiara)

Photo 25 White sand savannah with a cover offerns and palmlets(campina), occuring asnarrow strips along rivulets inotherwise forested area. This savannah isalso ofedaphic origin.Below a layer, about one metrethick, of bleached sand, ahomogeneous dark brown hardpan (Ortstein) occurs: Ground WaterPodzol. The bleachedsand at this spotis excavated for constructionpurposes. The Ortsteinforms thefloor of the excavation (AM-I, 10km ca.from Ilacoatiara)

formerriverbeds(th ecampinas ;a par to fth ecampina-ranas ;th ecaatinga samazonicas) . Some patches of savannah-forest occur on relatively high, freely draining terrains, where the soil isdeepl y white sandy. Such profiles are called White Sand Regosol, al though most ofthe m actually seem to havebee n very deeply and intensively bleached in situ, and 'Giant Podzol' therefore might be a more adequate classification (cf. III.3.4). The caatinga amazonica and the campina apparently occur on comparable types of terrain and have identical soils.I t is likely that the difference in composition of the vegetativecover s isdu et oa differenc e inclimate .Caating a amazonica isreporte d only for thenorthwester n parto f Amazonia,wit h Af climate;campinas ,o nth eothe r hand, occur apparently throughout the other parts of Amazonia, which have Am climate predominantly(cf. Fig. 2).

221 IV.2.2.2 Savannahs andSavannah-Forests outside thePlanicie Forest encircled savannahs or savannah-forests outside the area of the Planicie are notuncommon .The yar eth efollowin g(cf. 1.5.2): 1. the savannahs and savannah-forests on the Brazilian shield at about 7° Slatitude . They arefoun d alongth elowe r Araguaia,alon gth e Xingu inth eregio n ofth e conflu enceo fth e Rio Fresco,an d alongth eTapajo s inth e region,o fth ejunction ofth e Sao Manuel and the Juruena (Campos de Cururu, Campos de Mucajazal). 2. The savannahs and savannah-forests on the Guiana shield, between the rivers Trombetas and Jari up to the frontier with the Guianas.T o these belongth e Campos deAriramba ,th e Campos geraisd e6bidos , and thecampo s ofth euppe r Paru.

Very little is known about the climatic conditions in the regions of these savannahs and savannah-forests. Local variations in climate are not to be expected, since pro nounced differences in topography do not occur; the terrains concerned are at maxi mally 600 m altitude, and ore often practically flat.1 It can therefore be said that the existence of the savannahs and savannah-forests under discussion is not due to local adverseclimati cconditions .I t may,however ,b etha t thegenera lclimat e inth eregion s involved has contributed to their extent. This may apply in particular to parts of the Guiana shield where the rainfall is believed to be comparatively unfavourable (cf. Figs.2 ,3 and4) . That anthropogene factors acted as the prime cause is unlikely, because the regions concerned are far from the navigable waterways, whereth e former Indian tribes were concentrated (cf. III.3.4). Human influence may,however ,hav ehelpe d locally inesta blishingth e present day composition and extent of the savannah-forests and particu larlyo fth esavannah sconcerned.SIOL Ian d KLINGE(1961),fo r instance,refer toactivit y of Indiansi nth eare ao fth eCampo sd eCururu . Asregard sth esavannah so fth euppe r Paru river, the data of DOST(1962 )ar e of interest. He describes the Sipaliwine savan nahs,whic har eth econtinuation , inth e south-western point of Surinam, of the upper Paru savannahs. DOSTfoun d artefacts offorme r Indianoccupation , and notedtha tth e terrainsar estil lfrequentl y burned. While little can be said with certainty concerning the climatic conditions and the degree of anthropogene influence, the samei s true for the edaphic conditions of the majority of the savannahs and the savannah-forests outside the Planicie. Definite soil data are only available for the forest encircled savannahs or savannah-forests along the lower Araguaia, viz. the campo, the campocom arbusto,th e arbustoan d the floresta comarbusto west of the 48°. 10' W. latitude, as described and mapped by GLERUM and SMIT(1962b) .A s isshow n on Appendices 2an d 6,the y occur largely on undulatingt omountainou sterrain ,wher eth esoil sar eLithosol so fon etyp eo ranother , viz.th e mapping units Lithosol, Quartzite substratum phase (L, QU), Lithosol, Sand stonesubstratu m phase(L ,ss )an d Lithosol, Cherty substratum phase(L , CH). A part of the savannah-forests is found onflat an d imperfectly drained terrain, namely in the

') Except for a part of the lower Araguaia savannahs.

222 area of mapping unit Hydromorphic Grey Podzolic soil, with high base saturation, Shallow Phase(HP A6,s). Theexten to fth esavannah salon gth emiddl ecourse so fth eXing u andTapaj6 s isun known. In part, even their existence is doubtful. Their geographic position suggests they are located on an EarlyTertiar y peneplanation surface{cf. 1.4.1). Their soilsar e therefore apt to be largely Lithosols and imperfectly drained soils (Ground Water Laterite soils,Groun d Water Podzols),bu t this is little moretha n a guess. SIOLI and KLINGE(1961 )collecte d 'Podzol'profile s from campo along the Cururii (a small river somewhat north of theSa o Manuel).Thi scamp oform s probably part of the extensive savannah areai n thisregion .Description s of the profiles, of their position, and of the degreet owhic hthe yar erepresentative ,ar ehoweve rno tgive nb yth ementione dauthors . Nofield dat a exist as to the extensive savannahs on the Guiana shield. They are be lieved to belocate d for the main part on a peneplanation surface supposedly of Early Tertiary age. KATZER (1903)describe d the region of the upper Trombetas-upper Paru as flat plateau land with numerous lakes.1I t seems likely that the main soils on such extensive,flat, an d imperfectly drained terrainswit ha savanna h coverage,ar e Ground Water Laterite soils of one type or another. DOST(persona l communication) studied soil profiles ofth epar t ofthes esavanna h areaswhic h islocate d in Surinam (Sipaliwi- nesavannahs) .Th e profiles areeithe r Lithosolic, or show signso f imperfect drainage, withstrong ,reddis h mottlingi nth esubsoil .Groun d Water Lateritesoi lan d intergrad- ings of thissoi l to Lithosol are believed to be involved. GUPPY (1958,relate d by HEY- LIGERS, 1963) sawa small ,isolate d patcho fsavanna h inth eregio n ofth euppe r Trom- betasnea rSerr a Irikoume\ whichha da whit esand ytopsoil . Insummarisin gthes efe windications ,i tseem stha tth eforest-encircle d savannahsan d savannah-forests on the Brazilian and Guiana shieldshav e predominantly Lithosol or Hydromorphic soilsa sa substratum.

IV.3 The Lowlands with Forest Cover

The relationships between lowland forests and lowland soils are often distinct. Soil conditions on the lowlands are determined directly by the character of flooding or submergence.Ther ear eman yvariation s inth echemica lcompositio n ofth ewater ,an d thequantit y andqualit y ofth eminera l materiali nsuspensio n(dguapreta, dgua branca, dgua limpa, cf. 1.4.4).Th e frequency, length and depth offlooding var y from place to place (igapo;vdrzea da chuva, -do rio, -do mare, -do mar, cf.1.5.1.1) .Fo r the influence of these variations on the composition of the forest coverage please refer to the short

*)Quote : 'eine Hochebene die von zahlreichenSeen und Lagunen bedeckt wird und so wenigausgesprochene Abdachungen besilzt dass zur Regenzeit Verbindungen zwischen den nach Norden abfiiessenden guyanischen und den naeh Suden abfliessenden brasilischen Fliissen bestehen' (KATZER 1903,p.2) . It is in the region of the upper Paru - upper Trombetas that the early maps of South America show the presence of a huge lake or interior sea, as well as of the legendary El Dorado and the tribe of the Amazonas.

223 Foto 26 Um dos tiposde vegetacao queocorrem nasfaixas estreitasde baixada permanentemente mal drenada, aolongo de arroios em areas de terra firme (igapo). Nestetipo predominant o Guarumd (Ischino - siphon aruma),a alta erva no/undo, e aPaxiuba (Iriarti aexorrhiza) ,a palmeira com as raizes esliradas a direita.O solo e esponjoso e a sua camada superior turfosa: SoloMeio-Turfoso (BR-14,km 110m. ou m.j

Photo26 One of thetypes of vegetation thatoccur onthe permanently poorly drainednarrow stripsof lowlandalong rivulets inupland areas (igapo). In this type, Guarumd((Ischinosipho n aruma), thetall herb at the back, and Paxiuba (Iriartiaexorrhiza) , thepalm withthe stilt-roots onthe right, arepre dominant. The soilis spongy, andhas apeaty top: HalfBog soil(BR-14, km 110 ca.) description ofth elowlan d forests in 1.5.1.1,whic hi sbase d largelyo n noteso f DUCKE and BLACK(1954) .N o forest inventorieswer eexecute d in the lowlands,excep tfo r an area in the lower Tocantins river (GLERUM, 1962;are a 25o f Fig. 22).Thi s inventory had as a specific purpose to determine the quantity and the growing site of the Ucuubabranca (Virolasurinamensis). Thetimbe r ofthi stre especie si suse d in theply wood industry, and thefloating seed s are collected to serve as a raw material for the localsoa pindustry .Th emajorit y of the Ucuuba treeso f thevarzea si nth estudie dare a werefoun d on theolde r islandsi nth eriver ,betwee n Baiaoan d Curu£umbaba.Thes e islands, which have Low Humic Gley soil, areflooded t o a shallow depth with river water(dgua branca) inth erain yseason ,an dth eactio no ftide scause sa dail yvariatio n inwate rleve lo f 1 to2 m (vdrzea do mare). Buriti palms(Mauritia flexuosa) ar ei nthi s areaa nassociat eo fth especies . Ucuuba ishoweve rals ofoun d inigap 6 stretches,wher e thesoi li softe n peaty(Bo gsoi lo r Half Bogsoil) .Dens estand so fth especie swer esee n in the lowlands alongth e lower courses of thetributarie s ofth e river Capim that are

224 crossed by the BR14-highway. These lowlands are intermediate between varzea and igapo (meio-igapd) and their soil was classified as Humic Gley soil, intergrade to Ground WaterPodzol . In many of the lowland forests, palms constitute a prominent feature. Agai(Euterpe oleracea), Buriti(Mauritiaflexuosa), Ubucu(Manicaria saccifera) andPaxiuba (Iriar- tiaexorrhiza) aresom eo fth epalm sthat , aloneo ri nvaryin gcombinations , givesigni ficant appearancest olowlan dforest s(cf. Photo2 6an d27) .

IV.4 The Lowlands with Savannaho r Savannah-Forest Cover

SUTMOLLERet al. (1964)giv eman y data ast o the soilsan d thevegetatio n on thelow landswit hsavanna h orsavannah-fores t coveragei neaster nAmazonia .Therefor e only a fewaspect swil lb ediscusse dhere .

IV.4.1 TheLowlan d Savannahs ofth eLowe rAmazo n Region

Many discussions have been made ast o thecaus e ofth e extensivenatura l floodplain pastures (campos de varzea dorio) alon g the lower Amazon river between Parentins and the mouth ofth eXing u river(cf. SIOLI, 1956).Bot h upstream and downstream of thisstretc hther ei sa fores t coverage,althoug hth esoil si nth ewhol estretc h aresimila r when classified to no great detail, namely Low Humic Gley soil or Humic Gley soil. DUCKE and BLACK (1954) suggest that the comparatively dry climate in the stretch concerned, accounts for the savannah coverage. The author is, however, convinced that itspresenc ei sdetermine d onlyb ya nadvers elengt han d deptho fannua l flooding; thusb yedaphi cfactors . The savannahs are'hydrological ' savannahs. Upstream ofth e savannah stretch, the annualflooding i sno t longenough , and downstream theflood ing isno t deep enough to impede forest growth. A strong argument for this point of viewi sth efac t that on the highest parts ofth efloodplain i nth e savannah stretch (the leveesan d point bars)th evegetatio n issavannah-fores t or forest. Nextt oth esavannah so nlowland sprope ri nth eLowe rAmazo n region,ther ear eals o savannahs on terrains that are slightly above the normal high water level, namely in theare abetwee nOroximin aan dFar o(Terr a Santa,campo so nEarl yHolocen emassa - peterrains , cf.1.4.4. )Advers e soil conditions (Ground Water Laterite soil; cf. Profile 3)ar e the cause of these savannahs. Frequent burning, however, conditions their present-dayvegetativ ecomposition .

IV.4. 2 The Lowland Savannahs of Eastern Marajo Island

Thereca nb elittl edoub ttha tth eextensiv enatura l grasslandso nth elowlan d partso f eastern Maraj6 (campos devarzea dachuva) are of edaphic origin. Patches of sandy

225 Fote 27 Acaizal.Neste tipo de igapo a pecaconstituinte quaseunica de vegetacio & a palmeira Acai (Euterpe oleracea),que cresce em gruposcompactos. Umacamada defolhas palmeiras cobre o solo limido e esponjoso, quee" profundamente turfoso(Solo Turfoso)

Photo 27 Acaizal.In thistype of igapothe nearly sole component of thevegetation is thepalm Acai (Euterpe oleracea), heregrowing in bigclumps. A layerof dead palm leaves covers thewet, spongysoil, which ispeaty toa considerable depth(Bog soil) low uplands (tesos) within the area of the lowlands have a forest coverage, when its vegetativecove ri sno tartificiall y altered. These lowlands, of presumably Early Holocene age, are submerged with rain water during several months of the year, in some parts up to 2 m in depth. The soils are heavy textured and havea bad structure. In part, they have an adversechemica l com position,Na + and Mg++ beingpredominan t onth eexchang ecomplex .Th esoil swit ha predominance of Na+ and Mg++,whic h often are the most deeply submerged by rain water (Solonetz, Coastal phase; cf. Profile 52;solonetzi c Humic Gley soil, intergrade toGroun d Water Lateritesoil ,cf. Profil e 18), lackeve nshrub si nthei rvegetativ ecove r (campolimpo). The soils with more regular ratios of the exchangeable cations and shallow submergence (Ground Water Laterite soil, heavy textured phase, cf. Profile 11) have a coverage of savannah in which shrubs have a scattered occurrence, and Buritipalm s(Mauritiaflexuosa) mayb epresen t infairl y highquantities . A situation similart o that ineaster n Marajo, prevails in at least a part ofth e natural grasslands onth elowland si neaster nAmap aTerritory .

226 V Chemical and Physical Qualities of theMai nAmazo nSoils , andthei rAgricultura l Occupation

INTRODUCTION The Federal Republic of Brazil, seen as a whole, still has vast areas of virgin land available for its rapidly growing population. This population has hitherto been con centrated in some south-eastern States and in a broad strip along the Atlantic sea board. With the foundation of the new capital Brasilia, in the interior, and the con struction of highways radiating from thiscapita l inal ldirections ,condition s are being created to permit the occupation ofth e interior virgin lands.A t present, several sche mesfo r suchoccupatio n arebein gdrafte d byfedera l and regionaldevelopmen t boards. Particularly urgent isth e procuring of new land for a part of the rural population of theNorth-Easter nregion ,wher eperiodicall yrecurrin gsever edrought san dfloods ar ea serious drawback for sound rural development. Among the outlet areas for the popu lation of this region, parts of Amazonia are being given much consideration. It is- therefore of immediate importance to evaluate the capabilities ofth eAmazo n soils. Atman yagricultura l experiment stationsi n Brazil,trial swer ean d still are being exe cuted on new agricultural methods, such as the application of chemical fertilizers, animal,farm-yard , or green manures,an d the planting ofcove r crops.Th e manner of implementation however,an d theeconomic so fsuc hmethod sar estil llargel y unknown for many of the Brazilian soils. In some areas, modern techniques are already being largely applied by the farmers, for instance in many parts of Sao Paulo State. But for a large part of the rural population, the stage of development as an agricultural community is still relatively low. This many apply particularly to that section of the Brazilian people which ispredominantl y of Indian descent, as isth e case with the in habitantso fAmazonia .Thi si sbecaus eth eAmer-Indian swere ,an d stillare ,essentiall y hunters,fishermen an d foragers. In many parts of Brazil, chemical fertilizers are ex pensive,due ,amon g other reasons,t oth ecost so ftransportin g them overth evas tdis tances.Anima l manurei sscars eowin gt oth egrazin go fcattl eove rwide ,ofte n unfenc- edareas . For these reasons, the present-day agriculture of Brazil is found predominantly on soilswit h high natural fertility (cf. BARROS, DRUMOND, CAMARGOet al., 1958 ; LEMOS„ BENNEMA, SANTOSet al., 1960) .Th e shifting cultivation system isalso ,fo r that matter,, essentially autilizatio n oftemporaril y high'natural ' fertility. Invie wo fth eabove ,i ti sunderstandabl e that, inselectin gne wsettlemen t areas,soil s witha hig h natural fertility aremuc h more sought after byindividua lpionee r farmers,, than soils that have a potentially high fertility owing to good physical qualities. For large-scale settlement schemes, however, the expanse, the homogenity and th&

227 Foto 28 Arroz de terra firme cultivadoem solode 'plinthite' fossil. Mais da meiade do material de solo4 constituida depedras de 'plinthite'' fossil. Mesmo assim colheitas comumente sdosatisfa- toriasnestes solos, visto que o 'plinthite'so excepcionalmente constitui urn carapacoimpenetravel e a terra propria provavelmentei um pouco mats rica que aquelede solos compardveis sent 'plinthite' {Faienda Oriboca, perto de Belemj

Photo28 Drylandrice growing on soil with fossil plinthite. More than half of thesoil material consists of stones offossil plinthite. Cropsare nevertheless doing wellon these soils,since theplinthite rarely forms an impenetrable capand theinterjacent earth isprobably slightly richer than that ofcomparable non-plinthiticsoils (Oriboca Estate, near Belim) means of access to the area to be selected are of importance, while socio-political factors areals oinvolved . In Amazonia, soilswit hrelativel y high natural fertility arefoun d on partso fth elow lands, and locally on the uplands outside the Planicie. Generally speaking, such soils are therefore sparse, often small in area, of difficult accessability, or expensive to re claim.Th eterrain s ofth e Planicie,o n the other hand, constitute largetract s ofpredo minantly flat or gently undulating land which is largely in a freely draining position. Thesetract so flan dar enea rth emai nwaterway san dth eexistin groads .Th eforest s on theseterrain s have supposedly thegreate r grosstimbe rvolume s ofAmazonia , and the present-day Amazon agriculture is concentrated here. It isfo r these reasons that the Planicie part of Amazonia, although the natural fertility of itssoil s islow , showspro misefo r large-scale settlement schemes.Specia l attention will therefore be devoted to the chemical and physical qualities, the adequate soil management measures and the agriculturalcapabilitie s ofth e Planiciesoil swit h an unhindered externaldrainage .

228 V.l TheSoil so f the Lowlands

Asdescribe d in HI.3.5, the soils ofth e terrains with Holocene deposits are rather di verse. The igapo soils(Bo g soil, Half Bogsoi l a.o.) are more or less peaty, of spongy consistency,an d usuallyver yacid .The y occuri nnarro w stretchesan d are permanent lypoorl ydrained .Thes epropertie s maketh esoil sgenerall yunsuitabl efo r agriculture, or for cattle herding. Silviculture adapted to these soil conditions may be the most appropriate, cultivating a valuable oil bearing and/or timber producing tree, such as the Ucuuba (Virola surinamensis). The same maybe said of the bottom lands within grass-covered uplands,althoug h here,cattl eherdin gma yb ea feasibl e proposition too. The varzea soils (Low Humic Gley soil, Humic Gley soil, a.o.) are generally non- peaty, heavy textured soils. The natural fertility varies considerably. The cation ex changecapacit y isusuall y much highertha n that ofth e surrounding upland soils, but the base saturation is low, to very low, on many sites and considerable amounts of aluminum are'ofte n present{cf. pag e236) .A portio no f thevarze asoil si ssalin eand/o r has a very unfortunate predominance of Na+ and Mg"^ in the subsoil. The structure of the soils is often unfavourable; heavy textured, sticky and compact subsoils are rather frequent. After the second worldwar ,becaus ewit h some cropsan d with cattleherdin g favour able results were obtained on some varzea areas, Government agencies paid much attention to promoting varzea agriculture and grazing (CAMARGO, 1950; LIMA, 1956). In this respect may be mentioned the growing ofjut e in the Lower Amazon region, of rice and sugar cane in the region west of Belem, and of cocoa along the lower Tocantins river.Als onotabl e isth eherdin go fcattl ean d buffalos inth eLowe rAmazo n region and inth eeaster n part ofMaraj o island. Erroneously,a popula r impression has grown that all of the varzeas are much more favourable for occupation than the surrounding uplands. Actually, land capabilities vary much with the type of varzea. The main criteria for its evaluation will have to beth e chemical qualities of the soils, {c.q. the richness of the flooding water), and the length and depth offlooding. Excep t on narrow, relatively high stretches (varzeas alias,restingas) and for some especially adapted crops such asjute , a large scale,al l year cultivation on the varzeas of eastern Amazonia ispossibl e onlyafte r artificial drainage and reclamation.Thi s requirescon siderable investigation, organization and capital investment. Favourable conditions for pasturing may be obtained more easily and in some parts already exist naturally {cf. SUTMOLLERet al., 1964).

V.2 TheSoil so fth eUpland sOutsid eth e Planicie

The uplands outside the Planicie are the undulating terrains of outcropping crystal linebasemen t and Paleozoic, Mesozoico r EarlyTertiar y deposits,a swel la sth epene - planation areas ofCretaceou s or Early Tertiary age.Thes eterrain s are largely located far from the population centres.Thei r accessibility isdifficult , due to the rough topo-

229 graphy, the occurrence of rapids in the rivers, and the presence of dangerous wild Indians. These are reasons why hitherto little agricultural use has been made of the land concerned. Asdiscusse d in III.3.1, III.3.2 and III.3.3, the soils are, according to indications, very diverse. They vary much in natural fertility, in texture, structure, depth, stoniness and in drainage condition. Each individual soil is probably present either rather scattered or incomplexes ,an d the reliefca n bebroken . Onlya fe w parts, •whichar e ofcomparativel y easyaccessibility , areoccupied . Examples areth e Alenqu- er-Monte Alegrearea , Fordlandia on the lower Tapajos river, and somepart s around Rio Branco do Acre.Thes e are all areas where soils occur with a comparatively high natural fertility. The principal criterion for land capability evaluations ofth e vast unoccupied expan ses must betha t of high natural fertility of the soils.Thi s isowin g to the high cost of transportation, which make it necessary to obtain maximum returns from the mini mumo f effort.

V.3 The Freely Draining Kaolinitic Soilso f the Planicie

The terrains ofth e Plio-PleistoceneAmazo n planalto and ofth e Pleistoceneterraces , together called Planicie(c/ .1.4) , arelargel y ina positio n offre e drainage. Comparativ ely many and consistent data are available ast o the soils on these freely draining ter rains.Th esoil sconcerne d are:

KaoliniticYello wLatoso l(,Ortho ) KYL KaoliniticYello wLatosol ,Compac t phase KYL, c KaoliniticYello wLatosol ,Concretionar y phase KYL, CR KaoliniticYello wLatosol ,intergrad et o Dark Horizon Latosol KYL-DHL KaoliniticYello wLatosol ,intergrad et o RedYello wPodzoli c soil KYL-RP Kaolinitic Red Latosol KRL Kaolinitic LatosolicSan d KLS RedYello wPodzoli csoil ,intergrad et o KaoliniticYello wLato sol RP-KYL RedYello wPodzoli csoil ,intergrad et o Kaolinitic YellowLato sol,Concretionar y phase RP-KYL,C R (For textureclasse scf. Table9) .

Thechemica lcharacteristic so fthes esoil sar esimila rt o a large extent.Th esoil s have incommo n atota lo rnearl ytota labsence ,eve ni nth edeepe rsubsoil ,o fprimar ymine ralstha t areeasil yweatherable ,whic hwoul d function asa nutrien t reserve.Th ecatio n exchange capacities are small,an d the basesaturatio n percentage ispracticall y always low.Th ecla yfractio n isstrongl y kaolinitic;th eK ivalue sar enormall y between 1.8an d

230 2.0, only sometimes as low as 1.5 or as high as 2.3; the Kr values are normally between 1.5 and 1.8,onl ysometime sa slo wa s1. 4o ra shig ha s2.0 . Together, thesoil s willb ecalle dfreely drainingkaolinitic Planicie soils.

V.3.1 Chemical and Physical Qualities of the Freely Draining Kaolinitic PlanicieSoil s

In the following, the chemical and physical qualities will be discussed of the freely draining kaolinitic Planicie soils, under their natural vegetative cover or varying degree ofhuma n influence. It islargel y acompilin g ofal lrelevan tfield an d laboratory data availablea tpresent ,a st oenabl ea tentativeassessmen t to be made ofth eagricul tural potential ofth e soilsan d their adequate management, and to provide a basis for later, detailed research. For this purpose, extreme and mean valuesar e given for each soilcomponent , and thecorrelatio n withothe rcomponent s isshown ,partl yi ngraphs . The causal factors for these correlations are not discussed. Such a discussion would not well fit in the context of the present chapter. The available data are, moreover, often tooelementar y for that purpose.

V.3.1.1 TheSoils under Primeval Forest Cover CHEMICAL QUALITIES For study of the chemical qualities of the freely draining kaolinitic Planicie soils un der primeval forest cover, IQA analytical data of 35 relevant profiles are available. Half ofthe m are of the Guama-Imperatriz area. The others are from places scattered overth ePlanici ei neaster n Amazonia. Organic Matter. Analysis shows that the easily available plant nutrients are highly concentrated ina thi n superficial layero fth esoi lwhic hcontain sth ebul k ofth eorgan icmatter . For all horizons of the profiles there is a tendency for the percentage of organic mattert oincreas ea ta highe rpercentag e ofclay . Thisi sillustrate d inth e Figs.27,28 ,an d 29(th edat a for Figs. 27an d2 8wer eobtain edmainl yb ygraphica linterpolatio n from theanalytica ldat ao fth ehorizon sinvolved) . Theincreas eo fpercentag eo fCarbo n with increasei npercentag eo fcla yi sgreates ti n the upper part ofth e profile. At 10 cm depth, for instance(Fig .27) ,th e increasecom prises roughly 0.16% Carbon per 10%clay . The variations, at equal percentages of clay are, however, fairly large. This is because (1) the graphical interpolation of the data doesno t givea naccurat e picture of theactua l situation inth e soil,(2 )th esampl ing of the upper, sometimes very thin horizon was often carried out inaccurately, (3) nomixe dsampling ,t ocompensat efo r thehorizonta lvariation si norgani cmatter ,wa s made and(4 )th edispersio n of the clay fraction on analysis may not have been com plete, due to the comparatively high organic matter content. Apart from these, the variation isals ocause d bydifference s inorgani cmatte rconten t whichar edu edirectl y toth ecompositio n ofth efores t coverage(cf. cipoal ,IV.1.2) .

231 I T Fig. 27 Relacao entre a 10cm percentagemde Carbono e apercentagem deargila a 10 cm de profundidade, emperfis de solos caolini- ticosdaPlanicie de drena- h gemlivre

2 --

1 •--

• forest profile/per/// de mala Fig.27 Relation between X savannah profile//«r/i/ de campo percentage of Carbonand f with plinthite/com 'plinthile' percentageof clay at 10 cm depth,in profiles of free ly -+- —I 1 20 40 60 eo 100% drainingkaolinitic Plani- > Clay/Argila cie soils

• forest profWclperftl de mala Fig.28 Relacao entre a I 1 T |100cm| x savannah profile/perftl decampo percentagemde Carbono e % apercentagem deargila a a r with plinthite/com 'plinthile' 100 cm deprofundidade, emperfis de solos caolini- ticosda Planicie de drena- 0.5-• gemlivre

x* Fig.28 Relation between percentageof Carbon and percentageof clay at 100 cm depth, in profiles of freely draining kaolinitic 20 40 60 80 100% » Clay/Argila Planiciesoils

i- • forest profilejperftl de mala Fig.29 Relacao entre a e percentagemde Carbono e Xsavanna h profile/y^r/7/ die campo -|% • apercentagem de argila no C Q8 f with plinthite/com 'pUnthite* • horizonte B, de perfis de • | •• solos caoliniticos da Pla 5 06 niciede drenagem livre • • • • 0.4--

. x •* Fig. 29 Relation between 0.2-• . . X. I percentage of Carbon and \ x • percentage of clay in the 1 1 1 1 . 1— -.—i—.—i Bt horizon, of profiles o 20 40 60 80 100% freely draining kaolinitic > Clay/Argila Planicie soils

232 Fig. 30 Relacaoentre a perceniagem deCarbono e o valor Tparaperfis desolos caolinlticosda Planlcie dedrenagem litre

mg/IOOg 20 n K .a is- • Q £ 1-16- • forestprofile Iperfil de mala • s x savannah profile/perjil de campo • > H- f with plinthite/com 'plinlhite' ' V 0 12 - • • • • • 10- • • / 8- • • • • • • •• • X 6- A t . • • * *. x . • :••..• * ' x* 4- - • • t *

-2! . 1 . 1— 1 1 1 2 3 °/o ™—•»Carbon/Corion o irjjp.J 0 Relation between percentage of Carbon andvalue T,for profiles offreely draining kaolinitic Planiciesoils

m.e./IOOg Fig. 31 Relacao entre a * forest profileIperfil de mala perceniagem deargila eo valorT nos horizontes B, •1 1 X savannah propel perfilde campo ^6- deperfis de solos caolinl Twilh plinthite/rom 'plinlhite' ticosda Planicie de drena gem livre 4-- ** 2- *4 Fig. 31 Relation between percentage of clay and value T,in B t horizons of 20 40 60 80 100% profilesof freely draining E>CfayjArgila kaolinitic Planicie soils

Probably morereliable ,an d anywaymor eclea r isth etren d at somedept h inth epro file,fo r instance at 100c m (Fig. 28). The increase of percentage of Carbon with in crease in percentage of clayi sher emuc h slower,namel y roughly 0.035% Carbo n per 10%clay .Th e small increase isdue , in part, to thefac t that in sandy soilsth e percen tage of organic matter decreases more gradually with increasing depth than in clayey soils. The former have generally deeper profiles, which shows up in the depth of the horizon of maximal clay content, i.e. the B2 horizon. More interesting therefore, in somerespect , isth esituatio n for the B2horizon s ofth eprofile s (Fig.29) , where thein creasei sroughl y0.0 6% Carbo n per 10% clay .

233 Table14 Cationexchange complex offreely drainingkaolinitic Planicie soils,under primeval forest cover (18 profiles;Guamd lmperatriz area)* orunder anthropogenic savannah (6profiles; Amapd Territory)

T Hori Potential cation Exchangeablemetalli ccations ,i n %o fT zon exchangecapacit y basestrocdveis, em"/ adeT hori- capicidadetotal zonte de troca Ca++ Mg++ K+ Na+ m.e./lOOg)

FORESTPROFiLES/per/ w demala Aj (3.9-14.8)' 8.6(3.0-23.0)' 8.0(3.0-17.2)' 1.6(0.9-2.6)' 0.7^0.4-1.9)' A.-A, (2.2-6.9) 6.8(3.5-16.8)** 6.8(3.5-16.8)** 2.4(1.2-6.3) 1.0(0.5-3.0) B (1.4-4.6) 9.2(5.0-20.0)** 9.2(5.0-20.0)** 2.9(1.0-6.2) 1.6(0.5-5.5) C*** (1.4-3.6) 11.1(6.7-19.5)** 11.1(6.7-19.5)** 3.1(1.9-4.5) 1.5(0.8-2.3)

SAVANNAH PROTiLts/perfis de campo A, (2.2-5.7) 6.7(4.5-10.8)** 6.7(4.5-10.8)** 1.4(0.3-4.3) 0.8(0.3-1.0) A,-A, (2.1-3.5) 7.1(4.5-9.7)** 7.1(4.5-9.7)** 0.9(0.8-1.0) 0.6(0.2-1.0) B (2.1-2.9) 8.1(5.2-10.9)** 8.1(5.2-10.9)** 1.5(0.3-3.5) 0.9(0.4-2.2)

') Datawithi nbrackets :rang eo fvalue s * Kaolinitic Red Latosol profiles not included dadosentre os parenteses: variafoes dos valores Perfis de Latosolo Vermelho Caoliniticonio incluidos Databefor e brackets: main values dados em frente dos parenteses: medios dos valores

In Fig.3 0th e potential cation exchange capacity at pH 7(valu eT ) iscompare d with theorgani cmatte rcontent ,usin gth e relevantdat ao fal lhorizon so fth ethirty-fiv e pro files.Th etren d isver yclear ;th e increase ofT i sabou t 3.9m.e .pe r 1 %Carbon . TheT valueshow sals oa positiv ecorrelatio n withth epercentag e ofclay .I nth etopsoil ,wit h its concentration of the organic matter, the T values is always considerably higher whenth e soili sheavier . For the B2horizon , the increase ofT wit h the increase of per centage of clay isonl y moderately high, as shown in Fig. 31. In horizons with practi cally no organic matter (Chorizons) , T isalway s very low,eve n when the percentage of clay ishigh .Thi s means that practically allth e existingcatio n exchangecapacit y is due to the organic matter in the soil. The chemical activity of the clay-sized mineral particlesthemselve s isver ylo w(a n estimate of the latter may beobtaine d byth ecom bination of the Figs.29 ,30 ,an d 31; the potential cation exchangecapacit y per 100 g pure, i.e.organi c matter-free, clay-sized mineral material, isbetwee n 1.8 and 4.0 m.e. approximately). It isapparen t that with regard to chemical qualities, themain advan tage ofthe clayey over the sandysoils istheir ability tocreate abetter milieu for preser- vationof the comparatively veryactive organic matter. The C/N values,whic h are often an indicator of the type of organic matter, arever y regular in the profiles. In the At horizon they are highest. The average value in this horizon is about 10.5 in the relatively heavy textured profiles. The relatively light textured profiles have an average ratio of 13.5i n their Ax horizons. In the subsurface and subsoilhorizon sther ei sn odifferenc e inaverag eC/ N value betweenth e relatively heavytexture d and therelativel yligh t texturedprofiles . At 50 cmdepth ,th eC/ N value

234 Tabela 14 Complexo de Irocacaiionica de soloscaoliniticos do Planicie de drenagem livre, sob coberturafloreslal primitivaf18 perfis; areaGuamd-lmperatriz)* ousavana antropogenica (6 per/is; Territdrio doAmapd)

S (Al)+ H+ (Al)++H+ S+(A1)+ S Hori Sum ofmet . Active pH-dep. Potential Act.catio n Sumo fmet . zon cations acidity activity acidity exch.cap. cations hori- (%ofT) (%ofT) (%ofT) (%ofT) (%ofT) (%ofS+(Al)+) zonte soma das acidez acidez acidez cap. de somadas bases aiiva pH-depend. polencial Irocaativa bases (%deT) (XdeT) (%deT) (XdeT) (V.deT) (%deS+(Al)+)

FORESTPROFiLES/per/f ade mala A, 19(9-43)' 15(2-24)' 66(50-73)' 81(57-91)' 34(27-50)' 56(28-95)' A,-A, 17(10-37 ) 23(11-37 ) 60(46-76 ) 83(63-90) 40(24-54) 43(21-77) B 23(14-39) 23(6-59) 54(23-76) 77(61-86) 46(28-73) 50(24-87) C*** 27(18-56) 13(3-39) 60(39-74) 73(44-82) 40(28-77) 68(45-95)

SAVANNAH PROFILES//W./?J de campo A, 16(10-24) 21(9-27) 60(41-69) 81(67-90) 37(31-45) 43(29-69 ) A,-A, 16(10-21) 25(21-30) 59(55-63) 84(79-90) 41(38^4 ) 37(24-51) B 19(16-25) 18(10-35) 63(49-68 ) 81(75-84) 37(32-51) 52(31-71)

** t(Ca++ + Mg++) The half of jointly determined bivalent cations Ametade dos cations bivalentes dtterminados em conjunto *** Data of 10profile s only Dados de 10 perfis sdmente

averages9.0 , at 100c m depth8. 0 anda t200c mdept h6.5 . Variationsfro mthes eaverage s are up to 4.0unit. 1Fo r the deeper layers,th evalu e of the ratio as an indicator of the type of organic matter may however be limited, because of the possibility that N-in- mineral-form constitutes there a more than negligible part of the totaldetermine d N (see below).

Exchangeable cations and acidities. Withth eexceptio n of the Kaolinitic Red Latosol profiles, the sumo f the exchangeablecation s(valu eS )comprise sonl ya smallpar t of theT value .Th ebas esaturatio npercentag e(valu eV ) isalway sbelo w 40%. Inth etop - soil(A xhorizons) , thevariatio no f V isbetwee n 5% and 40 %, witha naverag eo f 15 %. In the subsurface soil(A 3o r A2horizons )th e variation isbetwee n 5%an d 35%, with an averageo f 14%. In the subsoil( Bhorizons )th evariatio n of V isbetwee n 10% and 40 %, witha n averageo f 23 %. The deepersubsoi l (C horizons),insofa r it was sampled, showsa variation between 15% and 50%, with an averageo f 29%. In agreement with the low base saturation, the soils show an 'extremely' or 'very strongly' acid reaction (terminology of SOIL SURVEY MANUAL, 1950). The pH-H^O slightly increaseswit h increasingdepth . The variation of this value in the topsoil (A, horizons)i sfro m 3.7t o4.7 ,i nth esubsurfac e soil(A 3o r A2horizons )fro m 3.8t o 5.2,

') KLINGE (1962) gives higher values for Amazon forest soils. In two Braunlehm profiles under high forest, the ratios are as high as 20. Other methods of sample storage and analysis may account for this difference.

235 in the subsoil (B horizons) from 4.0 to 5.4, and in the deeper subsoil (C horizons), insofar sampled,fro m 5.0t o 5.5. ThepH-KC li sgenerall yles stha n oneuni t belowth epH-H 20,namel y onth eaverag e 0.6 inth e upper part ofth e profiles, and onth e average0. 8i nthei r lower part. The cation exchange complex was analysed in detail for the relevant profiles of the Guama-Imperatriz area. The data are given in Table 14.1Becaus e it is suggested by COLEMAN et al. (1959), among others, that base saturation percentages calculated on the active cation exchange capacity are of more value for the study of the chemical qualities of the soil, in relation to the plant growth, than those calculated on the po tential cation exchangecapacity ,th eforme r percentages areals ogiven . In the Aj horizons, the percentage of Ca++ is usually slightly higher than that of Mg++. Topsoils in which Mg++ slightly predominates over Ca++ are however also found. Whether the same applies for the other horizons of the profiles is not certain, because there the bivalent cations were determined collectively. The absolute amount of K+i nth etopsoi lrarel yexceed s0.1 5 m.e./100g o fsoil . It can be seentha t the exchangeable2(Al) +, or the active acidity (cf.III .1) ,i sno t re markably high, especially not soi nth etopsoi l and inth e deeper subsoil.Th e absolute amounts of (Al)+ are between 0.1 and 2.2 m.e./lOOg o f soil, with an average value of 0.7 m.e./lOO g. Mention may be made of the aluminum content of a number of pro files from the centre of the Planicie,whic h wereanalyse d by the Royal Tropical Insti tute of Amsterdam, Holland. In eight profiles, the 'easily available' (Al)+ (extraction with Na-acetate/acetic acid of pH = 4.8i n apast ewit h a soil:solutio n ratio of 1:2.5, shaking time 30 minutes; MORGAN-VENEMA extract) turned out to be nearly al ways less than 100mg/ 1 in the Ax horizons, and on the average about 60mg/ 1 in the B horizons. That (Al)+ values of the freely draining kaolinitic Planicie soils are not remarkably high, appears also when they are compared with the data of a number of other Ama zon soils. For example, acid hydromorphic soils (those without, or with only little, annualenrichmen t bymateria l suspensed inwater ,a sther ear eGroun d Water Laterite soils,a part of the Low Humic Gleyan d the HumicGle y soils) have often 75%(Al) + in their B2ff horizons.Thi s is associated with comparatively large pH-H20-pH-KCl differences (average of 17 profiles; cf. Appendix 9). The 'easily available' (Al)+ is 150-200mg/1 ,i nth esam ehorizo n ofsuc hsoil s(averag eo f7 profiles) .

Nitrogenand Phosphorus. The IQA analysis data as to nitrogen concern only the totalN (N-in-organic-complexes + N-in-mineral-form) in the soil.Th e percentage of total N decreases gradually with increasing depth, in correlation with the decrease of thepercentag e oforgani cmatter .Apar t from this,th epercentag e oftota l Ni sgeneral lylowe rwit hlowe rpercentage so f clayi nth esoil . KLINGE (1962)give s a few data on mineral N in Amazon forest soils,admittin g that

*)See n in their entirety, the profiles of this area seem to have a slightly higher base saturation in their upper horizons than those of the centre proper of the Planicie. *)Wit h the analysis method used, if soluble (Al)+ is present, it is included in the data.

236 the values may be incorrect because of possible reactions during the storage of the samples. In two Braunlehm profiles under forest coverage, the mineral N decreases from about 5m gpe r 100g o f soili nth e topsoil, to about 3m gpe r 100 go f soil in the subsoil. The litter layer contains about 8m g per 100g . The percentage of mineral N compared totota l N increaseshoweve r with increasing depth,namel yfro m about 3% to 8%. With regard to the form of the mineral N, KLINGE gives values for the ratio N-NH4+ :N-NO3 -tha t arebetwee n 8 and 25,wit h 40i nth elitte r layer. For eight forest profiles of freely draining kaolinitic soils of the Planicie, the 'easily available'N wasdetermined , byth e RoyalTropica l Institute ofAmsterdam , Holland, in the MORGAN-VENEMA extract (see before). In the topsoil (0-20 cm; A1 horizons), thevariatio n ofNH 4i s from 1 to 14mg/ 1(averag e7 )an dtha to f NO3from3to30mg/l (average 15).I n the subsurface soil(20-6 0cm ; A3horizons) , the variation of NH4 is from 1 to 12 mg/1(averag e 3.5)an d that ofNO s from 'traces't o 11 mg/1.I nth esubsoi l (60-100cm ; Bhorizons) ,th evariatio n of NH4 is from 'traces' to 12,an d that of N03 from 'traces't o 18 mg/1.I tca n be seentha t therei sa grea tvariation .Thi si sonl yasso ciated with textural differences to a minor extent. Apart from seasonalfluctuations i n the amounts of NH4+an d N03~ in the soil,nitrificatio n and denitrification processes are likely to have occurred in the samples before they were analysed. Such processes + probably have changed, in an uneven manner, the contents of easily available NH4 and N03~a si twa si nth esoil si nthei rnatura l position. It isnote d that, inth e Forest Inventory areas in the Planicie (cf.IV.l) , normally be tween 15 and 30% of the enumerated trees belong to the Leguminosae order (the highest percentages are found in the region between the lower Xingu and the lower Tapajos; cf. GLERUM, 1960). No data are however available as to the effective root nodulation ofthes etrees . No data exist on the relative importance of nitrogen fixation, i.e. the conversion of atmospheric N into acombine d form and itsincorporatio n into the forest soil.Thi si s so, whether non-symbiotic bacteria (Clostridium, Beijerinckia) are concerned, or bac teriatha tar ei nsymbiosis ,vi aroo t nodules,wit hLegumes .

The IQA data on Phosphorus are more interesting. For the thirty-five forest profiles mentioned before, the availableP wa sdetermine d according to the TRUOG-methodo r the BRAY-method, or both (Truog on 25 profiles, often only their upper horizons; Bray on 18 profiles ofth eGuama-Imperatri z area).I n the Figs.3 2an d 33th e data on available P are compared with the depth inth e profile. It can be seen that only in the topsoil(0-1 0cm )th eamount s of available P are appreciable, though the variation is rather large.Belo wth etopsoi Lth eamount s ofavailabl e Pdecreas erapidly , especially when considering the Bray calues, to very low, and approximately constant, values, namely about 0.6 mgP 2Ospe r 100g o f soil(Truog )an d about 0.15m gP 205pe r 100 g of soil (Bray). No clear correlation emerges between the amounts of available P and the soil texture. For the eighteen profiles of the Guama-Imperatriz area, there are also data ontotal P. Contrary to the situation with the available P,th e values for total P arefairl y con-

237 Fig.32 Relacdo entre a quantiihule (ng/ 100g defosforo disponivel(me'todo Truog) 5-r forest profile clay --50% perfilde mala argita e a profundidade, em perfis desolos forest profile ''">>., >50% caoliniticos da Planicie dedrenagem *-•. perfil de mala livre savannah profile perjit decampo 3- r with piinlhile com "piinlhite* 2- o >: 1 . 9* o • Fig. 32 Relation between theamount 0 *eiff ° .450 of available phosphorus (Truog H r— method) and the depth, inprofiles of 40 80 120 160 200 240 280 cm freely draining kaolinitic Planicie —O Depth/Profundidade soils

Fig.33 Relafdo entre a quantidade fores! profile clay -^50% defosforo disponivel (metodo Bray) per/it de mata argila e a profundidade, em perfis de solos forest profile clay perjit de mala argila >50% caoliniticos da Planicie dedrenagem savannah profile livre perjit decampo . with piinlhile ' com 'piinlhile'

Fig. 33 Relationbetween theamount ,450 o°'*f 0\> a'ftga'f KthptgJrL' o j £ r T -i- -I 1 of availablephosphorus (Bray meth 40 80 120 160 200 240 2B0 320 cm od) and thedepth, in profiles of free •Dept h/ Profundidade ly draining kaolinitic Planicie soils stant throughout each individual profile, including the topsoil. It is only in the very heavytexture d profiles that adifferenc e may occur in total Pbetwee ntopsoi l and sub soil,an d thisi sno t moretha n 20m gP 205pe r 100g o f soil.Th etota l Pincrease swit h heavier texture of the soil. For instance, in the medium textured profiles, the amount is about 30 mg P2Os per 100g of soil, and in the very heavy textured profiles about 1 55m gP 205pe r 100g o fsoil. Acompariso n of the data for total P with those for available Pgive sa picture of the degreet owhic h the Ptha t actually occurs inth esoi l isfixed. Th e Brayvalue sar e used for the calculation of the ratio P2Os-total: P2Os-available,fo r eighteen of the profiles. The ratio generally increases with increasing depth in theprofile . In Fig.34 ,th e ratios of all horizons arecompare d withth econten t oforgani c matter, takingth etextur e of thehorizo n concerned intoaccount . Itca n besee ntha t theratio ,o r thedegre eo f fixa tion ofth esoi l phosphorus, increasessharpl y whenth epercentag e ofCarbo n drops to below approximately 0.5.Th efixation, a s determined in this way, isgreate r when the texture is heavier, but the percentage of Carbon the same. For the relatively light

') In eight forest profiles, for the most part of medium to rather heavy texture, the amount of P extracted with 25% HO was approximately 10 mg/100 gof soil(analysi s of Royal Tropical Institute, Amsterdam, Holland).

238 forcse profile clay Fig.34 Disponihilida- H 50° 9» T|g per/it de mala argila de do fosforo de solo forest profile flay emrelafdo com aper- 0 50% r perjil de mala argila cenlagemde Carbono eoo savannah profile e apercentagem de ar perjil decampo gila,em perfis de solos I 700 + • with plinlhite H ctm'plinlhite' caolinilicosda Plani- cie de drenagemlivre 600 -

500 ..<£ o o o o El

400

300 -f "eoD*0 o o" Fig. 34 Availability 200 - - •• OOO

PHYSICAL QUALITIES Seen in their entirety, the freely draining kaolinitic Planicie soils under forest cover havea 'good ' structure,whic happlie sals ot othos esection so fth eprofile s that arelo w in organic matter. This structure is usually weakly coherent porous massive, or subangular blockyt ogranular ,o fvaryin gstrength .Th egoo d structureo flatosoli csoil si n general is thought, by many, to be due to the nature of the clay fraction, on the one hand, and to a large activity of thesoi l fauna, in particular termites, on the other. The author believes that in the Planicie soils concerned, with their small amounts of ses quioxides,th e soil fauna certainly iso f great importance in bringing about, and main taining, the structure as it is under the primeval forest coverage. The activity of ter-

239 Foto29 Cigarras, As construfoesdas larvas deuma cigarra (Cicadidae/mJ asuperficie dosoloflores- tal. As crisdlidas no seu ultimoestado, escavam no solopocos de urn metro ou maisde pro/undidade, cobrindo-os commaterial desolo revestido de2 dm deallura m. oum. O capuz decerto modo constitui micromonolito viradopara cima do perfildo solo

Photo29 Cigarras. The constructions of the larvasof a cicadeat the surface of theforest soil. The larvadigs shafts of one metre depth andmore and closes thiswith ahood of coaledsoil material of about twodm height.The hood forms akind of upturnedmicro-monolith of thesoil profile mitesan d associated fungi, that of ants(notabl y theparaso l ants),o f larvaso f beetles and crickets, of forest crabs, and also that of rodents, isver y apparent everywhere in the forest soils(cf. Photos 29an d 30).Thei r effect on porosity and homogenisationo f the soil profile must be enormous. No quantitative data are available regarding this effect. It is, however, noted that in the relatively heavy textured profiles, termites (Cupins: Isoptera order) and larvas of a certain cicade (Cigarras: Cicadidae family) predominate, while in relatively light textured profiles the activity of parasol ants (Sauvas: Attaspp.) and rodents (e.g. the Tatu: Dasypoda sp.)i s most striking. Earth worms seem to becompletel y absent.

Permeability. The discussion is limited to the freely draining soils, thus those with good external drainage. There are differences in the internal drainage of these soils. The light and veryligh ttexture d ones(KLS )ma yhav ea somewhat excessive internal drainage,du et orapi dpercolatio n ofth erai nwater .Fo rsevera lo fth esoils ,namel yth e KYL,c , the RP-KYL, and the RP-KYL, CR, the subsoil (Bhorizons ) has apparently a somewhat slowpermeabilit y for water, resulting in only a moderately good internal

240 Folo30 Os montoesde terradeitado por umacoldnia de salivas(Att aspp.). Estas formigas manejam hortas defungosem liteira armazenada emcavidades no subsolo. Aalividade das saliva? e demaisfauna desolo, 4 urn fat or import ante na homogenizacao doperfildo solo sobfloresta mm

Photo 30 Theheaps of earth thrown out by a colonyof parasol ants {sauvas, Atta spp.). These ants maintainfungus gardens on litter which isstored in excavatedchambers in thesubsoil. Theactivity of the ants,together with thatof other soilfauna isan importantfactor in the homogenisation ofthe forest soil profile

drainage.Ther ear en omeasurement s availableo fth epermeabilit y ofth evariou shori zonso fth esoil sunde r discussion.

Moisture equivalents andmoisture tension curves. Of great importance isth e amount ofmoistur etha t canb estore d inth esoi lt ob euse db yth eplant sdurin gth edr yseason , sincethi sseaso n is fairly well defined in most parts ofth e Planicie(cf. 1.2). It israthe r common tofind, o n forested terrain, at the end of the dry season, seemingly very dry subsoils,eve nwhe nth eprofile s arerelativel yheav ytextured . Thecapacit y of soilst o retain moisture in available form depends upon the structure and porosity of the soil,th e percentage and type of clay-sized particles,an d upon the amount of organic matter. The moisture equivalent (M.E., given in gramso fmoistur e per 100g of soil), which gives an indication of the amount of moisture that can be stored in the freely draining soil, was determined for the disturbed samples of thirty- five forest profiles.I n Fig.3 5th e M.E.dat a of the B2horizon s are compared with the claycontent .Th e trend isver yclear .Th e increaseo f M.E.wit h the increase in percen tageo f clayi sno t directly proportional for the whole clay-range,bu t for generaldis -

241 g/IOOg Fig. 35 Relacdo entre 40 -r o equivalente de umidade e a percentagem de argilano horizonte x* *• B, de perfis de solos caotiniticosdaPlanicie 11 dedrenagem livre n 30-- ° i 25

20 -- J '

10 -• • forest profileIperfil de mala X savannah profile/pw/i/ de campd Fig. 35 Relation be~ /with plinthite/rom 'plinlhite' tween moistureequi valent and percentage of clay, in the B, horizonof profilesof -i r- H 20 40 60 60 100% freely draining kaoli- E»ClayMrgi/a nitic Planlcie soils

I Fig. 36 Relacdo entre g/100g o equivalente deumi- 40 \EH dade e a percentagem de argilano horizonte Ax de perfis de solos caotiniticosda Plant- ciede drenagem livre

•|30-- I!

20 --

{ x

• forest profileIperfit de mata . • * 10 -• X savannah profile!per/it decampo r with plinthite/rom 'plinlhile' Fig. 36 Relation be tween moistureequi valentand percentage of clay, in the A\ horizon of profilesof 20 40 60 60 100% freely draining kaoli- > Clay,'Argila nitic Planicie soils

242 cussion itca nb esai dtha t theincreas eamount s to3. 6g/10 0g pe r 10%clay .Th e M.E. is generally low in comparison with that of less weathered soils, as for instance the LowHumi cGle ysoi lo fth eAmazo nfloodplains. Th edat a arestil llo wwhe ncompare d with those of other latosolic soils rather than the kaolinitic ones of the Planicie. For latosolic-B horizons ofSa oPaul oState , it isreporte d that inth e medium rangeo f textures the M.E. value is approximately half of the clay percentage, or 5g/10 0 gin creasepe r 10%cla y(cf. LEMOS,BENNEMA , SANTOSel al., 1960 ,p . 71). The M.E. data for the topsoils of the thirty-five profiles are only slightly above the M.E.-% clay curve of the B2 horizons.I nFig .3 6th e situation for the Ax horizons is given. A calculation,fo r samples with approximately the same clay content, reveals that 1 %Carbo n accounts for about 1.7 g/100 g of the M.E.Th epar t ofth e M.E.du e to the presence of organic matter istherefor e small in comparison to the part due to theminera l material,i.e. thecla y fraction. M.E. data alone do not give much insight in the availability of soil moisture for the plants,especiall y sinceth e data are for disturbed samples.Th e availability is actually determined byth edifferenc e between theso-calle dfield capacit y (F.C.)an d thewiltin g point percentage (W.P.) respectively.Th eforme r isth e maximum amount of moisture that can bestore d inth e freely drainingsoil ,th e latter theamoun t of moisture held by the soil particles with a force larger than the maximally possible suction by the plant roots ('hygroscopic water').Thi s maximum suction is generally taken to be about 15 atmospheres. A method much in use for detailed study of soil moisture availability isth e determi nation of a so-called moisture tension or pF-curve of a natural soil sample (pF being the common logarithm of the soil moisture tension in cms water column). The Insti tute for Land and Water Management Research (I.C.W.) of Wageningen, Holland, kindly determined such curves on a dozen Amazon samples. The ones of subsoils (Bhorizons )o fth esoil sunde r discussion arereproduce d in Fig.37 .Th emoistur e freed betweenp F2. 0an d pF4.2 ,markin gth e F.C.an d theW.P .respectively ,i stake n asth e amount of available moisture (also called 'capillary water')1. This moisture may be taken to be stored in the smaller pores ('effective' pore diameter 20-0.2 micron). The lowerpar t ofth ecurve ,belo wp F2.0 ,concern sth ewate rcontaine d onlya t full satura tion of the soil ('gravitational water'), stored in the larger pores ('effective' pore dia meter > 20micron) . Only a few determinations are involved, and the samples not all remained quite na tural during transport. The curves of samples 231-3, 219-3 and 303-3 suggest never thelesstha tth eAmazo nkaoliniti cLatosol shav eth efollowin g moisture characteristics: 1. Aconsiderabl e part of thesoi l moisture isno t available for plant growth because it isto o strongly held byth e soil particles.Whe n thepercentag e ofcla y ishigh ,th e fixed moisture isu pt o3 5vol .% .

') Sometimes the M.E.,whic h isa t about pF 2.5, is taken as the lower limit, instead of the F.C. of pF 2.0. If this would be adopted for the Amazon soils, then the afore mentioned M.E. data should be converted, via apparent bulk density, to vol. %. It may be recalled that these M.E. data concern disturbed samples, not natural ones.

243 2. Therei sa hig hpercentag eo flarge rpore s(2 5vol . % ca.) which will befilled wit h airwhe nth esoi li si nfreel y drainingcondition , andtherefor ed ono tcontribut et oth e effective soilmoistur estorage .

Fig. 37 Curvas detensao deumidade deamostras de alguns solos caoliniticos da Planicie de drenagem livre e de alguns outros solos amazonicos. For obsiqu'to dolnstltuto I.C. W. de Wageningen, Holanda

231-3 279-3 219-3 303\ Pp 4.2-1- o V X +

3.4--

2.5-

2.0-- V \ x\

1.5-

0.4- 213-3 331-3

10 20 30 40 50 60 70 •WaterMgua, vol. %

Fig. 37 Moisture tension curvesfrom samples of a number offreely kaolinitic Planicie soils, and of some other Amazon soils. By courtesy of the Institute for Land and Water Management Research (I.C.W.) of Wageningen, Holland

No. Classification Location Depth Clay C Vol. Available sample (c/.Tab.9) org. weight1 moisture amostra classificacao localizacao profundi- argila agua dade disponivel (cm) (%<2(x) (%) (g/cm») (vol. %)

231-3 KYLm BR-14,km58 100 24 0.22 1.34 10.8 219-3 KYLt,» BR-14,k m32 4 80 80 ca. 0.45 ca. 1.09 9.3 303-3 KYL„A Curua-una, km 14 80 86 0.50ca. 1.24 10.5 213-3 RP-KYLr* BR-14,k m25 8 100 40ca. 0.30ca. 1.56 9.0 ca. 279-3 DL,s Araguaia(Ri oCorda ) 50 40 0.64 1.54 20.2

*) Volume weight = weight of 1cm *natura l sample in completely dry condition (approximately equal to apparent bulk density) peso de 1cm' de amostra natural em condifao totalmente seca (aproximadamente iguala massa especifica aparente)

Table 15/Tabela IS Description of the samples of Fig. 37/Descricdo dasamostras da Fig.37

244 3. The amount of soil moisture available for plant growth (the part between pF 2.0 and pF4.2 )i ssmall .I nth esample sunde rdiscussio n iti sonl yabou t 10vol . %.A fully rootedsubsoi lo f 1 metrethicknes swoul dhav ea neffectiv e soilmoistur ereserv ewhic hi s equivalent to 100m m rainfall. PEERLKAMPand BOEKEL(1960 ) giveavailabl e moisture data rangingfro m 13.5t o 33vol . %,fo r a number of Dutch soilso fvariabl e clay and humus content. Only for very sandy, humus-poor soils they give values that are the sameo rsmalle rtha n theAmazo nones . 4. Differences in texture have less effect on the amount of available moisture than might beexpecte d consideringth ever ydistinctl y higher M.E.value s at higher percen tageso fclay ,an d keepingi nmin d that alsoth erelativel yheav ytexture d soilshav eth e 'good' structure inherent to Latosols. The two very heavy textured subsoils (219-3, 303-3)even ,d o not have a greater amount of available moisture whatso-ever than the medium textured subsoil (213-3).Bot h former subsoils are actually relatively compact (they are both from central sections of planalto stretches, cf. IV.1.2) although this compactness is not so outstanding as to classify the soil as KYL, Compact phase.A really compact subsoil would have comparatively high volume weight and a steeper pF-curve, as borne out by the subsoil of the Red Yellow Podzolic soil, intergrade to KaoliniticYello wLatoso l (sample213-3) . Heavyan drathe rheav ytexture d subsoilstha tar edefinitel y non-compact arelikel y to have some more moisture available than non-compact medium textured subsoils asi s sample231-3 ,an d non-compact light or veryligh ttexture d subsoilslikewis eles savail able moisture. Preliminary field observations namely, suggest that for non-compact subsoils there isa n increase in porosity with increase in clay content. This will apply also toth e quantity of smaller pores,whic h largely determine the soilmoistur e availa bility. The field observations are being confirmed by the data on the B horizons of Table 16.Thoug h the porosity data of this table have limited value, because they do notconcer nnatura lsamples 1,th etren d ofincreasin gporosit ywit hincreas ei ncla ycon tent seems clear enough (the B2horizon s of the profiles 216an d 112/42A ma y still be compared with theA x horizon of profile 226,sinc ethe y have about the same percen tageo fCarbon) . The influence ofth eporosit y on the soilmoistur e availability isstil l illustrated byth e pF-curve of sample 279-3, which is of a Dark Red Latosol. This sample has nearly twiceth e amount of available moisture asthos e of the kaolinitic Latosols. According tofield observations ,thi s Dark Red Latosol profile isstrikingl y porous.2 Its pF-curve bears this out, and the difference in porosity isapparentl y largely because of a higher amount ofth esmalle rpores . Thenon-compac t subsoilso f the kaolinitic Latosols neversho wa shig ha porosit y as that of the discussed Dark Red Latosol. It seems therefore safe to assume that the available moisture ofth eforme r maximally variesbetwee n 5 vol. % (for the veryligh t

') Apparent bulk density andreal bul k density were determined on only a limited number of samples, to which the ones of Fig. 37 do not belong. *)Tha t the volume weight of the sample is nevertheless high, is probably due to its much higher content of iron oxides.

245 Table16 'Natural' porosity of freely drainingkaolinitic Planicie soils withnon-compact subsoil, all under primevalforest cover.'Natural' porosity is100 X (I - apparent bulk density/real bulk density). Determinationson disturbed samples

No. field Classifi Location Depth Org. Clay Bulk density 'Natural' descr. cation localizacao profun- Carb. argila massaespecifica porosity numero (c/Table9 ) didade apparent real porosidade descr. classificacao aparente real 'natural' decampo (cf.Tabela9) (cm) (%) (%<2|x) (g/cm') (g/cm8) (%)

A thorizo nfhorizonte A i

169 KLS Curua-una, km 3 0-20 0.73 5 1.55 2.59 40.2 226 KRLm BR-14,km429 0-20 0.69 7 1.49 2.63 43.3 206 KYU BR-14,km201 0-15 1.23 7 1.36 2.63 48.3 201 KYL-RPr&BR-14,kmll7 0- 5 1.20 14 1.30 2.57 49.4 221 KYU BR-14,km341 0- 5 1.34 19 1.30 2.61 51.2 216 KYUn* BR-14,km291 0- 2 3.60 70 1.08 2.49 56.6 112/42A KYUH* Curua-una, km6 (MO 3.16 81 0.96 2.45 60.8

B2 horizon/ horizonte B a 169 KLS Curui-una, km 3 300-400 0.08 15 1.46 2.65 44.9 226 KRLm BR-14,km429 70-120 0.20 15 1.61 2.66 39.1 206 KYU, BR-14,km201 130-210 0.19 20 1.49 2.66 44.0 201 KYL-RPr»BR-14,kmll7 65-220 0.14 32 1.43 2.66 46.2 221 KYU BR-14,km341 70-140 0.28 64 1.26 2.70 53.3 216 KYU** BR-14,km291 30-80 0.50 91 1.11 2.67 58.4 112/42A KYU** Curua-una, km6 100-140*''' 0.60 82 1.06 2.62 59.5

• edgeo f planalto/borda de planalto **B,horizjon/horizonteB, Tabela 16 Porosidade'natural' de solos caoliniticos da Plankie de drenagem livre, com subsolondo compacto,todos cobertos defloresta primitiva. Porosidade 'natural'4100 X(1 - massa especifica aparen te/massa especifica real). Determinaqao em amostrasdestorroadas textured kaolinitic subsoils) and 15vol . %(fo r the very heavy, if not the heavy and ratherheav ytexture dkaoliniti csubsoils) .I tshoul db enote dthat ,fo r theligh ttexture d ones,th efineness o fth esan dfractio n mayb eo fmor eimportanc efo r thesoi lmoistur e availability than the percentage of clay(cf. PEERLKAMPan d BOEKEL,1960) . The soil moistureavailabilit y situation for thetopsoil s may bedifferen t from thato f the subsoilsbecaus eo fth econcentratio no f theorgani cmatte r there.Th einfluenc e of the humus on the soil moisture availability is described by several authors. That at higherlevel so f humusther e isgenerall y more moistureavailabl ei sapparentl y not so much because of the presence of organic matter as such. BAVER(1956) ,fo r instance, showstha t thesimpl eaddin g oforgani cmatte r (peat)t o soilmaterial ,thoug h sharply raising the maximummoistur e holding capacity of the mixture, only little raises the for plants available moisture, since most of the water isdraine d belowth e pF theo f M.E. and also the W.P.percentag e somewhat increases.Th e influence of the organic matter should be largely indirectly, via its positive influence on the structure, i.e. the porosity. Forth eLatosols ,wit hthei r inherentgoo dstructure ,thi s influence cannot be

246 very large. From Table 16 animpressio n mayb egaine d on theexten t ofthi s influence for the Amazon kaolinitic Latosols,unde r their natural forest cover.Ther ei smostl y a fair difference in'natural 'porosit y of Axhorizon s and B2horizon s respectivelywit h about the samecla y content, but different percentages of Carbon. The topsoils under discussion therefore, are liable to have somewhat more moisture available than sub soils of comparable clay content. Secondly, for topsoils there is likely to be a more distinct increasei nth e amount of available moisture at an increasei nth e clay content than insubsoils ,owin gt o alarge r increaseo fth epercentag eo forgani cmatte rwit hin crease of the percentage of clay(se e before).Ther e were no pF curves determined on topsoils and therefore, exact data about their volume percentages of available moist ure arelacking .B ycombinin ghoweve rth erelevan t data onth epar t ofth e M.E.whic h isaccounte dfo r byth epresenc eo forgani cmatter ,o nmaximu mpercentag e of Carbon intopsoils ,an d onth eapparen t bulkdensities ,i ti sestimate dtha tth eadditio n ofavail ablemoistur ei ntopsoils ,i ncompariso nwit hsubsoils ,varie smaximall ybetwee n 1 vol. % (when very light textures are involved) and 5vol . %(whe n heavy and very heavy textures are involved). The range in available moisture of topsoils therefore may be maximallybetwee n6 an d 20vol . %. Bywa y of summary, it may be said that the kaolinitic freely draining Planicie soils under forest coverhave ,a sa whole,onl y smallamount s ofavailabl e moisture pervo lume unit. The advantage of the clayey aboveth e sandy soils with regard to moisture availability is fairly distinct only for topsoils. For subsoils the difference is smaller, and evennon-existen twhe nth eclaye yone sar ecompact .

The phreatic leveli sver y deep in practically all freely draining Plaru'cie soils.There fore thegroun d water doesnormall y not constitute asourc eo fmoistur e for the plants duringth edr y season.Onl y onth eyounges t Pleistoceneterrace s(3- 4m +)i t mayb eo f somevalu efo r thevegetativ ecover . Forth euppe rpar t ofth eGuama-Imperatri zarea , for instance, there are indications that the grosstimbe r volume ofth e primeval forest on such terraces is slightly higher than on adjoining, higher terraces with soils of the sametexture .

Penetration of roots. The bulk of the roots, and especially the rootlets, of the forest vegetation, is found very near to the soil surface. A number of the rootlets are even aboveth esoil ,particularl y inligh ttexture d soils,namel yi nth elitte rlayer .The ycollec t thenutrient sfro m thedecayin glitter ,an d alsorainwater . Butth edeepe r rootsar eals o very important, not only in providing ground support for the trees. The available moisture inth e soilca n only beutilise d toth eful l ifth eroo t systemi sdens e anddeep . With thelo wnatura l fertility ofth esoil sunde r discussion, adee prootin gsyste mi sal soo fimportanc efo rcollectin gwhateve rnutrient sar estil lpresen t inth edeepe rsubsoil . No quantitative data are available on the rooting systems of the components of the primevalforest , onth evariou s freely draining kaolinitic Planicie soils.Durin g the re connaissance soilsurvey sreporte d inthi sthesi si twas ,however ,notice dtha to nsevera l ofthes esoil sth epenetratio n ofroot si shampere dt oa degree .Thi sapplie sparticularl y

247 to the RP-KYL,th e RP-KYL, CR and the KYL,c . The mentioned soils have Bhori zonso f acomparativel y largecompactnes s and considerable resistance to penetration with a soil hammer. These Bhorizon s have afirm o r rather firm consistence and the colourtransitio n from theA horizo n isofte n clear. In profile pits and on recently cut road embankments, the hampering of the root development in these subsoil was easily noticed. Also, fallen trees provided for in dications; although the root systems of the various tree species can be very different, thelum po feart h thrownou t withth eroot srarel ycontaine d moretha n oneo rtw od m of such acontrastin g B horizon. Notably onth e RP-KYL, CR ofth e northern part of the Guama-Imperatriz area,o nth e RP-KYLrft around km27 0o fthi sare a and onth e KYL, C„A of the km 140-km 190 section the rooting is relatively shallow. For the KYL, cVh moreover, the frequency of fallen trees was observed to be comparatively large.Fo rth erootin gsituatio n inth ever yheav ytexture d soilsi ngenera l(Belterr acla y soils), reference should be made to the discussion on the cause of the cipoal and the 'cipoalic'fores t (IV.1.2) .

The soil KYL-RPr*o f the section around km 110o f the Guama-Imperatriz area, in contrast, is deeply rooted. This soil has also a clear colour change from the A to the B horizon, but thelatte r isfriabl e and hasver y littlecompactness . A deep rooting is a consistent feature of the medium and light textured soils, such asth e KYLm and the KLS.I n thelatte r itma ycompris e severalmetres .

V.3.1.2 TheSoils under Influence of Man THE SOILS UNDER THE SHIFTING CULTIVATION SYSTEM For Amazonia, no data are available ast o the amounts of nutrients accumulated in theprimeva lfores t cover.I ti spresume d thatthes eamount sar elarge rtha nth enutrien t storage in the forest soil itself (cf also NYE and GREENLAND, 1960),i ftimbe r volumes are not exceptionally low.I t iswel lknow n that immediately after felling and burning of the forest, the fertility of the soil increases considerably. This is, because a part of the nutrients accumulated inth e vegetation and litter are stored suddenly,vi ath e ash, inth euppe r soillaye ri na neasil yavailabl eform . Thetemporar y highnutrien t content ofth etopsoi lan d theincrease d rateo f mineralisation of thehumu scaus ea vigourou s growth of thefirst crop s (cf.Phot o 31).However , duringth e rainy seasonstha t follow after the burning, the rains, falling on the unprotected surface, cause surface wash ('fertility erosion') and leaching of the soluble salts,tha t werefree d from the ash and the mineralising humus, to deepersoi l horizons.Surfac e wash and leaching, the irre trievable loss of nutrients by crop removal, and the gradual decrease of soil organic matter content, are causes of a fairly rapid decline in fertility of the soil once it iscul tivated. Only the planting of fast-growing and deeply rooting perennials may keep such a decline in check. Normally however, cultivated plots are abandoned, also be cause of the proliferating of weeds,afte r two or three years,an d a newroca is started elsewhere. The abandoned plotgraduall y becomes fully covered with weeds and fast growing,sunlovin gsaplings .Thi si sfollowe d bybush ,an d ultimately secondary forest:

248 capoeira when young, capoeirao when old. This fallow, besides smothering of the weeds,i sthough t to beabl et o replenishth ebod yo f stablesoi l humus.I t alsoregain s gradually,b yit sdee prooting ,th enutrient sleache dt oth edeepe rsubsoi lan daccumu lates them again in thevegetativ e cover. Eventually,th e equilibrium of the primeval forest andfores t soilwil lb enearl yreache d(accordin gt o LAUDELOUT, 1962,th ereplen ishment of the soil organic matter isensure d withina fewyears ,whils t a full re-accu mulation of the nutrients in the vegetative cover requires much more time). After a varying number of years, the secondary forest is felled and burned anew, and ane w croppingcycl estarts . Notenoug hdat a isavailabl eo nfreel y drainingkaoliniti cPlanici esoil sunde r shifting cultivation to ascertain general trends in the changes in soil qualities that may take place under the successive cultivation cycles.A n establishment of such trends, and a comparison with the situation under primeval forest cover, may be possible onceth e data of the reconnaissance soilsurve yo f the Bragantina area (FILHOet al., 1963)ar e fully known. Reference may also be made to several publications on the effects of

Foto 31 O sistema de agricultura itinerante. Exemploda cullura doprimeiro ano numa rocada • •* 1^'^^J* dentro dafloresta primitiva. Entre ostroncos de drvores meio-queimados espalhados pela superficie do solo, plantam-se arroz e milho. No ^VS,:.', Unfa fundovise umafaixade mandioca. Perlo da casa sefaz umaqueimada

Am. • .. >^o^ l^JTCHrSM

Photo31 The shifting cultivation system. An example of the first year crops on a plot of agriculture {rocada)within theprimeval forest. Between thehalf-burned tree stems lying strewn on the soil surface,rice and maize isplanted. Inthe background a patch of cassava can be seen. Some burning is taking place nearthe dwelling

249 shifting cultivation elsewhere in the tropics, in particular those concerning West and Central Africa, which regions have, for a part, soils comparable with those of the Amazon Planicie (cf. II.2.3). In particular the previously mentioned, comprehensive studies of NYEan d GREENLAND(1960 )an d LAUDELOUT(1962? )contai n many data on the effects of cropping periods, and of fallow of secondary forest or grassland, on thequalitie sfreel y draining,usuall yhighl yweathere d soilso ftropica l Africa. If the fallow periods are allowed to last long enough, then the long-range decline in agricultural potential of the soils isapparentl y very limited. With relatively short fall ow periods however, definite degeneration of the land is unavoidable. Re-burning is done too soont osecur e full recoveryo f thelan d under the fallow, which now reaches onlyth ebus h stage.I never yne wcroppin g period therefore, loweryield sar e obtained. The latter isapparentl y the case on parts of the Bragantina area, and onth e uplands alongth elowe rTocantin s (Cameta),whic h arebot h regionswher eshiftin g cultivation has been practised for more than fourty years. BIARD and WAGENAAR (1960),fo r in stance, mention that in the Bragantina area much capoeira of six to ten years old is being used at present, although people are aware that much better results would be obtained iffellin g weredelaye dunti lth ecapoeir a istwent yyear sold .A plo to fa neigh t years old capoeira,i na n areawher eth e virgin forest wasfelle dabou tfourt y yearsago , wouldyiel donl yhal fth ecro pvalu eo fa virgi nfores t plot. Although in Amazonia, asa whole ,lan d isstil labundant , there isa n actual shortage of land around the established rural population centres, such as those of the Bragan tina area.Thi sresult si na loca lovercroppin g of thelan d underth eshiftin g cultivation systemi nit spresen t day form.

There seemst o be noclearl y defined preference for shifting cultivation on one of the varioussoil sdiscussed .Ver yligh t textured ones,however , are avoided, partly because theyeasil yerode . The majority of the present-day Amazon shifting cultivation is found on light and medium textured soils (KLS; KYLm). Concretionary soils (KYL, CR and RP-KYL, CR) are also used. The stoniness of the latter isapparentl y not a major disadvantage. That the relatively heavy textured soils are little used, islargel y because they are nor mally found farthest from the land and water transport routes. For the very heavy textured soils (KYLCA; KYL, cVh), a difficult supply of drinking water, and a slightly less easy tillage, may constitute additional factors accounting for their sparse occu pation.Whereve rsuc h Belterra clay soilsare ,however ,i n agricultural use,fo r instance south of Santarem,th eyield sar ecomparativel y very satisfactory.

THE SOILS UNDER ANTHROPOGENIC SAVANNAH To whichchange si n soilqualitie s a repeated burning ofth evegetatio n maylea d ulti mately, may be ascertained by studying the freely draining kaolinitic Planicie soils under savannah. Asha sbee ndiscusse d in IV.2.1,th e savannahs on such soilsar econ sidered to be anthropogenic.The y occur in Amapa Territory,i n parts of the Lower Amazon region,an d locallyo nth eeaster n part of Marajo island.

250 Analytical data are available on nine profiles of freely draining kaolinitic Planicie soils under savannah, largely from Amapa Territory1 (cf. Figs. 27t o36) . Thepercentage s ofCarbo n in thetopsoil sare , onth eaverage ,abou t0. 5 % Lowertha n those of forest topsoils (cf. Fig. 27).A t 100c m depth, and in the B2horizon s respec tively, the difference with the forest profiles seemst ob eless ,i f it exists at all (cf. Figs. 28 and 29).Th e ratios between T values and the percentages of Carbon, considering all horizons, are equal or only slightly below those of forest profiles (cf. Fig. 30).Th e above implies that, when compared with the ratios of forest topsoils and forest sub soils,th eratio so fT : %cla yar eprobabl ylowe ri nth etopsoil so f the savannah profiles, but the same,o r only slightly lower, inth e savannah subsoils (cf. Fig.31) . TheC/ N valueo f the savannahtopsoil s(A t horizons) is, onth eaverage , 13.0,bot h for the relatively light and for the relatively heavytexture d ones.A t 50 cmdept h theaver agevalu ei s 12.0an d at 100 cmdept h 10.0. TheC/ N valueso fth esavanna h profiles are therefore 2t o 3unit s higher than those of forest profiles2. The exchangeable metallic cations comprise a small part of the T value. The base saturation isapproximatel y the same astha t in forest profiles. The savannah profiles are usually slightly less acid than the forest profiles, possibly owingt o the repeated burning. The pH-H20 value in the topsoil (At horizons) of the nine profiles variesbetwee n 4.7 and 6.3.I n the subsurface soil (A3o r A2horizons ) the variation isbetwee n 4.8 and 5.9, and in the subsoil (Bhorizons ) it isbetwee n 5.2 and 6.2. The difference between pH-H20 and pH-KCl is slightly larger than in the forest profiles. Both inth e upper and the lower part of the savannah profiles it averages 1.0. Thevariatio n ishoweve rcomparativel y large. For six of the nine profiles, thecatio n exchange complex was analysed in detail (cf. Table 14).Th e relative proportions of the exchangeable metallic cations are approxi mately equal to those of the forest profiles, except for K+whic h isclearl y lower. The absoluteamount so f K+ rarelyexcee d0.0 6m.e./lO Og . The data for active acidity and pH-dependent acidity (Table 14)d o not suggest any consistent difference on the acidity situation with that of the forest profiles. The few data on available P (mainly Bray method), suggest that there is not such a great difference between theavailabilit y inth etopsoi l and that in the subsoil as in the forest profiles; the amount of available P iscomparativel y low also in the topsoil (cf. Fig. 33). Thisi sprobabl y related withth epreviousl y mentioned difference inth edistri bution of the organic matter. The values for total Par e comparable to those of forest profiles. The ratios P2Os-total: P2Os-Brayd o not exceed 150,eve n in the subsoil with itslo wleve lo forgani cmatte rconten t (cf. Fig.34) . Thefe wdat a suggesttha tth edegre e of fixation of the soil phosphorus may be somewhat less than in forest profiles (the non-plinthiticsavanna h samples,indicate d in Fig.34 ,hav e all lesstha n 50%clay ,th e twoplinthiti cone sbot h moretha n 50% clay).

•) The data of profiles AP, and AP„ described in the study of CARNEIRO (1955) on the Amapa savannahs, might also be used for comparison. *) KLINGE (1962) gives much higher values for C/N ratios of Braunlehm or Podzoliger Braunlehm under savannah, namely up to 90. See also the note at page 235.

251 The structure of the savannah profiles is, generally speaking, comparable to that of the forest profiles. However, the surface is sealed owing to rain splash. In the light textured profiles gully erosion may be severe, as can be observed at Santarem and Monte Alegre towns.Th ever y heavy textured ones,o n the other hand, may besub ject to considerable sheet erosion. With the exception of the very light textured ones, allsavanna h soilsunde r discussion havea comparativel y largecompactness ,especiall y in the topsoil. Pores are sparse and fine, and the consistency isofte n slightly firm, to firm. There is also a somewhat larger textural difference between topsoil and subsoil, when compared with forest profiles of the same over-all texture. The larger textural difference and thecompactnes s islikel yt o berelate d toth efac t that, possibly termites excepted, the activity of the soilfaun a isles stha n infores t profiles. No pF-curves for samples of savannah profiles were determined, so that no definite data can be givenfo r the amounts of available moisture. In viewo f the soil compact nessan d the lower levelo f the organic matter content inth e topsoil the amounts may beexpecte d to belowe rtha n infores t profiles ofcomparabl e texture.I n this respect it is, however, noteworthy that the data for moisture equivalent of savannah subsoil (B2horizons )ar ethesam eo rslightl yhigher,an d thoseo fsavanna htopsoil(A xhorizons ) the same or only slightly lower, than those of forest subsoil and forest topsoil respec tively(cf. theFigs .3 5an d36) . Bywa y of summary it can be said that the qualities of the freely draining kaolinitic Planiciesoil sunde rlong-lastin ganthropogeni c savannah arestil lfairl y wellcomparab let o those of forest soilso f the same character. The main differences are a somewhat lower organic matter content of the topsoil - and consequently a somewhat lower potential cation exchange capacity (T) and lower amounts of exchangeable cations and anions- , alarge r soilcompactness , and presumably a lower quantity of available moisture. Slight differences are found in the C/N ratios, the values for pH-H20, and the degrees of phosphorous fixation. But the picture changes, of course, much to the disadvantage ofth esavanna h soilsi fon etake sint oaccoun t thenutrient s accumulated inth erespectiv evegetativ ecovers .

THE SOILS INFLUENCED BY PRE-COLUMBIAN INDIAN OCCUPATION Throughout the Planicie of Amazonia patches of so-called Terra Preta soil can be found. InIII.3. 4i tha sbee ndescribe dtha tthes epatche sar ea kin d of'kitchen-midden', which have acquired their specific fertility from dung, household garbage, and the refuses of huntingan dfishing. Th eTerr a Preta (TP)soi l ofth ePlanici e iskaoliniti ci n character and usually freely draining. It gives therefore an excellent opportunity for study of the permanent changes in the qualities of freely draining kaolinitic Planicie soilswhic h maytak eplace ,unde rth eprevailin gwar m and wetclimate ,a sth eresul to f soilimprovemen t activities ofma nove ra lon gperio d oftime. 1 Although, understandably, there is a great variation in the chemical and physical

') Their study is also of importance because of an allegedly smaller susceptibility of rubber trees, planted on patches of TP, for the notorious fungus Dothidelta ulei.

252 qualities of the TP profiles, related with stronger or weaker and shorter or longer lasting Indian influence, somespecifi c trends can beclearl y discerned. Analytical data offive profile s of TP are available.Th e profiles are all from patches now cultivated but not settled on. Two of the profiles are of the common relatively light textured variant of the soil, having 12 to 20%'sil t + clay'; theothe rthre ear eo f the infrequently found very heavy textured variant, having more than 80% 'silt -f clay'. (The silt and clay fractions are taken together, because of the probability that, with the relatively high organic matter content, a good part of the determined 'silt' is actually non-dispersed clay). The blackish toplayer, containing pieces of ceramics, is normally less than 50 cm thick, sometimesa sthic k as 100cm .I n thever yheav ytexture d variant, the percentage of Carbon is 4 to 5% i n the upper 20c m of this blackish layer, and 1t o 2% i n the part below. In the relatively light textured variant, the percentage of Carbon is 1 to 2% in the upper 20c m of the blackish layer, and 0.5%i n the lower part. Thus,des pite the very humic appearance, the present organic matter content of the blackish layeri sonl ymoderatel yhigh .I ti sroughl ytw otime sth eaverag evalue sfo r non-enrich edfreel y draining kaolinitic Planicie soilso fcomparabl ecla yconten t(cf. Fig.27) .Th e blacknesso fth ecolou r isprobabl y due to acomple x formation of organicmatte r and Ca++,whic h may form a coating on the soil particles. In the yellowish subsoil of the TP, the percentage of Carbon isnearl y identical to that of the B2horizon s of non-en riched soils(cf. Fig .29) .I t isabou t 0.6%i nth ever ytexture d variant, and 0.3% inth e relativelyligh t textured variant. The C/N values are slightly higher than those of non-enriched profiles. This is true for the whole of each profile, but especially for the lower part of the blackish layer, where it is,o n the average, 18.I n the upper part of the blackish layerth e value is,o n theaverage , 14,an d inth eyellowis hsubsoi l 13. The potential cation exchangecapacit y (Tvalue )i srelativel y high,eve n whencomp ared with the higher organic matter content. Fig.3 8show stha t the points for the TP are all aboveth elin eo f the averagerati o T:%Carbo n of non-enriched forest profiles (cf. Fig.30) .I t isemphasize d that, when there isan y decrease of the Si02/Al203 ratio (Kivalue )wit h increasing depth at all, itcomprise s maximally 0.08 unit. Allhorizon s of allfive profile s have Ki data between 1.71 and 1.95, except for one subsoil sample which showed Ki = 2.2.I ncompariso n to this,value s slightly over2. 0les srar e inth e non-enriched kaolinitic Planicie soils. It can therefore be taken as certain that the higher T values of the TP are not due to a presence of small amounts of silicate clay minerals considerably more active than kaolinite. It might be concluded that theche mical activity of the organic matter of the TP is much higher than that of the non- enriched profiles. Thereare ,however , fairly great differences in Tvalue ,namel y upt o 10 m.e.,wher e percentages of Carbon and percentages ofcla yar eidentical . It isinter estingt onot etha tth eT value ssho wa fairl y goodcorrespondenc ewit hth eamount so f phosphate in the TP. The latter are strikingly high, though with a large variation. In figure 39,th e'surplu svalue 'o fT (i.e. theamoun t abovetha t ofth eT-%Carbo n lineo f non-enriched forest profiles) is compared with the 'surplus value' of P2Os-total (i.e.

253 the amount of P205-totalo f the TP, less the small amount of P20&-total which ispre sent in non-enriched profiles of comparable clay content; the latter is 0.02% for the relatively light textured variant, and 0.05% for the very heavy textured variant; cf. V.I.I). It can be seen that, below 0.30%, the 'surplus value' ofT increases about pro portionally withth e'surplu s value'o f PjOj-total.1 In thissection ,on epe rcen t 'surplus value' of P2Os-total corresponds with four m.e. 'surplus value' of T. Above 0.30% 'surplus value' of P20&-total, the increase of T is apparently more gradual. It maybe wondered how the high amounts of phosphates favourably influence the potential cation exchangecapacity . Forth etopsoils ,i tseem slikel ytha t thephosphoru s enlarges the activity ofth eorgani c matter, by havingentere d ina comple x formation with it. For the subsoils, however, which are poor in humus, such a complex forming alone seems no sufficient reason for their higher T values. Reference may be made to the publication of CATE(1960) . Hestudie d indetai l two freely draining kaolinitic Planicie soilsfro m the Curui-una centre,bein ga KLSan d a KYL^A respectively. Heobserve d that the potential cation exchange capacity increased considerably after equilibration of the soil samples concerned with 0.01N H3P04, for 25hours . Except for the topsoil of the KLS,th e increase was moretha n 100%,thi sals o beingtru e for thedeepe r sub soilwit h itslo worgani cmatte r content. Theexchangeabl e metalliccation scompris ea comparativel y largepar t ofth eT value . The base saturation Vi s between 30an d 85%.I n the blackish layer the average per centage is65% , and in the subsoil 50%.Th e relative proportions of the exchangeable metalliccation s arestrikingl y different from thoseo fnon-enriche d soils.Fo r 5 ofth e6 profiles (in one of the relatively light textured profiles no detailed analysis of the me tallic cations wasexecuted) ,th e Ca^+ varies between 20an d 77%o f the T value, and the average is 55%.I n the blackish layer, the percentage is often higher than in the subsoil; the absolute amounts of Ca++ in the former layer can be as high as 25 m.e./ 100g . For Nig"1-1-, thevariatio n isbetwee n 4an d 16%o f the Tvalue ,wit h an average of 8% . The percentage of Mg^ is normally less than one-fifth of the percentage of Ca++. The K+ and Na+ together always comprise lesstha n 2% of the T value.Thei r relative proportions and their absolute amounts are fully comparable with those in non-enriched soils.I t can beconclude d that theenrichmen t with metalliccation s con cerned very predominantly Ca++, to a small degree Mg4^, but no K+ and Na+. Understandably, the TPsoi l isles saci d than itsnon-enriche d relatives.Th e pH-H^O varies between 4.7 and 6.4. In the blackish layer the average value is 5.8 and in the subsoil 5.3.I n both the blackish layer and the subsoil thedifferenc e between pH-H20 and pH-KCL isabou t 0.9.Th e active acidity, or the (Al)+, issmall , when it exists at all. In the three profiles of thever y heavy textured variant,(Al) *comprise s less than 3% of the T value in the blackish layer and less than 15%i n the subsoil. In these three profiles thepH-dependen tacidity ,o rth eH v ,i sabou t 25% o f theT valu e in the blackish layer, and about 40% inth esubsoil .

') The determination of P,Ot-total for the two relatively light textured profiles is less reliable than the determination for the very heavy textured profiles. The latter were analysed afew year slater ,whe nth emetho d for determination of P,04-total was improved.

254 Fig. 38 RelacSo entre apercentagem deCarbono eo valor Tnosperfis deTerra Preta da Planicie

mtJIOOg 40-r

1 2 3

Fig.38 Relationbetween percentage of Carbonand value T, inTerra Preta profiles of the Planicie

Fig. 39 RelacSo entreo valor'excedeme' Te o valor'excedente' de P tO%-totalem perfis de Terra Preta da Planicie

m«/100g 20 -r Terra preta

10--

5-- 4 very heavy tenured irxtvra muito pnada

A reUlivcly lifhl miurcd trxlura ttiau*amtntt Jrnr j 025 OS 075 10 125 15% Surplus v»lue of „ „ . ,

Fig. 39 Relation between'surplus' valueof Tand 'surplus' value of PtOt-total, in TerraPreta profiles ofthe Planicie

255 Asalread y mentioned,th eamount so fphosphoru sar ecomparativel y veryhig hi nth e TP soil. Inth e very heavy textured variant, the amount of available P (Bray method) varies between 6.5 and 66.0 mg P2O5/100g o f soil (average 40)i n the blackish layer, and between 3.8an d 65.1mg/10 0g o f soil(averag e35 )i n thesubsoil . Inth e relatively light textured variant, the amount of available P (Truog method) varies between 3.2 and 98.8m gP 2O5/100g o fsoi l(averag e40 )i nth eblackis h layer, and between 6.7 and 31.2mg/10 0g o fsoi l(averag e20 )i nth esubsoil .

The amounts of total P are between 120an d 1.350 mg P2Os /100 go f soil (average 950)i n the blackish layer of the very heavy textured variant, and between 80an d 300 mgP 2O5/100g o f soil(averag e 150)i n its subsoil. In the relatively light textured vari 1 ant , the valuesfo r P2Os-totalar e between 60an d 370mg/10 0g of soil(averag e 150) in the blackish layer, and between 70 and 140 mg/100 g (average 100)i nth esubsoil . To assessth e degree to which the phosphorus isfixed i nth e TP soil,th e P2Os-total : P205-Bray ratios were calculated for the three profiles of the very heavy textured variant. The values are below 25i nal l horizons,eve n inth e subsoil where the percen tage ofCarbo n may belowe rtha n0. 5% .Henc e thevalue sfo r theT Psoi ld ono t agree with the curves for non-enriched kaolinitic soils {cf Fig. 34). It seems that the total amount of phosphorus of the TP soil isconsiderabl y greater than thefixing powe r of thehumus-poo rsoi lmaterial . The relatively high C/N values imply that the nitrogen may be in short supply. It is noted that KLINGE(1960) , whodetermine d Carbon and Nitrogen data for twoT Ppro files, givesvalue s for mineral N (in percentage of the total N) that are comparable to + those of non-enriched forest profiles.Th eratio sN-NH 4 :N-N0 3~ are howeverlower , especiallyi nth eT Ptopsoil ,du et o highervalue sfo r N-N03~. The moisture equivalent data for the three very heavytexture d profiles vary from 36 to 41g/10 0g an d from 34t o 37g/10 0g ,fo r topsoils and subsoils respectively. In the relatively light textured profiles, the range isfro m 8t o 12g/10 0 gan d from 7t o 9g / 100g respectively. These values are comparable to those of topsoils and subsoils of non-enriched forest profiles {cf the Figs.3 5 and 36).Th eamoun t ofavailabl e moisture inth eT Pi slikel yt ob eslightl y largertha n that ofnon-enriche dfores t soils.A sregard s other physical qualities,th eT P seemscomparabl e with itsnon-enriche d relatives.

THE SOILS UNDER MANURING AND FERTILIZING Chemical fertilizers and animal and green manures are applied locally to the freely draining kaolinitic Planicie soils, namely on the limited acreage with permanent agriculture. The latter involves the growing of vegetables and fruits on plots near the towns, of pepper mainly by Japanese small farmers, and of rubber on recently established smallholdings and afe w commercial estates(Pirelli ,Goodyear) .A system aticcollectio n ofdat aconcernin gth eresponse sto ,an dth eeconom yof ,th eapplication s offertilizer s andmanure si sbadl ylacking . Fundamental experiments on the effects of fertilizers or manures with annual or

*)Cf. note page 254.

256 perennial crops, on the various soils, and with different rotation systems, were com pletely lacking until afe w yearsago .However , after theestablishment , in 1939, ofth e Instituto Agronomieo do Norte(IAN )a t Belem,wit h several Experimental Substations throughout Amazonia, research inthi s respect has gradually been taken up.Th e data obtained so far are still piecemeal and little conclusive. Future fullyfledged researc h will certainly provide for more definite data on the best systems for fertilizing and manuring. The IANtrial so nfreel y drainingkaoliniti cPlanici esoil sconcer n mainlythos eo nth e terrafirme are a ofth eIA N terrains on the outskirts of Belem.Th e soilsconcerne d are relatively light textured (KYL„,) predominantly, while a small part is concretionary. Much of the area wascleare d of itsprimeva l forest coverman yyear sago .Th e follow ingdat a arepartl ysummarise d from expositionsa tth ePrimeira Reunido de Agronomia do Norte do Pais, Belem 1962 (the full descriptions of these trials are being published in the Boletins Tecnicos of the Institute).

Beans (Feijdo:Phaseolus vulgaris), were fertilized with powdered lime one month before planting.Statistic sdi dno t revealan ysignifican t differences inyield so fth econ trol plots and those with 2,4 , 6 or 8ton/h a of lime. The application of N, P and K (100 kg/ha of chili saltpetre, 500 kg/ha of super phosphate, 200 kg/ha of potassium chloride),i nvariou scombinations ,di d not showan ysignifican t increasei nyield .

Corn (Milho: Zea mais), planted on plotsjus t cleared from forest, wastreate d with 2 ton/ha of lime. The liming resulted in a highly significant increase in yield. On the other hand, the application of N, P and K. (300kg/h a of chili saltpetre, 400kg/h a of super phosphate, 100kg/h a of potassium chloride), invariou s combinations, did not result in any significantly higher yields. The same crop was planted on a plot that alreadyha d beencultivate d for threeyears ,no wbein gadde d 15to no fanima lmanure , 1to n oflime ,30 0k go fchil i saltpetre,40 0k go fsupe r phosphate and 100k go fpotas siumchlorid e per ha. The yieldcompare d veryfavourabl y with the lowyield so fcor n normally obtained onsuc hexhauste d plots,bu t thecos t ofth efertilizin g wasalmos ta s higha sth egros srevenue .

Oilpal m (Dende)i sno t indigenous, and not commercially cultivated asye t inAma zonia. At IAN however, various plots have been planted with the palm (Elaeis guineensis, Elaeismelanococca, and thevariou s varieties and hybrideso f these). From the manner of the growth of the plants it has been concluded by technicians of Uni lever Cie and of IRHO1, among others, that the palm adapts itself well to Amazon conditions. Dura x Dura crossings, for instance, have given fairly high yields, al though foliar analysis,a tth e RoyalTropica l Institutei nAmsterdam , Holland,showe d a deficiency of K+ and probably also of Ca++. Fertilizer trials on these plots are at present in execution.

') Institute de Recherchespour les Huiles et Ole'agineux, Ivory Coast.

257 Though the planting of rubber (Borracha: Hevea brasiliensis and crossings) in Ama zonia hasbee n greatly stimulated,ther e is no exact data on which to base fertilizing schemesfo r thisperennial .O nIA Nterrain sth e fertilizing takes placeaccordin gt o the formula N : P : K = 8 : 12 : 10. At Pirelli Estate near Belem, where the soils are largely concretionary (KYL, CR) the formula applied is 10 : 10 : 10; at Goodyear Estate,whic h islocate d 100k meas t of Beleman d has sandy soils(KYL OT; KLS),th e fertilizer application iso na 8 :1 6 :9 basis(HUFFNAGEL ,1964) .

Pepper (Pimentado reino: Pipernigrum) was introduced in Amazonia inth e thirties, by Japanese colonists. In recent years, this crop has rapidly expanded, notably at Tome-Agu,an d inth e surroundings of Belem.Fertilizin g isessentia l for this intensive crop; without it only 1 kgyiel d per plant may be obtained, whilewit h liberal applica tion offertilizer s 8k gca n bereached .A tTome-Ag u colony,wher eth e soilsar e rather heavy textured mainly (KYLrft), the avarage yield of full grown plants is 4 kgwhe n both inorganic fertilizers and manures are applied. Manuring inth ecolon y comprises the mulching with grass fodder (Capim-da-colonia: Panicum nimidium) as wella s the application ofrefuse s ofoi lrenderin g palm fruits. No usei smad eo flegumes ,becaus e ofassociate d dangerssuc ha scontaminatio n ofth esoi lwit h nematodes(b yIndigofera pascuorum) and stranglingo fth epeppe r plants(b yPueraria phaseoloides). DAY(1961 ) estimates yields of 2t o 3 kg per full grown pepper plant, on relatively light textured soils near Belem (KYLTO or KLS), with an application of 0.5 kg of mixed chemical fertilizer perplan t peryear .

For silviculture,th eapplicatio n offertilizer s isbein gtrie d too. PITT(1961 ) reports on the effect of phosphate addition on the artificial regeneration of several tree species. The application of 'half a dessert spoon' of rock phosphate or super phosphate did not showan yeffec t on plotsa tCurua-un a centre.Th eplot sha djus t beencleare d from their primeval forest cover,an d werelocate d either on theplanalt o with itsver yheav y textured soil (KYL„A), or on the lower terrains with their light textured soils (KLS). The application of the same amount of these phosphates was, however, essential for the growth oftre e seedlings on plots in the upland savannah area of Amapa territory, withheav ytexture d soil (KYLA).

Asregard spasturin g thefollowin g can besaid . Data ast o the mineral imbalancesi n Amazon cattle and the nutrient content of the soilso n which they graze, are given in the combined veterinary- pedological study of SUTMOLLER et al.(1964) . The study reveals that the soils ofth e uplandsavannah si nAmap a territory,i nth eLowe rAma zon region and on eastern Marajo island are deficient in several micro nutrients (Co balt, Manganese),a swel la si n phosphate and potassium. Thesoil sconcerne d mainly arefreel y draining kaolinitic Plaru'ciesoil s{cf. IV.2.1).A st o the improvement of pas tures on kaolinitic Plam'cie soils, trials may be mentioned on teso terrains1 at the

') The soils concerned are not quite freely draining (cf. DAY and SANTOS, 1958).

258 artificial insemination station of Sao Salvador, near Soure (NIEUWENHUIS, 1960). IntroducedPangola grass (Digitaria decubens) ,Elephant grass (Pennisetum purpureum) , Mucund (Stizolobiumdeeringianum) and tropical Kudzu(Pueraria phaseohides) wer e growingwel l on these terrains, and responded favourably to the addition of chemical fertilizers. Aninterestin gexampl eo fmixe d farming onfreel y draining kaolinitic Planicie soilsi s practised locally in the Bragantina area.Tobacc o isplante d inthi s area on plotsuse d previously as temporary pens for cattle which graze on nearby lowlands (cf. SOARES, 1956). Reference may be made to the favourable results of additions of chemical fertilizers on the growth of sugarcane and other crops on the taboleiro uplands near Campos,i n Rio de Janeiro State. These uplands have soils that are much comparable to the Amazon soils under discussion.

V.3.1.3 TheSoils with Horizonsof Fossil Plinthite The freely draining kaolinitic Planicie soils with horizons of fossil plinthite, hard or soft, are KYL, CR,an d RP-KYL, CR. Seen as a whole, their chemical and physical qualities arecomparabl e with these ofnon-plinthiti cfreel y draining kaoliniticsoils . It isespeciall y noteworthy that plinthitic soilhorizon sd onot hav ea higherdegre eo f fixation ofth e soilphosphoru s present(cf. Fig.33) .Analyse d plinthitic topsoils under forest cover have a base saturation which varies from 34 to 20%, and is therefore slightly above the average value for non-plinthitic topsoils. For plinthitic subsurface and subsoil horizons, no consistent difference in this respect appears. Plinthitic topsoils seem to have a comparatively high percentage of organic matter (cf. Fig. 27), plinthitic subsoils however have a relatively low percentage (cf. the Figs. 28 and 29). The potential cation exchange capacity (T value) of these subsoils is nevertheless not consistently different from what may be expected at a given percentage of clay (cf. Fig. 31). In other words, plinthitic horizons with low percentages of organic matter have T values which are slightly higher than the average for non-plinthitic horizons with similar percentages of organic matter. This seems to indicate that, although Ki data isapproximatel y thesam ea stha t of theothe r Planicie soils, the clay fraction of the plinthitic horizons is slightly less chemically inactive. In this respect it may be mentioned that on alarg erubbe r estatenea r Belem,th estrippin go fth e non-plinthitic surface layer(whic hcontain s mosto fth e organic matter!)fro m the plinthitic soils has beenfoun d to beadvantageou sfo r thegrowt h ofrubbe r saplings. Owingt othei r compactness,horizon so ffossi l soft plinthite,suc ha sth e B horizon of the RP-KYL, CR, are not favourable for rooting. Horizons of fossil hard plinthite however, such asth e Ahorizo n of the RP-KYL, CR and the Aan d Bhorizon s of the KYL, CR, normally are excellent rooting media. Such horizons consist very predomi nantly of concretions in friable earth (cf.Phot o28) ; exceptionally the segregates of sesquioxides form slag-like,impenetrabl emasses . Reference may bemad e to theremark s in ChapterI Vo n thenatura l vegetativecove r of the freely draining plinthitic Planicie soils.Th etimbe r wealth iscertainl y not lower

259 than onnon-plinthiti cfreel y drainingkaoliniti c Planiciesoils ,an d at leaston evaluabl e treespecie s predominates on one ofth e plinthitic soils.

V.3.2 Agricultural LandCapabilit y Evaluationfo r the Freely Draining Kao linitic Planicie Soils,an d their Adequate Management

V.3.2.1 AgriculturalLand Capability Evaluation The mainconclusio n from theforegoin g discussion istha t the natural fertility islo w in all freely draining kaolinitic Planicie soils(excep t the Terra Preta). In such a situa tion,th edifference s whichoccu r insoi lqualitie ssuc ha sorgani cmatte rstorage ,availa blemoistur e reserve,an d root penetrability, becomever y important. Agricultural land capability evaluation inth e Planicie will havet o be based largely upon the differences inth elatte rqualities .Suc hexterna lfactor s asth etopograph y and drainage pattern of the land, its accessability, and the degree of geographical coherence of the individual soils,pla ya part .Apar tfro m thishowever ,an dth esoi lpreference s ofindividua lcrops , it may be stated that the most favourable conditions for crop production are to be found onthos erelativel y heavytexture d forest soilstha t havenon-compac t and friable subsoils. The KaoliniticYello w Latosol (,Ortho),an d Kaolinitic YellowLatosol ,intergrad e to Red Yellow Podzolic soil which are of rather heavy, or heavy texture (KYLr» and KYL»; KYL-RPrA and KYL-RP*) meet these requirements best. On thesesoil s the highest gross timber volumes of the primeval forest are found (cf. IV.1.2). Where the above soils are used for shifting cultivation they often have a smaller po tential because of the tendency of the small farmers to gradually reduce the length of thefallo w period. The soils under savannah coverage have definitely less agricultural potential than those under primeval forest, because of the near absence of a nutrient reserve in the vegetation, less organic matter in the topsoil, and some degradation in the soil struc ture which affects unfavourably the available soil moisture and the root penetration. The soils with horizons of fossil plinthite have about the same potential, when hand culture methodsar eapplied ,a s non-plinthitic soilso f thesam edegre e ofanthropi c in fluence.Onl y when machinery is to be used do they obviously become much less suitable,becaus eo fthei rstoniness .

V.3.2.2 So/7Management The preceding examination of thechemica l and physical qualities of the freely drain ing kaolinitic Planicie soils, under forest or under human influence, enable an assess mentt o bemad ea st oth ebes tway sfo r management ofthes esoil safte r clearingo fth e primeval forest cover, aswel l asfo r their recuperation after adverse human influence.

MAINTENANCE OF SOILORGANI C MATTER It is evident that the organic matter in the soils is of paramount importance. It is

260 practically the sole bearer of the available plant nutrients, determines the nutrient storagecapacit y( Tvalue )t olarg eextent ,i sth emai nsourc eo fnitrogen ,an d makesth e phosphorus of the soil easily available. The soil organic matter also has a favourable effect onth esoi l moisture availability. In forest soils,th e soilorgani c matter isconcentrate d ina thin superficial layer.The refore, at the commencement of agricultural activitieso n these soils,grea tcar e has to betake n to leave intact, as much as possible,th e original superficial layer. Any effort to improve the fertility of the forest soils, and to recuperate the soilsunde r savannah and those under long-lasting shifting cultivationwit hshor t fallow periods,wil lhav e to taketh emaintenance ,an d if possibleth eenlargement , ofth esoi lorgani cmatte r hasa basis. Theoretically, it would be possible to attain gradually a level of soil organic matter which is comparable to that of the Terra Preta soil. It has been shown before that already underth enatura lcove rth epercentag eo forgani c matter ishigher ,bot h intop - soil and in subsoil, with increasing percentage of clay.I n the relatively light textured forest soils,th e layer in which the organic matter isconcentrate d isofte n up to 40c m thick. A relatively large mass of soil is involved in storing this organic matter. In the heavytexture d soilshowever ,a ver ythi n superficial layer,ofte n only2 t o 5cm , suffices tostor ea goo d parto fth enaturall y present organicmatter .I nth elatte r soilsi twil lno t be very difficult to maintain, under permanent cultivation, the natural organic matter level.Thi swil lb epossibl e byarrangin gtha t theorgani c matter becomesabou t equally distributed in the total tilth layer,throug h increasing the sub-surface section with the sameamoun t oforgani cmatte rwhic hi slos t inth ethi n superficial layer.Eve na nabso lute enlargement of the organic matter level in such soils seems feasible, namely by securing an organic matter level inth etota ltilt h layer which isa shig h asi n the super ficial section of forest profiles. In fact, some of the heavy and very textured profiles under forest, notably those with very high timber volumes, have remarkably thick At horizons.A nexampl eo f thisi sshow nb yth e profile at the planalto edgeo f Curua- una centre(11 2o f Fig. 11).Thi s profile, having 80%clay ,contain s about 3% Carbo n downt o4 0c mdepth , resultingi na T valu eo f 15 m.e.pe r 100g o fsoi lthroughou t this layer! Whether an increase in the organic matter level in the relatively heavy textured soils can be realised under non-intensive agriculture, and whether it will be economically justifiable, is questionable. It is, however, evident that the heavier the texture of the soil,th e better itsabilit y to maintain the level of organic matter inth e newlyoccup ied forest soil and to restore it after degeneration of the soil under adverse human influence. There aresevera l methodso fensurin gmaintenance ,an d possibly increaseo fth eleve l of soil organic matter, such as the planting of cover crops and the application of ani mal and green manures and compost. Application of some specific green manures seemst o beth e most effective and economical.Ther e are however hardly any compa rativetrial savailabl e toconfir m thisbelief .

261 EXPLOITING OFSOI L MOISTURE RESERVE It hasbee nmentione d that a large part ofAmazoni a hasa climatewit h adr y season, in which the rainfall is too small and too irregular to provide for unhibited plant growth. The soil moisture has to beexploite d in such periods.Th e total effective soil moisture reserve depends upon the amount of available moisture per unit of soil, as wella s upon the intensity and depth of rooting. It seems hardly feasible to obtain an enlargement ofeffectiv e soil moisture reserve by mechanical improvement of porosity and soil penetrability. In topsoils these qualities are normally quite favourable, and breaking of compact subsoil layers will not be economically justified for most crops. Becauseo f the influence of thesoi lorgani c matter on soil moisture availability and on root penetrability however, via itsinfluenc e on the activity ofth e soil fauna, the main tenance or increaseo fth e organic matter level is an important management measure also for theeffectiv e soilmoistur ereserve . Strange as it might seem for a region with a 'wet' climate, some irrigation, during a shortperio do fth eyear ,ma ywel lb ea neconomi cpropositio n fora numbe ro fdrought - sensitive crops. In this case, owing to the supplementary character, irrigation with flexible equipment sucha ssprinklers ,seem sadvisable .

TILLAGE AND EROSION CONTROL Normally, the topsoils are friable and easily penetrable, and the dangers of large- scale erosion are very small. Exceptions to this are the very heavy textured and the very light textured soils. The former may be subject to surface sealing and sheet ero sion, and on the latter gully erosion can be severe. On these soils,tillag e and erosion controlmeasure srequir especia l attention.

FERTILIZING Immediately after felling and burning of the forest cover, an addition of chemical fertilizers seems superfluous, owing to the temporary enrichment of the soil by ash from thevegetation ,an d therapi d rateo f mineralisation ofth esoi lhumus . Afterwards however, the removal of nutrients from the soil system by crop production or cattle rearingha st o berecompense d byne wadditions ,i fth e nutrient statuso f thefores t soil ist ob efull y maintained. Suchadditions ,togethe r with measurestake n to maintain or increase the organic matter content, will also be required for recuperation of ex hausted soilsan d when general soil improvement isenvisaged . With the foregoing soil data, and the few data of fertilizing trials, it is possible to assess tentatively to what degree applications of chemical feitilizers are necessary and what their effects will be:

Theapplicatio n oflarg eamount sof limei sprobabl y notadvantageous ,a t least noti n so far as it is intended to lessen the acidity. It istru e that the soils are generally very strongly to extremely acid, but the active acidities, or the exchangeable aluminum percentages, are nevertheless only moderately high.A sligh t increaseo f the pH (about one unit, to approximately pH-H20 = 5.5) isprobabl y sufficient to eliminate the un-

262 favourable effect ofthi saluminu m on theavailabilit y of other nutrients. Owingt o the fact that the potential cation exchange capacities are generally low, such an increase mayb eeffecte d withmoderatel y smalldose so flim e(a sa roug hestimate , 1.5t o2 tons / ha for a tilth layer of 20cm , not counting possible leaching). Afurthe r raising of the pH is likely to again decrease the availability of several nutrients. This may apply particularly to micronutrients ,o fwhic hth etota l amounts seemt o bealread y natural ly low both in forest and savannah soils. The often very gradual effect of liming on crop production has to be kept in mind, when thefirst year' s results of the fertilizing areevaluated .

The available amount of nitrogen inth e soils isunlikel y to bedeficien t for sometim e after the clearing of the forest, because of the relatively high rate of mineralisation of theorgani c matter.O n longcultivate d plotsan d on savannah areas however, an addi tiono f N-fertilizers mayhav econsiderabl eeffect . Itwil lb enecessar yt oappl y nitrogen whengree n manuring isexecute d withsoi lcover sothe r than N-fixating legumes.

It has been shown before that theconten t of totalphosphorus is low.Th efixation o f thiselemen t issomewha t larger inclaye y soilstha n insand y soils,bu t itis ,a sa whole , probably lessstron g than in Latosols other than the main Amazon ones.A n enlarge ment, if necessary, of the content of available phosphorus in the soil, can therefore probably be obtained with only moderate amounts of phosphates, particularly in the sandy soils.Thi s isals o true, because the exchangeable aluminum isonl y moderately highan d easilyeliminate dwit hsom eliming .Becaus eo fth e relation betweenth eavail ability of phosphorus and the percentage of organic matter, an enlargement of the apparentlyto olo wamoun to favailabl e phosphorus inth esavanna hsoil simplie stha ta n enlargement of theorgani c matter content must beeffecte d at the sametime . Without a considerable quantity of organic matter, it will bever y difficult to obtain a satisfac tory level of available phosphorus (1 mg P2Os-Truog per 100g of soil/7). It is most probably uneconomical to apply such liberal doses of phosphates that the amount of total phosphorus inth esoi lbecome sgreate r than thefixing powe r of the humus-poor soil material, as is apparently the case in the Terra Preta subsoil (more than 100m g P2Os-total per 100g o f soil).Suc h largeadditions ,however , may havea n enlargement ofth epotentia l cation exchangecapacit ya sa n interestingsid e effect.

Fertilizing withpotassium seems to be generally useful, at least when perennials are involved. SUTMOLLER et al.(1964 ) found that potassium deficiences in Amazon cattle may occur if the amount of exchangeable potassium in the soils concerned is below 0.20 m.e. per 100 g of soil. Reference is made to the evidence from foliar analysis executed for Amazon oil palms. Experiments with the crop on comparable soils in West Africa have shown that for this perennial the level of exchangeable potassium should be at least 0.15 m.e. per 100g o f soil, in the upper 30t o 40c m of the profile (IHRO-OLLAGNIER, personal communication). Another minimum value estimation is 1.5 to2 %o fth eT value .I fthi si sgenerall y true forcondition s inth eAmazon ,the n not

263 onlyar eth eexhauste d savannahsoil sdeficien t inpotassiu m forth egrowt h ofoi l palm, butals oth efores t soils(th estorag e inth e vegetation not taken intoaccount) ;eve n the Terra Preta soili sat ,o rbelow ,th eminimu m necessarylevel .

The nutritional imbalances in cattle, grazing on the savannah soils, suggest that fertilizing with several mesoan d micronutrients wil l be effective. On non-exhausted soilsalso ,whe nsupportin g perennials or sustained cultivation ofannuals , deficiencies of mesoan d micronutrient sar ea possibilit y ifn ocompensatio n ismad ewit hchemica l fertilizers or compost.

It is stressed that any fertilizing recommendation, made wit h the aid of the present- day knowledge of soil- plant relationships in Amazonia, is at best a rough estimate. Therear eals oth edifferen t requirementso f individualcrops .Th e method ofsoi ltillag e and the system of fertilizer application can have a tremendous influence upon the effect, ifany ,o f added materials.Also ,a shortag e ofon e element,eve nwhe n a micro nutrient, may causea n apparent lack of response to a fertilizer application which is totally misleading. With the generally weak buffering of the soils, especially of the sandy ones,i t iseasy ,o n theothe r hand, to unwittingly add an excesso f one nutrient, thereby diminishing theavailabilit yo fothe rnutrients .

V.3.2.3 Systemsof Crop Production At present, the crop production on the freely draining kaolinitic Planicie soils is carried out very predominantly by native small farmers, with hand culture methods, and theapplicatio n ofth eshiftin g cultivation system.I tha sbee nconclude d inV.3.1.2 , that this system, as it ispractise d nowadays in Amazonia, reduces gradually the agri cultural potential of the land. In viewo f the low standard of livingo f the rural popu lace,an d theenvisage d immigration from otherpart s of Brazil,th eexpandin g of crop production on the Planicie soils is a necessity. This must be done both by improve ment of the crop production in the present-day agricultural areas such asth e Bragan- tina area,an d bycreatin go fagricultura l settlementsi nprimeva l forest areas.Stimulat ing of the establishment of commercial agricultural estates, belonging to specialised companies,migh t bea n effective wayt o achieve higher production and they mayb eo educativevalu efo r therura l population on their peripheries.Governmen t agricultura policieshoweve rwil lai m mainly at improvingan d enlargingth e meanso fexistenc e o the nativesmal l farmers. For this,th e shifting cultivation system can not bediscarde d as'obsolete' .I ngenera lth esyste msecure sa living ,thoug h ona lo wlevel ,fro m tropical soils with a low natural fertility, in areas where land is superfluous. The choice and succession ofcrop safte r clearingo fth efallo w isofte n the ingeneous result of century- long experience by the native population in tropical areas where animal manures, chemical fertilizers and insecticides were not available in the past, nor are they today an economic proposition. Proposals to introduce 'modern' agricultural practices for peasant farming havet o take full account, not only of theeconom y of the agricultural products themselves, but also of the socio-economic level of the rural populace.

264 Under the prevailingconditions ,th e improvement andexpandin go f peasant farming on the freely draining kaolinitic Planicie soils of Amazonia has to take the shifting cultivation as a basis.Ther e seem to besevera l ways inwhic hth e system, as practised nowadays, may beimproved , and gradually transformed int o asyste m of settled agri culture. Oneo f thefirst an d easiest measureswoul d bet o allot more landt o original settlers, to strictly control the area of forest felled each year, and to allow only the felling of sufficiently old (15 to 20 years) fallow. Secondly, a stimulating of the growth of the fallow may be feasible, as has been tried for instance at Yangambi station in Congo, withth ecorridor system . A main gap in Amazon agriculture is believed to be the practical absence of peren nialssuc h as rubber, cocoa and oil palm.Suc hcrop sar ea firm ste pt o settled agricul ture.The y will provide for considerable and stablecas h without requiring much tech nical skill, form a good protection against degeneration of the land, and are likely to react the most favourably to applications of manures and chemical fertilizers. To givea n encouragement toth e planting of such perennials, on apercentag e ofth e plots allotted for ordinary shifting cultivation, seemst o bea n inherent part ofa sound agri cultural policy. In fact, this has been propagated in recent years, for the agricultural colonies established by the lnstituto Nacional de Imigracao e Colonizacao (INIC) in general and for rubber in particular, via thejoin t project ETA '54 of U.S. A.I.D., SPVEA and other institutions. However, the strict precautions that have to be taken against the leaf blight Dothidclla uleiar e one ofth e hindrances to expansion of rubber planting. Both technically and economically it seemsquit e possiblet o promote thees tablishment ofsmal lgrove so foi lpalm salon gsimila rline sa si sbein gdon efo r rubber. Smallholding of this perennial seems promising even in those parts of the Planicie where climatic conditions may not be optimal for the crop. Apart from this, one or more of the many indigenous oil rendering palm species may prove to be a valuable perennialcrop . Alsoth e prospects for increased growing ,b ysmal lfarmers , of specialssuc h as pepper look favourable. Such new crops constitute, in fact, a more intensive useo f the land. The use of fertilizers, green manures and cover crops will bejustifie d by higher and morestabl e profits, and also annualcrop s maygraduall y begrow n continuously. It ispossibl e that an adapted system of mixed farming wil l prove to be a stable base for rural settlement. Besides the profits of livestock husbandry itself, the grass or legume ley may improve soil conditions and the manure of the penned cattle may be fully usedfo r theenlargemen t ofcro pproductio n onth esam efarm . Inthi scase ,muc h attention has to be given to the quality of the grass ley.Th e application of adequate chemical fertilizers, and the introducing of new and highly productive grasses and palatable legumes,o n artificial pastures,see mt ob epromising . Another possibility for improvement ofth estandar d of livingo fth erura l population 'n Amazonia isth ecombinatio n ofcro p production with a modern exploitation of the forests. Several forest production reserves on Planicie areas are proposed or already established (cf. Fig. 22), in accordance with recommendations made after the FAO-

265 SPVEA forest inventories. The creation of agricultural settlements on the peripheries of such reserves would be a safeguard against uncontrolled cutting and felling, and secure a labour force for the exploitation of the reserve itself; at the same time, it would provideextr acas hfo r theagricultura l community. It is without doubt that theagricultura l research,a s started by the Institute) Agrono- micodo Norte, must be enlarged and deepened. It should comprise studies on crop rotation and intercropping schemes,gras sleys ,type so fgree nmanure san dcove rcrops ; trialso nth eapplicatio n ofchemica l fertilizers, animal manure,compos t andth ewast e from industrial production; trials on cultivation practices such as planting time and spacing; the introduction and testing of newcrops ,an d the selecting of the best varie ties of local crops; the assessment of the absolute production capacity of the various soilsan d theeconom y ofsoi lmanagemen t measures.Th eexperience s inthes e respects in other tropical areas with similar physical conditions, of which particularly West Africa and Congo(Yangambi )mus t bementioned , mayb eutilize d toa larg eextent . Rural extension work,t o a greater degree than beingdon e byth e Fomento Agricola, will have to teach the new methods which have proved valuable, and to produce and distribute seeds and other planting material. This may prove to be one of the bottle necks for the expansion of the Amazon agriculture, because of the preference of the present-day rural population for extractive activities on the natural vegetative cover, and the aversion oftrainee sfro m spending a considerable time in the interior. In this respect, the establishment of subsidised pioneer and demonstration farms should be highly beneficial. In fact, most measures in the technical agricultural field for improving Amazonian peasant farming arewel lunderstoo d byAmazo n agronomists.Othe rservice showever , areequall yindispensable .Thes ear eth esufficien t and regular releaseo fworkin g funds for agricultural research and rural extension work; cheap agricultural credit; promo tion of processing industries on the spot; price policies for cheap provision of fertiliz ers, insecticides and implements; easing of import and export regulations and theres triction of federal, state and municipal taxeso n theagricultura l products;th e promo tion of agricultural cooperatives, of transportation, education and other measures in thesocia field.l I ti softe n ratherth elac ko fthes eservices ,o rthei r unbalanced or short- term implementation, which is delaying the emergence of a prosperous agricultural community inAmazonia .

266 References

A.A.F. 1942 U.S.Ai rForc epreliminar yBas e Mapso fth eAmazo n Region (1:500.000). AB'SABER, A. N. 1959 Conhecimentossdbr ea sflutuac/>es climatica sd oquaternari o no Brasil. NoticiaGeomorfologica 7(1). ARENS, P. L. 1963 Reconnaissance soil surveyo f the Dapartments of Punoan d Madred e Dios, Peru. FAO report (inpress) . AUBERT,G . 1958 Classification des solstropicaux .Bull. An.ORSTOM, 4e trim. 1958. Paris. AUBERT, G. and 1956 Projet de classification des sols. Trans.Int .Congr .Soi l Sci., PH. DUCHAUFOUR Congr. VI, Paris 1956,Vol . E,Comm . V: 597-604. AUBREVILLE,A . 1958 Les forets du Bresil, etude phytogeographique et forestiere. Boiset Forets 59: 3-18; 60: 3-17. BAKKER,J . P. 1951 Bodem en bodemprofielen van Suriname, in het bijzonder van de noordclijke savannestrook. Landbouwk. Tydschr. 63: 379-391. BAKKER,J .P .an d 1949 Indrukken van de natuurwetenschappelijke expeditie naar J.LANJOU W Suriname 1948-'49. Tijdschr. Kon. Ned. Aardr.Genootsch. 66:538-557 . BALDWIN, M..C.F. KELLOGG 1938 Soil classification. In:Soil s and Men:979-1001 . U.S. Dept. and J. THORP Agr. Yearbook of Agriculture 1938. BARBO SA , O. 1959 Geomorfologia do Territ6rio do Rio Branco. Noticia Geomorfologica I. BARBOSA, O. et al. 1962 Relat6rio preliminar sobre o 'Projeto Araguaia*. SPVEA, Bclem-Brasil. BARROS,H.C.,J.L,DRUMOND,I958 Levantamento de reconhecimento dos solos do Estado do M. N. CAMARGOel al. Rio de Janeiroe Distrit o Federal. Commissao de Solos do C.N.E.P.A., Serv. Nac. Pesc. Agron Bol.II. Rio deJaneiro , pp 350. BAUER.M . 1898 Beitrage zurGeologi c derSeychellen , insbesonderezu r kcnntnisde s Laterits. Seuet Jb. Min.usw. 2: 163-219. BAVER,L ,D . 1956 Soil Physics, third edition. John Wiley andSon s Inc., New York-London. BEARD,J .S . 1955 The classification of tropical American vegetation types. Ecology36 (I): 89-100. BENNEMA,J . 1963 The Red and Yellow soils ofth e tropical and subtropical uplands. SoilSci. 95 (4): 250-257. BENNEMA,J. , M.N . CAMARGO 1962 Regionalcontras t inSout h American soil formation, inrela and A. C.S . WRIGHT tion to soil classification and soil fertility. Trans. Intern. Soil Conf. New Zealand 1962, Comm. IV and V:493-506 . BENNEMA,J. , R. C. LEMOS 1959 Latosolsi n Brasil. Proc. lllrd Inter-african soils conference, and l_ VETTORI Dalaba 1959.CCT A publ.50(1): 273-281. BENNIMA,J . and L. VETTORI I960 The influence of thecarbon'cla y and silicascsquioxide s ratioso n thep H of Latosols. Trans. Intern. Congr. Soil Sci. Congr. VII. Madison USA 1960. Vol. IV:244-250 . BIARD.J. and I960 Crop production inselecte d areas of the Amazon valley. G. A. W. WAGENAAR FAO, EPTA report 1254. Rome.

267 BONNET, J. A. 1950 Latosols of Puerto Rico. Trans. Intern. Congr. Soil Sci. Congr. IV, Amsterdam 1950. Vol. I: 281-285. BOTELHOD A COSTA, J. V., 1958 Carta geral dos solos de Angola; I - Distrito da Huila. A. L. AZEVEDO, E. P. C. Junto de Inv.d o Ultramar. Lisboa. FRANCO and R. P. RICARDO BOUILLENE, R. 1926 Savanes equatoriales en Amerique du Sud. Bull. Soc. Bot. Belg. 58:217-223 . BRAMAO, D. L. and 1958 Climate, vegetation and rational land utilization in the R. DUDAL humid tropics. Proc. IXth Pacific Sci. Congr., Bangkok 1958.

BRAMXO, D. L. and 1960 Soil map of South America. Trans. Intern. Congr. Soil Sci. P. LEMOS Congr. VII, Madison USA 1960.Vol .1:1-10 . BRAUN, E. H. G. and 1959 Estudo Agrogeologico dos campos Pucuari-Humait4. J. R. DEANDRAD E SPVEA, Commissao de Planejamento; IV. Serie: Recursos RAMOS Naturais. Belem. Also in: Rev. Bras, de Geografia 21 (4): 443-499. CAMARGO, F. C. DE 1949 Reclamation of the Amazonian floodlands near Belem. Proc. U.N. Scient. Conf. on the Conserv. a. Utiliz. of Re sources.Lak eSuccess ,Ne w York 1949.17:598-602 . CAMARGO, M. N. and 1962 Someconsideration s on themajo r soils of the humid tropics J. BENNEMA of Brasil. First Soil Correlation Seminar South Centr. Asia; World Soil Resources Rep. FAO/UNESCO: 87-91. CAMPBELL, D. F. et al. 1949 Relat6rioprelimina r sobrea geologi a dabaci a de Maranhao. Conselho Nacional de Petr61eo. Bol. 1. Rio de Janeiro. CARNEIRO, L. R. DA SILVA 1955 Os solos do Territ6rio federal do Amapa. SPVEA, Setor de Coordinacao e Divulgacao, Belem (Brasil). CATE Jr, R. B. 1960 Studies on gibbsite accumulations in soils. North Carolina State College. Department of Soils.Thesis . CLINE, M.G.et al. 1955 Soil survey of the Territory of Hawaii. Series 1939, 25. U.S.Dept . Agr., Washington. COHEN, A. and 1953 Klassificatie en ontstaan van savannen inSuriname . Ceologie J. J. VAN DERErn e en Mijnbouw. Nw.Serie 15:202-214. COLEMAN, N. T., S. B. WEED 1959 Cation exchange capacity and exchangeable cations in and R. J. MCCRACKEN Piedmont soilso fNort h Carolina.SoilSc. Soc.Am. Proc. 23: 146-149. DAY, TH. H. 1959 Report for the reconnaissance soil survey of the Caete- Maracassume Area. Stenciled report FAO/SPVEA Mission, Belem (Brasil). DAY, TH. H. 1961 Soil investigations conducted in the Lower Amazon valley. FAO, EPTA report 1395. Rome. DAY, TH. H. and 1958 Report on an excursion to the Rio Gurupi. Typewritten; J. BENNEMA FAO files Rome, pp 16. DAY, TH. H. and 1958 Levantamento detalhado dos solos da Estacao Experimental W. H. SANTOS de Sao Salvador, Maraj6. Stenciled report Inst. Agron. do Norte and FAO/SPVEA Mission, Belem(Brasil) . DAY, TH. H. et al. 1964 British Guiana soil survey. Report UNSF/FAO, Rome (in press). DERBY, O. A. 1877 Contribuicaopar a a geologia da regiao do baixo Amazonas. BoletimGeogrdfico, Ano VII, 80:830-849 . DOST, H. 1962 Verslagva nee nbodemverkennin g opd e Sipaliwine-savanne. *>* Dienst Bodemkartering, Paramaribo (Surinam), pp 32. DREWES, W. U. 1961 Explorations and land evaluation surveys for colonisation purposes in the tropical selva of southern Peru. Peruvian Times 21: 1096.

268 DUCKE, A and G. A. BLACK 1954 Notas sobre a fitogeografia daAmazoni a Brasileira. Inst. Agron. do Norte. Bol.Teen. 29,Bele m (Brasil). Also in english (1953): Phytogeographical notes on the Brazilian Amazon. Anais Acad. Bras, deCiencias 25 (1) : 1-46. EDELMAN, C. H. 1950 Soils of the Netherlands. North Holland Publishing Co., Amsterdam, pp 177. FILHO, J. P. S. O. el at. 1963 Levantamento de reconhecimento dos solos da Zona Bra- gantina. Inst. Agron. do Norte, Belem (Brasil).(i npress) . FRANKART, R., A. HERBILLON196 2 Carte pedologique et rapport preliminaire de Bugisira- and H. VERHOEVEN Mayaga-Busoni. Intern. Training Centre for Aerial Survey (I.T.C.), Consulting Department. Delft - Holland. GLERUM, B.B . 1960 Forest inventory in the Amazon valley- V(Regio n between Rio Caete andRi oMaracassum6) . FAO, EPTA report 1250.Rome . GLERUM, B.B . 1962 Forest inventory in the Amazon valley- VII (Survey in the Ucuuba-bearing region of the Tocantins river). FAO, EPTA report 1492. Rome. GLERUM, B.B .an d G. SMIT 1960 Forest inventory in the Amazon valley- VI (100%surve y in the Curua-una region). FAO, EPTA report 1271. Rome. GLERUM, B. B.an d G. SMIT 1962a Combined forestry/soil survey along road BR-14, from Sao Miguel to Imperatriz. FAO, EPTA report 1483. Rome. GLERUM, B.B .an d G. SMIT 1962b Survey of the Mahogany region. FAO, EPTA report (in press). GOUROU, P. 1949 Observagoesgeografica s naAmazonia . Rev. Bras, deGeogra- fia 11(3) : 355-408; 12(2); 171-250. GUERRA, A. T. 1952 Formacao de lateritas sob a floresta equatorial Amazonica (Territ6rio Federal do Guapor6). Rev. Bras, de Geografia 74:407^*26. GUERRA, A.T . 1954 Estudo geogrifico do Territorio doAmapa . Col. Bibl. Geogr. Brasileira 10:p p 366. GUERRA, A.T . 1959 Geografia do Brasil - Grande Regiao Norte. IBGE-CNG, Biblioteca Geografica Brasileira, Rio deJaneiro . Vol. 1, Ser. A. Publ. 15.pp421 . GUPPY, N. 1958 Wai-wai.Throug h theforest s north of theAmazon . London. HAMMEN, TH. VAN DER 1957 Climatic periodicity and evolution of South-American Maestrichtian and Tertiary floras. Boletin Geologico 5(2) : 49-91. HAMMEN, TH. VAN DER 1960 Upper Pleistocene and Holocene climate and vegetation of and E. GONZALEZ the 'Sabana de Bogota' (Colombia, South America). Leidse Geologische Mededelingen 25: 261-315. HARRASSOWITZ, H. 1926 Laterit. Fortschr. Geol.Paleontol. 4 (14): 253-566. HARRASSOWITZ, H. 1930 Boden der tropischen Regionen. (y) Laterit und allitischer (lateritischer) Rotlehm. In: Blanck, E.: Handb. Bodenlehre Vol III. Berlin 1930.387-436 . HARRISON, J.B . 1934 The katamorphism ofigneou s rocks under humid tropical conditions. Imp. Bur. Soil Sci.(Harpenden) . pp 79. HEINSDUK, D. 1957 Forest inventory inth e Amazon valley- 1 (Region between RioTapajo san d Rio Xingu).FAO ,EPT A report601. Rome. HEINSDIJK, D. 1958a Forest inventory in the Amazon valley- II (Region between Rio Xingu and Rio Tocantins). FAO, EPTA report 949. Rome. HEINSDUK, D. 1958b Forest inventory in the Amazon valley- I V(Regio n between Rio Tocantins and Rios Guama and Capim). FAO, EPTA report 992. Rome.

269 HEINSDIJK, D. 1958c Forest inventory in theAmazo n valley- III(Regio n between Rio Tapaj6s and Rio Madeira). FAO, EPTA report 969. Rome. HEINSDIJK, D. 1960 Dry land forest onth eTertiar y and Quaternary south of the Amazon river (Interim report). FAO, EPTA report 1284. Rome. HEINSDIJK, D. and A. DE 1963 Inventariosflorestais n a Amazonia. M.A. Servico Florestal, MIRANDA BASTOS Setor de Inventarios Florestais, Bol.6. Rio de Janeiro. HEYLIGERS, P. C. 1963 Vegetation and soil of a white-sand savanna in Surinam. Verh. Kon.Ned. Akad.Wetensch. ,afd . Natuurkunde.Twee - de reeks, deel LIV, 3. HlLBERT, P. P. 1955 A ceramica arqueol6gica da regiao de Oriximina. Publ. 9 Inst. Antropol. Etnol. Para, Museo Goeldi. Beldm(Brasil) . pp76. D'HOORE, J. 1949 Bull.Agr. Congo Beige40: 66. D'HOORE, J. 1954 L'accumulation de sesquioxides libres dans les sols tropi- caux. 1NEAC Ser. Sc.62. D'HOORE, J. 1959 Pedological comparisons between tropical South America and tropical Africa. Afr. Soils4 (3):4-20 . D'HOORE, J. 1959 Report - Head I, 1 and 2. Proc. Hlrd Interafrican soils conference, Dalaba 1959,CCT A publ. 50(1) : 59-68. D'HOORE, J. 1960 The soils map of Africa south of the Sahara (1:5.000.000). Trans.Int .Congr .Soi l Sci.,Cong r VII, Madison USA,1960 , Vol.4: 11-19. HUFFNAGEL, H. PH. 1964 Rubber production in the Amazon region. FAO, EPTA report 1808.Rome . I.Q.A. 1949 Metodos de analises de solo. Instituto de Quimica Agricola. Bol. 11. Rio deJaneiro . JACKS, G. V., R. TAVERNIER 1960 Multilingual vocabulary of Soil Science, second edition. and D. H. BOALCH FAO. Rome. JONGEN, P. and 1959 Les nappes de recouvrement de la Cuvette Centrale Congo- M. JAMAGNE laise.Proc .Hlr d Interafrican SoilsConference , Dalaba 1959. CCTA publ. 50 (1):413-420 . KATZER, F. 1903 Grundzuge der Geologie des unteren Amazonasgebietes. Leipzig. KELLOGG, C. E. 1949 Preliminary suggestions for the classification and nomen clature of Great Soil Groups in tropical and equatorial regions.Comm . Bur.Soi l Sci.Techn . Commun.46: 76-85 . KELLOGG, C. E. and 1949 An exploratory study of soil groups in the Belgian Congo. F. D. DAVO L 1NEAC Ser. Sc. 46.p p73 . KING, L. C. 1957 A geomorfologia do Brasil oriental. Rev Bras, de Ceografia 18(2) :147-263 . KLINGE, H. 1962 Beitrage zur Kenntnis tropischer Boden, V. Uber Gesamt- kohlenstoff und Stickstoff in Boden des Brasilianischen Amazonasgebietes. Z. Pflanzenerndhr. Dung.Bodenk. 97(2): 106-118. LAMEGO, A. R. 1960 Mapa geoldgico do Brasil, 1:5.000.000. Min.Agr. , DNPM, Divisao de Geologia e Mineralogia, Rio de Janeiro. LANJOUW, J. 1936 Studies of the vegetation of the Suriname savannahs and swamps. Ned. Kruidk. Arch. 46: 823-851. LAUDELOUT, H. 1962(?) Dynamics of tropical soils in relation to their fallowing techniques. FAO report 11266/E. Rome. LECOINTE, P. 1903 Le Bas Amazone. Ann. de Geogr. 12. LECOINTE, P. 1907 Carte ducour sd el'Amazon edepui sl'Ocea njusqu' a Manaus et del a Guyane bresilienne(1:2.000.000) .Ann. deGeogr. 16, PI. IV.

270 LECOINTE, P. 1922 L'Amazonie bresilienne, Le pays, ses habitants,se s resour ces; notes et statistiques jusqu'en 1920. Two volumes. Augustin Callamel (Editeur). Paris. LECOINTE, P. 1945 O estado do Para, a terra, a agua e o ar. Biblioteca Pedago- gicaBrasileira 5 (Sr5a): 15-16. LEENHEER, L. DE, 1952 Cartographie et caractensation pedologique de la catena de J. D'HOORE and K. SYS Yangambi. 1NEAC Ser. Sc. 55. LEMOS, R. C. DE, I960 Levantamento de reconhecimento dos solos do Estado de J. BENNEMA, R. D. DOS Sao Paulo. Commissao de Solos do C.N.E.P.A., Serv. Nac. SANTOS el al. Pesq.Agron . Bol. 12.,Ri o deJaneiro ,p p 634. LIMA, R. R. 1956 A agricultura nas varzeas do estuario do Amazonas. Inst. Agron. do Norte, Bol.Teen. 33.Bele m(Brasil) . LINDEMAN, J. C. and 1959 Preliminary survey of the vegetation types of northern S. P. MOOLENAAR Suriname. The vegetation of Suriname, Vol. I, part 2. Van Eedenfonds, Amsterdam. MAIGNIEN, R. 1958 Le cuirassement des sols en Afrique tropical de l'ouest. Ann. de la Carte d'Alsace-Lorraine; Thesis Straszbourg. MARBUT, C. F. 1932 Morphology of laterites. Trans. Intern. Congr. Soil Sci., Congr. II, USSR 1930.Vol . V: 72-80. MARBUT, C. F. and 1925 The topography of the Amazon Valley. Ceogr.Rev. 15: 617— C. B. MANIFOLD 642. MARBUT, C. F. and 1926 The soils of the Amazon basin in relation to their agricul C. B. MANIFOLD tural possibilities. Geogr. Rev. 16:414-442 . MILLER, G. 1954 Orellana. W. Heinemann Ltd, London. MIRANDA BASTOS, A. DE 1958 Floresta na regiao Rio Amapari - Rio Matapi - Rio Cupixi, SPVEA, Belem (Brasil). MOHR, E. C. J. and 1954 Tropical soils. Netherlands Royal Tropical Institute, Am F. A. VAN BAREN sterdam, pp498 . NlEUWENHUlS, W. H. 1960 Development ofgrazin gan d fodder resources in the Amazon valley. FAO, EPTA report 1238. Rome. NYE, P. H. and 1960 The soil under shifting cultivation. Comm. Bur. Soil Sc. D. J. GREENLAND Techn.Commun. 51. OLIVEIRA, A. I., el al. 1956 Brasil. In: Handbook of South American Geology, an ex planation of the geologic map of South America (Jenks, W. F.). Geol. Soc.Am . Mem. 65. PEERLKAMP, P. K. and 1960 Moisture retention by soils. Versl. Meded. Comm. Hydrol. P. BOEKEL Onderz.TNO 5: 122-138. PETRI, S. 1954 Foraminiferos fosseis da bacia do Maraj6. Bol. Fac. Fil. Cien. Let., Un. S.P., Geologia //. pp 170. PETROBRAS 1960 Preliminary map of the Amazon geologic basins. Petr61eo Brasileiro Ltda, Belem(Brasil) . PITT, J. 1961 Application of silvicultural methods to some of the forests of the Amazon. FAO, EPTA report 1337. Rome PLAISANCE, G. and I960 Contribution a 1'etude des limons des plateaux de la foret H. W. VAN DER MAREL de Chaux (Jura), Analises physiques, physico-chimiques et chimiques, deuzieme partie. Ann. Agron. 11: 661-711. PRESCOTT, J. A. and 1952 Laterite and lateritic soils. Comm. Bur. Soil Sc. Techn. R. L. PENDLETON Commun. 47.p p51 . ROBERTS, R. C. el al. 1942 Soil survey of Puerto Rico. U.S. Dept. Agr.Bur .PI . Ind.,i n coop, with Univ. of Puerto Rico Agric. Expt. Sta. Series 1936, 8. RUELLAN, F. 1945 Evolucao geomorfol6gica da baia de Guanabara e das regioesvizinhas . Rev.Bras, de Geografia 6(4) :445-508 . RUHE, R. V. and J. G. CADY 1954 Latosolic soils of central African interior high plateaus. Trans. Intern. Congr. Soil Sci., Congr. V. Leopoldville 1954, Vol. 5:401 -401.

271 SAKAMOTO, T. 1960 Rock weathering on 'terras firmes' and deposition on 'varzeas' inth eAmazon . J. Fac.Sc. Univ. Tokyo. Sec.II. 12(2) : 155-216. SCHULZ, J. P. 1960 Ecological studieso nrai n forest innorther n Suriname.Nort h Holland Publishing Co.Amsterdam , pp 267. SCHUYLENBORGH, J. VAN, 1951 Over de invloed van magnesium op de structuur van sedi- and J. S. VEENENBOS menten.Landbouwk. Tijdschr. 63(11) . SIOLI, H. 1951 Alguns resultados e problemas da limnologia Amazonica. Inst. Agron. do Norte, Bol. Teen.24. SIOLI, H. 1956 Uber Natur und Mensch im Brasilianischen Amazonas- gebiet. Erdkunde10 (2): 89-109. SIOLI, H. and H. KLINGE 1961 Uber Gewasser und Boden des Brasilianischen Amazonas- gebietes. DieErde 92 (3) : 205-219. SlVARAJASINGHAM, S., 1962 Laterite. Adv.in Agron. 14:1-60 . L. T. ALEXANDER, J. G. CADY and M. G. CLINE SOARES, L. DE CASTRO 1953 Limites meridionals e orientals da area de ocorrencia da floresta Amazonica em territ6rio Brasileiro. Rev. Bras, de Geografia 15(I) : 2-122. SOARES, L. DE CASTRO 1956 Excursion GuideBoo k no8 .XVIIIt h Intern. Geogr. Congr., Rio deJaneir o1956 . SOIL SURVEY MANUAL 1951 U.S. Dept. Agr., Handbook no 18.U.S .Govnt . printing office, Washington D.C. SOIL SURVEY STAFF 1960 Soil classification, a comprehensive system. Seventh Appro ximation. U.S. Dept. Agr.Soi l Conservation Service. SOMBROEK, W. G. 1962a Reconnaissance soils survey of theGuama-Imperatri z Area (area along the upper part of the Belem-Brasilia highway). Stenciled report FAO/SPVEA Mission, Belem (Brasil). SOMBROEK, W. G. 1962b Soils of Amazon areas with natural pastures. Stenciled report FAO/SPVEA Mission, Belem (Brasil). (also at: Netherlands Royal Tropical Institute, Amsterdam). SOMBROEK, W. G. and 1962 Reconnaissance soilsurve yo fth eAraguai a Mahogany Area. J. B. SAMPAIO Stenciled report FAO/SPVEA Mission, Belem(Brasil) . STERNBERG, H. O'REILLY 1950 Vales tectonicos na Amazonia?. Rev. Bras, de Geografia 7 (4): 511-534. SUTMOLLER, P., 1963 Mineral imbalances in cattle in the Amazon valley. Nether A. V. DE ABREU, lands Royal Tropical Institute, Amsterdam (inpress) . J. VANDE RGRIF T and W. G. SOMBROEK SYS, C. 1960 La carte des sols du Congo-Beige au 1:5.000.000. Serie INEAC, Carte dessol s et de la vegetation du Congo-Beige et du Ruanda-Urundi, (A) Sols. Brussels. SYS, C. et at 1961 La cartographie des sols au Congo, ses principes et ses methodes. INEAC, Serie Techn.66. SYS, C. 1962 A scheme for soil classification for the soils of Congo. Mimiographed report, Symposium on soil classification, Ghent,jun e1962 . THORP, J. and 1940 Laterite in relation to soils in the tropics. Annals of the M. BALDWIN Assoc. Amer. Geogr. 30(3) : 163-183. THORP, J. and G. D. SMITH 1949 Higher catagories of soil classification: Order, Suborder, and Great Groups. So/7Sc. 67(2) : 117-126. VARGAS, L. F. C. 1958 Contribution to the study and correlation of Tertiary and Pleistocene terracesan dplanatio n surfaces inth eLowe ran d Middle Amazon,th e Estuarine Deltaan d theestuarie so fth e tributaries of the Guajara bay,i n theAmazo n Region.Mi miographed report, FAO/SPVEA Mission, Belem (Brasil).

272 VETTORI, L. 1959 As relagoes Ki and Kr na fracao argila e na terra fina. Anais VII Congr. Soc.Bras. Ciencia do Solo, Piracicaba (Sao Paulo) 1959(i n press). VIEIRA, L.S .an d 1961 As caatingas do Rio Negro. Inst. Agron. do Norte. Bol. J. P.S .O . FILHO Teen 42. Belem (Brasil). WAMBEKE, A.VA N 1962 Criteria for classifying tropical soils by age. J. Soil Sc. 13 (1): 124-132. WILHELMY, H. 1952 DieeiszeitlicheundnacheiszeitlicheVerschiebungderKlima- und Vegetationszonen in Siidamerika. Tag. Ber. u. wiss. Abh. d.Dt . Geogr. Tage, Frankfurt/M. 1951: 121-128. WOLDSTEDT, P. 1958 Das Eiszeitalter; Grundlinien einer Geologie desQuartar s II. Ferdinand Enke Verlag, Stuttgart. ZEUNER, F.E . 1959 ThePleistocen e period. Itsclimate , chronology and fauna! successions. Hutchinson Scientific andTechnical , London.

273 Summary

The present study actually deals with the soils of the Brazilian part of the Amazon region. The boundaries accepted for this region are the transitions from the belt of equatorial forest to the savannahs of North-Eastern and Central Brazil and those of Rio Branco-British Guiana.Th eeaster n halfo fth eregio n isdiscusse d in particular.

PEDOGENETIC FACTORS Climate.Onl y a small, north western section of Brazilian Amazonia has a climate without anydr yseaso n(typ eAf inKoppen' s classification; Fig.2) .Th egreate r parto f theregio nha sa fe w months peryea rwit hlittl eo r no rainfall (type Am).

Geology. Brazilian Amazonia consistso fa low ,sedimentar y area,th eAmazo n valley proper, and parts of the crystalline shields of Central Brazil and the Guianas (Fig.5) . The crystalline shields are of Pre-Cambrian agean d mainlyconsist so f granites,gneis sesan d mica schists.Th e sedimentary part consists ofsevera lbasins ,namel y the basin of Acre,o f the Amazon proper,o f Maraj6,an d that of Maranhao (Fig.5 an d Appen dix 8). The sedimentary part of Amazonia has at its surface only narrow bands of Paleozoic-Mesozoic deposits, of very varying character; the greater part consists of Tertiary unconsolidated sediments which are kaolinitic clays and quartz sands. The Pleistocene sediments, which are similar in character to the Tertiary ones, are thin, and their extent has been reported to be limited. Holocene deposits comprise a small area, much lesstha n wasestimate d by early explorers of Amazonia.

Geomorphology. Inth ewatershe d regionst oth enort h and south ofth eAmazo n river systemth e following geomorphological unitsca n bedistinguishe d (Figs.8 ,9 an d 17): 1. Undulating terrains with outcropping crystalline basement. 2. Undulating terrains with outcropping Paleozoic, Mesozoic or Early Tertiary de posits. 3. Twopeneplanatio n surfaces insideth eare a designated ascrystallin eo nth egeologi cal maps.Thes esurface s occur at 400-600an d 250-400m altitud e and are assumed to be of Cretaceous and Early Tertiary are respectively. Little can be said with certainty about these surfaces. Thebroa d axialpar to fAmazoni a consists of: 4. Flat plateau land, known asAmazo n planalto,an d of Plio-Pleistoceneage . 5. Upland terrains at a lower level, fashioned to terraces at various levels, and of Pleistocene age.(Th e units(4 )an d (5)ar e designated together as Planfcie). 6. Lowlands,o f Holoceneage .

274 The height of the Amazon planalto is usually 150-200 m in the eastern part of the valley and somewhat lessi nth ewester n part. Acharacteristi c of this planalto istha t it hasa tit ssurfac e ate nt o twenty metresthic klaye ro funiform , veryheav y sedimentary clay. It issuppose d that thisso-calle d Belterra clay had itsorigi n inth eerosio n ofup lifted kaolinitic deposits ofth e Andes area. The clay must have settled very gradually and evenly on theflat botto m of a huge shallow inland sea.A larg e part of Amazonia was covered by this sea during an age of the Latest Tertiary or Earliest Pleistocene whenth e general sea levelwa shig h (Calabrian;cf. Tabl e 1). The upland terrains in the axial part of Amazonia that were fashioned during the Pleistoceneperio d havea grea texpanse .A 50 0k mlon gsectio no fth euppe r parto fth e recently constructed Belem-Brasilia highway, allowed for a detailed study of theloca l geomorphologic constitution (Appendices 4 and 5). Various terraces occur in this stretch. Someo f them are coastal-marine, the othersfluviatile. Al lterrac e levelsmus t havebee n formed during timeswhe n the existing sea levelwa shighe r than at present and therefore they areo f interglacial age.Fo r part ofth e stretch,a compariso n ofth e terrace levels with world wide interglacial sea levels enables the terraces to be tenta tivelydate d indetai l(Table s 1 and2) . It islikel ytha t in other parts ofth e Amazonvalle y proper theterrac e levels are also ofinterglacia lage . The very locally occurring, so-called massape terrains, and parts of eastern Marajo island arepresumabl y ofEarl y Holoceneage .Th eyounge rHolocen eterrain sar ecom monly divided into varzeas and igapos. In this publication,varze a is defined as low land that isintermittentl y waterlogged, and the igapo as lowland that is permanently waterlogged. The varzeas, which have the largest extent, are subdivided according to the character of the variation of the water levelan d to the composition of the water.

Levels offossil plinthite. Along the afore mentioned upper part of the Belem-Brasilia highway detailed observationswer emad eo nth eoccurrenc e oflevel so ffossi l plinthite (Appendices4 an d 5).Severa ltype so ffossi l plinthiteca nb ediscerne d inthi sarea , and each ofthe m hasa specifi c position and area ofoccurrence .I nth e stretchtraverse db y thehighway ,th e mainpar t ofth efossi l plinthitewa sforme d insitu. Itmus thav eorigi nated during at least two ages of theLat eTertiar ywit hflat lan d surfaces of imperfect drainage. An age of the Pleistocene produced plinthite of limited expanse. It islikel ytha tth efossi lplinthit ei nothe rpart so feaster n Amazonia wasforme d also largely during the Tertiary and the Pleistocene.

Vegetation. Isi sconclude d from thegeomorphologi cconstitutio n ofth ePlanici e part of Amazonia and the character of its sediments that both during the Tertiary and the Pleistocene there were several ageswit h relatively dry climates (interpluvials). During these agesther e wasn o forest coverage over large parts of Amazonia. The present-day vegetative cover is subdivided into several forest types, on the one hand, and several savannahs and savannah-forests, on theothe r (Fig.12) .

275 THE MAIN AMAZON LATOSOL Theconcept s on'latosol' , 'laterite'an d 'plinthite' haveha d already anevolution . The national Brazilian Soils Commission has recently given a detailed definition of the Latosol, in an elaboration ofth e U.S.A.concept . Attention isgive nt o the separation of Latosolsfro m otherre d andyello wsoil so fth etropic san d subtropics.I n particular, the characteristics of the diagnostic horizon of Latosols, the latosolic-B horizon, are contrasted (Table 5)wit h those ofth etextural- Bhorizon , which isdiagnosti c for Red YellowPodzoli csoil san d others. Manydifferen t termshav ebee napplie d toth evariou s Latosolsfoun d throughout the world. In Brazil, a subdivision of the Latosols is being worked out which takes as its principal criterion the composition of the clay fraction in relative amounts of silicate clay minerals (kaolinite), iron clay minerals (goethite, hematite), and aluminum clay minerals (gibbsite).Thes e amounts are assessed mainly on the molecular ratios Si02: Al203:Fe203. It is shown that the well-drained soils of the Planicie part of Amazonia have predo minantly a latosolic-B horizon as this is defined by the Brazilian Soils Commission. Themolecula r ratiosSi0 2:AI203:Fe203poin t toa stron gpredominanc e of kaolinite in their clay fraction, which isconfirme d by several X-ray curves of representative pro files (Tables 7an d 8,an d Fig. 14a-e). The soilstherefor e belongt o thesubgrou p ofth e kaolinitic Latosols and are named Kaolinitic Yellow Latosol. For thesand y relatives of the soils, with less than 15%cla y in their Bhorizon , the name used is Kaolinitic LatosolicSand . Thedescribe d Amazon soilsar ecomparabl e with severalo fth e soilsi ntropica l Afri ca,whic har egroupe d accordingt oothe rclassificatio n systemstha n used inthi s publi cation.

AMAZON SOILS WITH PLINTHITE The soils that contain plinthite in one form or another are discussed separately. In agreement with the Vllth Approximation, 'plinthite' here replaces the old word 'laterite', and a subdivision ismad e into hard plinthite (laterite concretions and slabs) and soft plinthite(Buchanan' s lateriteo r mottledclay) .

Soilswith recentplinthite. A t present plinthite in Amazonia mainly develops onflat land surfaces with intermittently imperfect drainage. Soft plinthite isforme d at some depth in the soil profile on these sites, viz.i n the zone of fluctuation of the phreatic level,whic h isfull y inagreemen t with theclassica lfindings o f MARBUT(1932) . In somecase sthi szon e occursa t such adept h that the solum ofth e soilprofil e isno t affected {cf. Profile description 1).Usually , however, the zone isneare r to the surface and clay-sized mineral particlesan d sesquioxides arecarrie d from the superficial layer downward. Animperfectl y drained soildevelop swit h aleache d Ahorizo n and a B ho rizon of soft plinthite, i.e.th e Ground Water Laterite soil.Ther e isa great variability incharacteristic so fthi ssoi luni t{cf. Profiles 2t o 18,an d Appendix 9),whic h isgovern ed by several factors. The ultimate stage of Ground Water Laterite soil development

276 appears to berepresente d byth e profile which hasa bleache d and light textured, thick A horizon over a heavy textured, dense and slowly permeable,thic k Bhorizo n of soft plinthite, whilst at the transition zone of the two horizons some plinthite material occurs that has already hardened (the Ground Water Laterite soil, Lowphase) .

Soils withfossil plinthite.I n Amazonia it is common to encounter layers of soft or hard plinthiteo nwell-draine d siteswher eth ephreati c leveloccur sa t a depth of many metres.I n viewo f the described present-day formation of plinthite, and in agreement with most of the recent literature on the subject, such plinthite is considered to be fossil. Layers of fossil hard plinthite underlain by fossil soft plinthite are likely to have formed insitu. Th e combined layers constitute the relics of a Ground Water Laterite soil which developed on a former land surface with imperfect drainage. After condi tions of imperfect drainage ceased to exist, the profile was geologically eroded to the point at which all the light textured and loose A horizon, and possibly part of the plinthitic B horizon, wasstrippe d off. Whilst the lowerpar t ofth eplinthiti c B horizon has remained soft, the upper still present part hasbecom e hard plinthite. Separate layerso ffossi l hard plinthite arenormall y ofcol luvial-alluvia lorigin . There are several supplementary characteristics that may help to establish whether plinthite isrecen t or fossil, and whether formed insitu o r ofcolluvial-alluvia l origin. Ifth efossi l plinthite isfoun d belowth esolum ,i ti sno t diagnosticfo r classification of the soil.If , however,th efossi l plinthite forms part ofth esolum ,the n the classification of the soil depends upon the degree of weathering of the parent material, i.e. in this caseth e plinthite. In Amazonia thefinal weatherin g product consists,wit h fewexcep tions,o f a mixture of hard concretionary elements andloos efriabl e eartho fwhic h the clayfractio n iskaoliniti ci ncharacter . Where there is shallow weathering of the plinthite layer the soil is classified as Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase. When the weathering isdeep , and in any case when thefossi l plinthite iso f colluvial- alluvial origin,th e soil isclassifie d as Kaolinitic Yellow Latosol, Concretionary phase (cf.Profile s 19 to23 ,an d Appendix9) . Agenera l schemefo r thedifferen t soilst o bedistinguishe d with theformation , trunc ation, and burying of Amazon plinthite isgive n in Fig. 15.

THE SOIL UNITS DISTINGUISHED, AND THEIR GEOGRAPHIC OCCURRENCE All discerned Amazon soil units(cf. summar y inTabl e9 )ar e described systematical ly,b ygivin ga genera l concept, the range incharacteristic , and a representative profile with itsanalytica l data. The soil mapso f twoo f thearea so f which areconnaissanc e soil surveywa sexecute d (Appendices 1 and 2),giv ea picture of the geographic occurrence of the varioussoils . The Guama-Imperatriz area may be taken as representative of the Planicie part of Amazonia with itsunconsolidate d sedimentso f kaolinitic claysan d quartz sands.Th e Araguaia Mahogany area may be taken as representative of the non-Planicie part of

277 Amazonia, with itswid e variation in geologic age and petrographic constitution. Fig. 17 isa sketc h of aprovisiona l soilma p for the wholeo f Brazilian Amazonia. The undulating terrains with outcropping crystalline basement probably have Red Yellow Latosol asth e predominant soil.Th e presenceo f Red YellowPodzoli csoi lan d of Lithosolsi sestablished . The undulating terrains with outcropping Paleozoic, Mesozoic or Early Tertiary deposits havea larg evariet y ofsoils .Lithosol sappea r tob ecommon . Very little soil data isavailabl e on thepeneplanatio n surfaces on thecrystallin e base ment. Most of the soil data of Amazonia relate to the Planicie and the Holocene terrains. The soilso fth e Belterra claycovere d Amazon planalto havea n uniform profile devel opment overlarg edistances .B yfa r themos tcommo n soili sKaoliniti cYello w Latosol (,Ortho), very heavy textured. Parts of the planalto in the south-western part of Amazonia have imperfect drainage, with Ground Water Laterite soil or Kaolinitic Yellow Latosol,intergrad et oGroun d Water Lateritesoil . The Pleistocene terraces in the eastern part of Amazonia have Kaolinitic Yellow Latosol (,Ortho), of medium, rather heavy, or heavy texture and Kaolinitic Latosolic Sand as the principal soils. Several aspects of profile development of these soils are related to the texture of the parent material. The terrace level also appears to have some influence (cf.Tabl e 10).Othe r well-drained soilsar e Kaolinitic Yellow Latosol, Compact phase; Kaolinitic Yellow Latosol, intergrade to Red Yellow Podzolic soil; and Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol; soils which are subdivided according to their texture. In addition, there are the plinthitic soils Kaolinitic Yellow Latosol, Concretionary phase and Red Yellow Podzolic soil, inter grade to Kaolinitic Yellow Latosol, Concretionary phase. Imperfectly drained soils are Ground Water Laterite soils and Ground Water (Hu mus)Podzol . There are many small patches with a peculiar dark and fertile soil which is called Terra Preta. This isa 'man-made' soil,develope d on the dwelling sites of the pre-Co lumbian Indian tribes. Judging from thedat a of MARBUTan d MANIFOLD(1926 )an d other informations, the soilso fth ePleistocen eterrace si nth ewester n part ofAmazoni a arelargel ycomparabl e withthos eo feaster n Amazonia. Onth evarzeas ,Lo w HumicGle yan d HurnicGle ysoil spredominate ,an d onth eiga - pos, Bogsan d Half Bogs. Also Ground Water Laterite soils are found on these low lands.I nth ecoasta l area Salinean d Alkali soilsoccur . For description of hitherto unknown forested areas at which the detailed geographic pattern of the various soilsca n beshown , the introduction of the concept Mand-unit* issuggested .Thi sdenote sa trac t ofcountr y withinwhic hth egeologic ,topographi can d hydrographic elements, the climate, pattern of vegetation and pattern of soils are approximately uniform.

278 THE SOILS AND THEIR VEGETATIVECOVE R The available data on the vegetative cover are of limited value for the assessment of the relationships between soils and vegetation, because methods of ecological re search propercoul d not beemployed . A discussion isnevertheles sdesirabl e invie wo f the near absence of published data on soil-plant relationships in Amazonia. Among the characteristics of the upland forests, the mean gross timber volume and the occurrence of individual tree species are discussed in particular, because it is on thesetha tth eFAO/SPVE A forest inventoriesprovid efo r data.

The influenceof non-edaphic factors ondifferences inforest growth. The highest timber volumes of the Planicie may be found in areas where the total annual rainfall is not excessively high and a short dry season occurs.A numbe r of differences established in the occurrence of individual tree species are likely to be determined by differences in climatea swell . Anthropogenic influences may have affected forests,eve n seemingly primeval ones, over large areas. An example is the selective cutting of certain very valuable tree species.A peculia r type of forest, whichconsist s largely of short-lived Guadua species (tabocal) and isfoun d in an area along the Bclem-Brasilia highway, isver y likely due to the burning practises of Indians.

Edaphically determineddefferences inforest growth.Nearl y all forest inventory areas are located inth e Planicie,th esoil so fwhic h arever y similar,a t theclassificatio n level employed. Established differences in forest characteristics within an area of uniform climatic conditions and non-existent or uniform anthropogenic influences, are there fore apt to show a correlation wit h lower category soil differences. Among these are themoistur eholdin gcapacity ,th etota lavailabl eamount so fth evariou splan tnutrients , and the penetration possibilities for roots. These qualities depend on such factors as thesoi ltexture ,th ecompactnes so fth esubsoi lan d thepresenc eo fplinthiti c materials. The coincidence of differences in forest characteristics wit h differences in these latter factors are discussed.I t is often uncertain what exactly are the causal factors of the coincidences established. Arelationshi p existsbetwee n soil texture and grosstimbe r volume.Th e rather heavy or heavy textured soils in particular support forest with high timber volume. Very heavy textured freely draining Planicie soils, the soils derived from Belterra clay, have in several parts forest of a low timber volume, because it consists fully, or for a considerable part,o fcreeper san d climbers:cipoa lan d 'cipoalic'forest . These forest types can be found on the completely flat central sections of planalto parts, as well as on the oldest Pleistocene terraces where these are covered with reworked Belterra clay. Cipoal and 'cipoalic* forest prevail on Belterra clay soils with a compact sub soil and a thin A, horizon compared wit h the soils of the immediate surroundings. With the aid of a number of examples the problem is discussed of any relationship between soils and individual tree species that have an established variation in occur rence.Acapi i (Vouacapouaamericana), forinstance ,i srestricte d to parts of the region

279 and there found in colonies. These colonies are however found on a variety of soils. One of the species found concentrated on very heavy textured soils is the Angelim pedra(Hymenolobium excelsum).Th eoccurrenc e of Pau amarelo (Euxylophorapara cusis) wasstudie d inth eGuamd-Imperatri zarea .I nthi sare ath especie sapparentl y has itsoptima l growing conditions on soil with concretions of fossil plinthite of a specific level andage . Aspecia l study wasmad e on the occurrence of the mahogany (Swieteniamacrophyl- la) ina part of thesouth-easter n transition zoneo f thebel t oftropica l forest, the Ara- guaia Mahogany area.I nthi sarea ,th especie smainl ygrow so nterrain s with imperfect drainage and well developed hydromorphic soils that chemically are relatively rich (Fig.2 5 andAppendi x6) . Theoccurrenc e of palmsi nth e Amazon forests seemst o be more often related to the variations in the conditions of climate and soil than is the case with the occurrence ofdicotylenou stre especies .

Non-edaphicsavannahs. Some upland savannahs within the forest belt mainly origin ated as a result of adverse non-edaphic conditions, usually with a predominance of anthropogenic factors. These savannahs are located inth e north eastern part of Ama zonia (upland savannahs in Amapa Territory, on south-eastern Marajo island and on thenort h bank ofth e LowerAmazo nriver) .

Savannahs andsavannah-forests ofedaphic origin. The other savannahs, and the sa vannah-forests, ofth e Planiciear eo fedaphi c origin.Al lth e savannahsconcerne d and most of the savannah-forests are found on terrains with imperfect drainage. These terrains either have a Ground Water Laterite soil (often the Low phase),o r a Ground Water (Humus) Podzol. Extensive savannahs and surrounding savannah-forests, on watershed areaswit hflat relie f(th ecampos ,an d part of thecampina-ranas) , havepre dominantly aGroun d Water Laterite soil.Patch y savannahsan d savannah-forests, on narrow strips of low sandy upland along rivers and on sand-filled former river beds (the campinas, the caatingas amazonicas, part of the campina-ranas), have predomi nantly the Ground Water Podzol as substratum. Some patches of savannah-forest (part of the campina-ranas) occur on relatively high, freely draining terrains, with WhiteSan d Regosol. Littledat a isavailabl eo nth esavannah san d savannah-forests ofth enon-Planici epar t of Amazonia.The yar e probably ofedaphi c origin,wit h Lithosolsan d Hydromorphic soilsa spredominan t substrata. The growing conditions of the variousdiscusse d savannahsan d savannah-forests are summarised inTabl e 13.

Soils and vegetation of the lowlands. As regards the lowlands, differences between their forests are often considerable and theyca n often becorrelate d with the character offlooding o r submergence. The lowland savannahs and savannah-forests of eastern Amazonia are considered to have originated edaphically.

280 THE SOILSAN D THEIRUS E An evaluation of the main Amazon soils from the viewpoint of their agricultural occupation is of immediate importance. Parts of Amazonia are namely being con sidered as outlet areas for the large rural population of the North Eastern region of Brazil with itsperiodicall y recurring severedrought s andfloods. With regard to the soils of the lowlands it isemphasize d that land capabilities vary much with the type of varzea. The main criteria for their evaluation must be the quemical qualitieso fth e soilsan d thedept h and lengtho f theflooding. Th e necessary artificial drainage and reclamation would require considerable investigation, organi sationan dcapita l investment. The soils of the uplands outside the Planicie are very diverse.Th e natural fertility of thesoil si scertainl y aprincipa l criterion for land capability evaluation inthes e distant and inaccessibleareas .

Qualities of thefreely draining kaoliniticsoils of the Planicie.Th e terrains of the Planicie constitute large tracts of predominantly flat to gently undulating land which islargel y freely draining.Thes e terrains arenea r the main water-waysan d theexistin g roads. Their forests have the higher timber volumes, and the present-day Amazon agriculture is concentrated here. For these reasons the Planicie part of Amazonia, despite the low natural chemical fertility of its soils, has a potential for large-scale settlement schemes. Special attention is therefore given to the chemical and physical qualities, the agricultural capabilities and the adequate management measures of Planicie soils with good external drainage, i.e.th e freely draining kaolinitic Planicie soils.

Qualities of thesoils under theirnatural vegetative cover. For the assessment of the chemical qualities of the freely draining kaolinitic Planicie soils under their natural forest cover, analytical data of 35 relevant profiles are available. These are arranged and studied onthei r relationship (cf. Figs.2 7t o 34).I t isapparent that with regard to chemical qualities the main advantage of theclaye y over thesand y soils isthei r capa city tocreat e a better milieu for preservation of thecomparativel y very active organic matter. Inthei r physicalqualitie s thesoil sca ndiffe r distinctly from each other.Th e moisture holding capacity of the soils is of importance in view of the fact that a large part of Amazonia hasa short dry season. From moisture equivalents, some moisture tension curves,an d otheranalytica ldat a (cf.Figs . 35t o3 7an d Table 16),i t isdeduce d that the soils under forest cover have, as a whole, only small amounts of available moisture per volume unit. Clayey soils have a fairly distinct advantage over sandy soils,wit h regard to moisture availability, only as regardsth e topsoil.

Qualitiesof the soils under human influence. Th echemica lan d physical qualitieso fth e Planicie soilschang e in several ways, when these soilscom e under influence of man. For the soils under shifting cultivation, only qualitative observations are evaluated.

281 For those under ancient anthropogenic savannah, however, a number of quantitative data are available. When the difference in the amount of nutrients stored in the re spective vegetative covers is not counted, it can be said that the qualities of these savannah soilsar e still fairly comparable to those of forest soils.Th e main differences are a somewhat lower organic matter content of the topsoil, and consequently a somewhat lower potential cation exchangecapacit y and smalleramount so fexchange able cations and anions, greater soil compactness and presumably lessavailabl e soil moisture. Slight differences are found in the C/N ratios, the valuesfo r pH-H20,an d thedegre eo fphosphorou s fixation. The permanently beneficial effect of human influence is illustrated by the chemical qualities of theTerr a Preta(cf. Figs.3 8an d 39).Th e potential cation exchange capa cityo fthi ssoi li srelativel yhigh ,als owhe nth eorgani cmatte rconten t islow .However , this high cation exchange capacity corresponds with a high amount of phosphates in the Terra Preta.Thi samoun t often seemsconsiderabl e highertha n thefixing powe ro f thehumus-poo rsoi lmaterial . The effects of chemical fertilizers and of animal and green manures can be assessed only tentatively, because of the limited number of experimental plots and sites with permanent agriculture.

Themost suitable soils.I t iseviden t that land capability evaluations for the Planicie should be principally based on such differences in soil qualities as organic matter storage, soil moisture reserve and root penetrability. For large-scale settlements, external factors should be considered, such as the topography and drainage pattern ofth eland ,it saccessibility ,an dth edegre eo fgeographi ccoherenc eo fth evariou ssoils . Apart from this,however , and thesoi lpreference s ofindividua l crops,th e most favou rablecondition sfo r cropproductio n aret ob efoun d onrelativel yheav ytexture d forest soils that have non-compact and friable subsoils. The Kaolinitic Yellow Latosol (,Ortho)an d Kaolinitic Yellow Latosol, intergradet o Red Yellow Podzolicsoi lwhic h are ofheav y or rather heavytextur e meetthes erequirement s best.

So/7 management. The examination of the chemical and physical qualities of the freely draining kaolinitic Planicie soils under forest cover or under human influence, enable an assessment to be made ast o the best methods of managing these soils after the primeval forest cover has been cleared, and of recuperating them after adverse humaninfluence .Th emaintenanc eand ,wher epossible ,enlargemen t ofth esoi lorgani c matter content iso f paramount importance. At the actual state of research, fertilizing recommendationsca nonl yb etentative . Shifting cultivation must constitute the basis for improvement and expansion of peasant farming on the Planicie soils. The present system of shifting cultivation in Amazonia can begraduall y improved in several ways.Th e introduction of perennials, especially oil palms, seems desirable. The combination of crop production with modern management of the forests isa feasibility.

282 Sumario

Os solos da parte brasileira de regiao amazonica constituem a materia do presente estudo. Como fronteiras daquela regiao sao tomadas as transudes do cinturao dafloresta equatorial para as savanas do Nordeste ed o Brasil Central epar a as do territ6rio do Rio Branco e da Guiana inglcsa. Em particular i discutida a metade da parte Leste daregiao .

FATORES PEDOGENETICOS Clima. Apenas ume pequena parte da Amazonia brasileira apresenta urn climase m qualquer estacaosec a(tip o 4/"naclassificaga o de Koppen, Fig.2) .N a maior parte da regiao hd pouca ou nenhuma precipitagaodurant e alguns mesesd o ano (tipo Aw).

Geologia. A Amazonia brasileira consiste de uma area baixa e sedimentar sendo o valeamazonic opropriament edito ,e parte sdo scontraforte s cristalinosd o Brasil Cen trale da sGuiana s(Fig .5) .Essa sformacoe scristalina sremonta m aoperiod oPre-cam - briano e sao constituidas principalmente de granitos, gneisses e micaxistos. Na parte sedimentar encontram-se quatro bacias,nominalment ea do Acre,a bacia amazonica propriamente, ad a ilhad e Marajo e a do Maranhao(Fig .6 e Apendic e 8).N a super- ffcie apresentam-se apenas faixas relativamente estreitas de depositos Paleozoicos- Mesoz6icos decarateristic a muito variavel; a maior parte da area sedimentar constitui-se de sedimentos Terciarios nao consolidados, que se compoem de argilas caoli- niticase arei a dequartzo . Os sedimentos Pleistocenos, cuja composigao seassemelh a aosTerciarios ,sa o poucoespesso s ea su a extensao,conform e aosdado sd eliteratura , 6 limitada. Tembem os sedimentos Holocenos apenas compreendem uma area limitada, muito inferior aestimativ a dos primeirosexploradore s da Amazonia.

Geomorfologia. Nas regioes dos divisores de agua ao lado Norte e Sul do sistema fluvial do Amazonas, pode-se distinguir as seguintes unidades geomorfol6gicas (Fig. 8,9e 17). 1. terrenos ondulados de embasamento cristalino aflorante. 2. terrenos ondulados com depositos aflorantes do Paleozoico, Mesoz6ico ou Ter- ciario Inferior. 3. dois niveis depeneplanaca odentr od aare a queno smapa sgeologico sfica indicad a como cristalina. Estes niveis ocorrem a 400-600m e a 250-400m d e altura respeti-

283 vamentee 6 provavel queseja m daidad eCretace a ed oTerciari oInferior . Destesnivei s ha,porem ,pouco sdado scertos . A larga parte axial da Amazonia consiste de: 4. terras planase m platoPlio-pleistoceno ,chamad oo planalto amazonico, 5. terras firmes mais baixas, modeladas em terracos a niveis diferentes, de idade Pleistocena (As unidades 4. e 5.conjuntament e sao indicadas como Planicie). 6. baixadas, de idade Holocena.

Na parte Leste do vale, o planalto tern em geral 150 a 200 m. de altura; na parte Oeste 6u m pouco mais baixo. Uma carateristica do planalto €qu e ele apresenta na sua superficie uma camada, de dez a vinte metros de espessura, composta de uma argila sedimentar homogena de textura muito pesada. Esta argila, chamada argila de Belterra, ha de ter sua origem em depositos caoliniticosnosAndes,que-ap6 s serem elevados- foram expostosa erosa oforte .A argil ater-se- adepositad o de maneira uniforme e gradativa na base plana de um amplo mar interior de pouca profundidade. Uma grande parte da Amazonia deve ter sido coberta por esse mar durante certo periodo doTerciari o Superior ou doPleistocen o Inferior,quand o onive lgera ld o mar era elevado (Calabriano). Os terrenos modelados durante o Pleistoceno,n a parte axial da Amazonia, cobrem uma superficie grande. O tracado de 500 kilometros de comprimento da parte superior da recem-construida rodovia entre Bel£me Brasilia ofereceu a possibilidade de estudar em detalhes a geomorfologia local (Apendices 4 e 5). Neste trecho apresentam-se varios terracos. Alguns deles sao costeiros, os demaisfluviais. Todo s devemte rsid oformado s emtempo se mqu eo nive ld ema rer a maiselevad oqu eatual - mente,e portant o devem serd e idade interglacial. Comparando osnivei sdo s terracos com osnivei sglaciai sd o mar, ate mesmoresult a possivel determinar provisoriamente emdetalhe sa s idadesdo sterraco s numa parte da area emaprec o (Tabelas 1 e2) . E' provavel que todos os terracos nas demais partes do vale amazonica propria- mentedito ,tambe m sejam deidad e interglacial. Os chamados terrenos de massape\ de ocorrencia muito local, e partes da Marajo oriental remontam provavelmented o Holoceno Inferior. Osterreno s Holocenos mais recentes sao comumente subdivididos em varzeas e igapos. Neste estudo a varzea i definida como baixada intermitentemente inundada, e o igap6 como baixada permanentemente inundada. As varzeas, que constituem a maiorpart eda s baixadas, sao subdivididas de acordo com ocarate r dasflutuagoes d onfve l deagu a esu acomposicao .

Niveisdo 'plinthite'fossil. Ao longo da referida parte superior da rodovia Belem- Brasilia foi feito um estudo detalhado da ocorrencia de niveis de 'plinthite' fossil (Apendices 4e 5).O 'plinthite* fossil duro nesta area podedividir-s e em varios tipos e cada um deles tern sua posigao e area de ocorrencia espccificas. Deduziu-se que a maior parte do 'plinthite' fossil foi formada insitu. Deve ter tido sua origem durante pelo menos duas epocas do Terciario Superior, quando nesta area havia superficies

284 planas de drenagem imperfeita. Durante alguma £poca Pleistocena tambem deve ter sido formado 'plinthite', seja em quantidade Jimitada. E' provavel quetambe m nasdemai sparte sd aAmazoni a o'plinthite 'fossi l mormente fosse formado duranteo Terciari oe o Pleistoceno .

Vegelacdo. Da composicaogeomorfol6gic a da Planicie amazonica ed o carater dos sedimentos presentes 6 deduzido que assim durante o Terciario como durante o Pleistoceno havia varias6poca sco mclima srelativament eseco s(interpluviais) , duran te as quais grandes partes da Amazonia nao tinham cobertura florestal. A cobertura vegetal atual & subdividida em varios tipos florestais por um lado e varias savanase florestas-savanas po r outro lado (Fig. 12).

O LATOSOLO AMAZONICO MAIS IMPORTANTE Osconceito s'latosol' ,Materite 'e 'plinthite'j a terntid o umaevoluca ohist6rica . A Co- missao de Solos do C.N.E.P.A. ha pouco deu uma elaboracao extensa do conceito 'latosol' na base usada nos Estados Unidos. Nela sepresto u atencao para a maneira de distinguir os Latosolos deoutr o solosvermelho se amarelo sdo str6pico se subtro- picos.Nomeadament ea s carateristicas do B-Iatosolico e do B-textural, que sao os horizontes carateristicos deLatosolos ,respetivament ed esolo s Podz61icos Vermelho- Amarelo e outros, sao contrastadas. Os varios Latosolos no mundo tern recebido as mais variadas denominacoes. Hoje esta sendo desenvolvido no Brasil um sistema desubdivisa o dos Latosolos, usando-se como principal criterio a composicao da fra<;ao de argila, a saber a miitua relacao entre as quantidades minerais-de-argila de silicato (caolinita), minerais-de-argila de ferro (goetita, hematita) eminerais-de-argil ad ealumini o(hidrargilita) .Esse sminerai s sao avaliadospo r meioda srazoe smoleculare s SiO^:Al 2Oa:Fe 203. Fica demostrado que a maior parte dos solos bem drenados da Planicie amazonica tern um B-latosolico conforme a definicao da Comissao de Solos. Ha neles uma predominancia muito forte de caolinita em sua fracao argilosa, o que fica comfirmado por alguns Roentgengramas de amostras representativas de solo (Fig. 14a-e e as Tabelas 7 e 8). Esses solos pertencem portanto ao grupo dos Latosolos caoliniticos, e neste estudo sao denominados Latosolo Amarelo Caolinitico. Os solos arenosos comparaveis com eles, ou seja os que no seu horizonte Bconte m menos de 15%d e argila, sao denominados Areia Latos6Iica Caolinitica. Os solos amazdnicos acima dcscritos sao iguais a varios solos da Africa tropical, agrupados segundo outros sistemasd e classificacao que o usado neste estudo.

SOLOSAMAZOMCO SCO M 'PLINTHITE' Os solos que contem 'plinthite' em qualquer forma sao discutidos separadamente. Conforme a Seventh Approximation,o 'plinthite' substitui a palavra antiga laterita. O 'plinthite' fica subdividido em 'plinthite* duro (cascalho e pedras de laterita) e 'plinthite* macio (laterita de Buchanan: argila mosqueada).

285 Solosde 'plinthite" recenie. Aformaca o atuald c'plinthite 'n aAmazoni a ocorrequas e exclusivamcnte nos terrenos pianos de drenagem intermitentemente impcrfeita. Em tais lugares i formado o 'plinthite' macio a pouca profundidade, a saber na zona na qual oscila o ni'vel freatico, perfeitamente de acordo com as achadas classicas dc MARBUT (1932). Em alguns casos a zona em apreco ocorre a ta! profundidade que nao pode afetar o solum do perfil (Perfil I). Usualmcntc, por6m, a zona acha-se t5o pr6xima a supcrficic que lixiviam as particulas minerais maisfina s e os sesqui6xidos da zona superficial. Forma-se entao um perfil de drenagem imperfeita com um horizonte A eluviado e um horizonte B que se compoe de 'plinthite*macio : o perfil Laterita Hidromorfica. Ha grande variabilidade nas carateristicas desta unidadc de solo (Perfis 2-18 e Tabela 9), regulada por varios fatores. Como condicao final da evolugao Laterita Hidromorfica pode considerar-se o perfil que se compoe de um horizonte Aespesso ,alvejad o e arenoso e um horizonte B espesso, argiloso e pouco permeavel de'plinthite *macio ,co m- nazon a detransica o- algum material de *plinthite'jaendurecido(neste estudo denominado solo Laterita Hidrom6rfica, fase Baixa).

Soloscom ''plinthite' fossil. Muitas vezes ocorrem na Amazonia zonas de 'plinthite' duro ou macio em lugares bem drenados onde o nivel freatico apcnas se encontra a muitosmetro sd eprofundidade . Tendoe mvist aa formaca o recented e'plinthite *com o descrita acima, este i considerado como fossil, o que 6d e acordo com a maior parte da literatura mais recente sobre este assunto. As zonas de 'plinthite* duro fossil, que sobrepoem uma de'plinthite* macio fossil usualmenteter n sido formadas insitu. A szona sconstitue m conjuntamente o relicario de um perfil Laterita Hidromorfica, que se desenvolveu numa superficie anterior, de drenagem imperfeita. Este perfil, apos superados os impcdimentos de drenagem, tera sofrido uma erosao de tal magnitude, que desapareceu todo o horizonte A arenoso e solto, e talvez parte do B argiloso de 'plinthite*. Enquanto a parte mais profunda do B de'plinthite ' ainda presente permaneceu macia,a parte superior vinha transformando-se em 'plinthite' duro. Camadas individuals de 'plinthite' duro fossil sao normalmcnte de origem aluvial- coluvial. Ha varias carateristicas suplementarcs utcis para esclarecer sio 'plinthite' i recente ou fossil, es i foi formado insitu o u 6d e origem aluvial-coluvial. Quando o'plinthite ' fossil apcnass eencontr a por baixod o solum,na o importa para a classificacao do solo. Quando o 'plinthite* f6ssil constitui uma parte do solum, a classificacaodess esol odepend ed o graud emcteorizaca od o'plinthite* ,qu ee m talcas o constitui a rocha-mae. Na Amazonia o material mctcorizado final consistc quase sempre de uma mistura dc elcmentos concrccionarios duros e terra friavel, cuja fracao de argila tern caratcr caolinitico. Onde a mcteorizacao da zona de 'plinthite*apcna s tern pouca profundidade, o solo e classificado como solo Pod/61ico Vermclho-Amarclo, 'intcrgradc* para Latosolo AmarcloCaolinitico ,fas eConcrccioniria .Quand oa mcteorizaca od a zona eprofunda , o solo 6 denominado Latosolo Amarelo Caolinitico, fase Concreciondria. A ultima

286 dcnominacao €scmpr c aplicavcl quando o 'plinthite' f6ssil e de origcm aluvial-coluvial(Pcrfi s 19-23e Apendice9) . Na Figura 15 i dada uma esquema geral para os difcrentes solos que se podem distinguir na formacao, respctivamcntea truncaca oe cobrimcntaca od c'plinthite ' na Amazonia.

AS UNIDADES DE SOLO DISTINGU1DAS ESU A OCORRENCIA GEOGRAF1CA As diferentcs unidades de solo da Amazonia conhecidas (veja resumo da Tabela 9) sao dcscritas sistematicamcnte, dando o conceito geral, as possiveis variacdes nas caratcristicas e um perfil representativo com seusdado s analiticos. Os mapas de solo de duas das areas de levantamento expedito (Apendices 1e 2) dao uma idcia da ocorrencia espacial das diferentes unidades de solo. A primeira area de levantamento expedito, a area Guama-Imperatriz, pode ser tomada comou m exemplo da Planicie, com seus sedimentos nxioconsolidados , compostos de argilas caoliniticas eareia sd equartzo . Aare a Araguaiana de Mognoi exemplo da parte da Amazonia fora da Planicie.com sua amplavariaca oe midad egeo!6gic ae constituica o petrografica. AFig . 17& umesbdc od cu m mapa desolo spar a toda a Amazonia. Nos terrenos ondulados com embasamento cristalino aflorante, ocorre como solo provavelmente mais importante, o Latosolo Vermelho-Amarelo. Ocorrem tambem solos Podz6IicosVermelho-Amarel o e Litosolos. Ha grande variedade de solos nos terrenos ondulados de depositos aflorantes do Paleoz6ico, Mesoz6ico ou do Terciario Inferior. Provavelmente os Litosolos sao os que ocorrem com maior frequencia. Contamos com muito poucos dados sobre os solos dos niveis de peneplanacao na area do embasamento cristalino. A maior parte dos conhecimentos pedologicos sobre a Amazonia sao concernentes a Planicie e as baixadas Holocenas. Os solos do planalto amazonico coberto de argila de Bclterra aprescntam um desenvolvimento de perfil uniforme sobre grandes distancias. Predomina o Latosolo AmareloCaolinitic o(,Orto )d etextur a muito pesada. Algumas partes do planalto na zona sudoeste de Amazonia apresentam uma dre- nagcm impcrfeita, de forma que la o solo i de Laterita Hidrom6rfica ou Latosolo Amarelo Caolinitico, 'intergradc' para solo Laterita Hidrom6rfica. Os solos principais dos tcrracos Picistocenos da parte leste da Amazonia sao Lato solo Amarelo Caolinitico (,Orto) de textura m£dia, meio pesada ou pesada, e seu equivalcnte dc textura leveo u muito levc,a Areia Latosolica Caolinitica. Varias ten- dencias no desenvolvimento do perfil desses solos demostram uma relacao com a textura dos sedimentos. Tambcm o nivel de terraco faz um papel, o que possibilita uma avaliacaod a inMuenciad o fator tempo na formacao de solo (Tabela 10).Outro s solos bem drcnados sao o Latosolo Amarelo Caolinitico, 'intergrade' para solo Podz6lico Vermelho-Amarelo; Latosolo Amarelo Caolinitico, fase Compacta e solo Podz61icoVermelho-Amarelo.'intergrade ' para Latosolo AmareloCaolinitico ; solos esses que sao subdivididos de acordo com a textura do horizonte B. Em adicao ha ossolo sco m'plinthite ' f6ssil:o Latosol o AmareloCaolinitico .fas eConcrecionari a co

287 solo Podz61ico Vermelho-Amarelo, 'intergrade' para Latosolo Amarelo Caolinitico, fase Concrecionaria. Os principals solos imperfeitamente drenados sao os solos Laterita Hidrom6rfica e osPodzol s Hidromorficos. Nos terracos pleistocenos de Amazonia oriental ocorrem ocasionalmente solos que sao marcadamenteescuro se ferteis , denominadosTerr a Preta.Sa osolo santiopogeni - cos, desenvolvidos nos provoados das tribos de lndios pre-colombianos. Conside- randoo sdado sd eMARBU Tan d MANIFOLDe outro ssobr eest aarea ,o ssolo sdo sterraco s Pleistocenos na parte oeste da Amazonia devem ser por grande parte iguais aos do lested a Amazonia. Nasvarzea s predominam ossolo sGle i Pouco Hiimico eo ssolo sGle i Hiimico,e no s igap6s, solos turfosos. Tambem ocorrem solos Laterita Hidrom6rfica. Na zona costeira ocorrem solossalino se alcalinos . Para descrever areasflorestais at e hoje desconhecidas, prestando-se atencao para o padrao detalhado dosdiferente s solos,e introduzido oconceit o de 'unidade de terra'. Como tal e considerada uma area dentro da qual as condicoes geologicas, topogra- ficas ehidrograficas , oclima ,o padra o davegetaca oe o padra odo ssolo ssa oiguais .

RELACOESENTR EO SSOLO SE A VEGETACA O Para determinar as rela<;6es entre os solos e a vegetacao da Amazonia, os dados disponiveisapena ster nvalo r limitado,vist oqu eaind ana opudera m serfeita snalgum a escala pesquisas decarate r puramente eco!6gico.Todavia , por faltarem quase inteiramente dados publicados sobre a relacao solo e vegetacao na Amazonia, 6 aconsel- havel uma discussao do assunto. Entre ascarateristica s dafloresta, nomeadament e sao discutidos ovolum e bruto de madeira ea ocorrenci a deespecie s individuais dearvores ,poi se sobr eeste selemento s que existem dados, devidos a uma serie deinventario s florestais.

A influencia defatores ndo-edaficos nas diferencas na vegelafdoflorestal. O s mais elevados volumes de madeira da Planicie podcm ser encontrados nas areas onde a precipitacaoanua l nao €muit oelevad a eond e ha estacaoscc ad ecurt a duracao. E provavel quepart e dasdiferenca s na ocorrencia deespecie s individuais dedrvore s tambem sejam devidas a diferencas climatologicas. E possivel que areas extensas de florestas aparentemente primitivas podcm ser afetadas por fatores antropogenicos. Umexempl o 6 o corte seletivod ecerta s madei ras muito valiosas. Um tipo peculiar de floresta constituida em grande pane de es peciesGuadua d ecurt ocicl od evid a (tabocal) equ e ocone emcert o trecho darodovi a Belem-Brasilia, com grande probabilidade pode ser atribuido a praticas de incendi- mentod epart edo slndios .

Diferencas deorigem eddfica na vegetacaoflorestal. Quas e todas as areas de inventario florestal acham-se localizadas na Planicie, onde os solosdedrenagemlivre.no nivel aplicado de classificacao.apresentam um alto grau de semclhanca. Assim, dife-

288 rencas estabelecidasna scarateristica s defloresta, qu e ocorrem em areas decondicoe s climatol6gicas uniformes e de falta ou ocorrencia uniforme de influencias antropogenicas,ha od ese restudada sn asu acorrelaca oeventua l coma sdiferenca s encontradas nascarateristica sd esolo sd ecategori a maisbaixa , comoa capacidade deretenca o de dgua, a soma total dos varios nutrientes disponiveis para asplantas , ea s possibi- lidades de penetracao de rafzes. Estas qualidades dependem de fatores tais como a textura do solo, a compacidade do subsolo e a presenca de materia de 'plinthite' . Resulta que diferencas em carateristicas florestais freqiientemente coincidem com diferencas nestes fatores. Em geral €dificil , porem, estabelecer quais sao os fatores motivadores destas coincidencias. Existeum arelaca oentr etextur a do solo e volume bruto de madeira. Em particular ossolo sd etextur a pesada emeio-pesad a freqiientemente suportamflorestas co m alto volume de madeira. Os solos de textura muito pesada da Planicie, porem, ou seja os solosdesenvolvido s na argila de Belterra,e mvario strecho sesta ocoberto sco m flores tas de baixo volume de madeira,po rconsisti r inteiramente oupo rpart econsiderave l decip6 s (cipoal oufloresta cipoalica) . Estestipo s defloresta ocorre m no centro com- pletamente piano dos terrenosd e planalto,be mcom o nos terracos Pleistocenos mais antigos nas partes onde estes estao cobertos com argila de Belterra remodelada. Deduziu-se que ocipoa l ea floresta cipoalic a prevalescem nos solos de argila de Bel terraque ,comparado sco mo ssolo sna simediacoes ,apresenta m umsubsol o compacto e um horizonte Aj delgado. Parece que tambem em outras partes da Amazonia as florestas ricase mcip6 sprevalesce m emsolo sd epouc a penetracaod eraizes ,be mo sd e umabaix acom oo sd eum aalt afertilidad e natural. Com algunsexemplo s6 descutida arelaga opossive lentr e solose especie s individuals de arvores de distribuicao desigual. O Acapii (Voucapoua amerkana), por exemplo, apenas ocorre em algumas partes da regiao, onde 6 encontrado em grupos. Estes grupos,porem ,ficam e m solosvariados .A Angeli m pedra (Hymenolobium excelsum) £um ada sespecie sencontrad a mormente emsolo sd etextur a muito pesada. Aespeci e Pau amarelo (Euxylophoraparaensis) na area Guama-Imperatriz, onde foi estudada sua ocorrencia, encontra as suas melhores condicoes de crescimento em solos de concrecoes de 'plinthite' fossil de determinado nivel e idade. Uma investigacaoespecia l foi feita dos lugares decresciment o de Mogno (Swietenia macrophylla),e mpart ed azon ad etransica osudest ed ocintura o dofloresta equatorial . Resulta que ncsta area a especie cresce quase exclusivamente em terrenos de drena- gemimperfeit a comsolo shidrom6rfico s bemdesenvolvidos ,especialment e osqu equi - micamente sao ricos (Fig.2 5e Apendice 6). A ocorrencia de especies de palmeiras nas florestas amazonicas parecequ emai ss e acha relacionada com as variacoes nas condicoes de clima e solo do que e o caso na ocorrencia de especies de arvores dicotilas.

As savanas nao-edaficas. Algumas savanas de terra firme dentro do cinturao de floresta principalmcntefora m causadas por fatores adversosnao-edaficos , usualmente com uma predominancia de influencias antropogenicas. Essas savanas acham-se

289 localizadas na parte nordeste da Amazonia: o territ6rio do Amapa, a parte sudestc da ilha de Marajo ea part e ao Iadonort e do BaixoAmazonas .

Savanase florestas-savanas de origem eddfica. As outras savanas da Planicie e as florestas-savanas sao de origem edafica. Essas savanas eflorestas-savanas acham-s e em terrenos de drenagem imperfeita. Os solos dessessa o ou solo Laterita Hidrom6r- fica - muitas vezes a chamada fase Baixa dele - ou Podzol Hidrom6rfico. Exten- sas savanase asflorestas-savanas qu e ascircundam , encontradas nos terrenos pianos nos divisores de agua (os campos; parte dascampina-ranas) ,encontram-s eprincipal - mente em solos Laterita Hidrom6rfica. As savanas e florestas-savanas de pouca extensao,qu e ocorrem em faixas estreitas deterren o baixo earenos o ao longod erio s ou sobre antigos leitos derio senchido s de areia (ascampinas ; ascaatinga s amazonicas; parte das campina-ranas), ocorrem sobretudo em Podzols Hidrom6rficos. Alguns lugares de floresta-savana (parte das campina-ranas) acham-se em terrenos relativamente altos de drenagem livre, com Regosolo de Areia Branca. Ha" poucosdado sdisponfvei s sobrea s savanase a sflorestas-savanas for a da Planicie. E provavel que sejam de origem edafica eficam morment e em Litosolos e solos hi- drom6rficos. Na tabela 13 aparece um esquema das condicoes de crescimento das diferentes savanas eflorestas-savanas discutidas .

Solose vegetacdo nasbaixadas. As mutuasdiferenca s entre asflorestas da s baixadas sao muitas vezes claras, e e possivel relaciona-las com o carater da inundacao. As savanas eflorestas-savanas qu e ocorrem nas baixadas do este da Amazonia sao de origem edafica.

A POTENCIALIDADE DOS PRINC1PAIS SOLOS AMAZON1COS PARA FINS AGRICOLAS Uma avaliacao dos principals solos amazonicos parafins agricola si deimportanci a imediata. Isto e porque partes da Amazonia sao consideradas como podendo ser usados de escoadouros para a densa populacao rural do Nordeste, regiao esta perio- dicamente sujeita a sccas ou inundacoescalamitosas . Com respeito as baixadas ha de salientar que seu valor agricola depende muito do tipo de varzea. Oscriterio s dominantes para sua avaliaclo terao de ser asqualidade s quimicas de seus solos e o carater da inundacao. Os melhoramentos hidraulicos da terra requereriam investigacao, organizagao e investimento de capital numa escala vultosa. Os solos de terra firme fora da Planicie sao muito diversos. A fertilidade quimica natural do solo ecertament e um criterio essencial para avaliar as potencialidades da terra naquelasdrea sremota se d e dificil acesso.

Qualidades dos soloscaoliniticos da Planicie de drenagem livre. APlanici ecompoe-s e mormente de terrenos pianos ou levementeondulado s dedrenage m livre. Estesterre nos acham-se nas proximidadcsdo s principals cursos de agua e das poucas rodovias

290 existentes na regiao. Asflorestas naquela s areas tern um volume de madeira relativamente alto e asrireas agricola s amazonicasatuai sficam concentrada s nestesterrenos . Assima Planicie amazonica, embora a fertilidade quimica natural dosol osej a baixa, representa umapossibilidad e para acolonizaga o agricolae m grande escala. Portanto convem prestar aten<;aoespecia l para as qualidades qufmicas efisica s daqueles solos da Planfcie que tenham boa drenagem externa, os chamados solos caoliniticos da Planfcie de drenagem livre.

Qualidades dos solossob suavegetafdo natural.Par a discutir as qualidades quimicas dos solos caoliniticos da Planicie de drenagem livre sob sua vegetacaoflorestal natural, contamosco m dados analiticos de 35perfis . Estesdado s sao agrupados ees - tudados em suas miituas relacoes (Figs 27-34) E' evidente que a capacidade de criar um meiomai sadequad o a preservacaod a matdriaorganic arelativament e muitoativ a e a principal vantagem dos solos argilosossobr e as solos arenosos quanto as qualida desquimicas .A sdiferenca s entreo sdiverso ssolo sreside m por parte consideravel nas suas qualidades fisicas. Importa a capacidade de retengao de agua dos solos, visto que grande parte da Amazonia tem uma estacao seca de alguma importancia. De equivalentes de umidade, algumascurva sd e tensao de umidade eoutro s dadosanali ticos(Figs .35-3 7e Tabel a 16)deduz-s e que ossolo s sobsu a vegetacaoflorestal, vist o no conjunto, apenas tem pequenas quantidades de agua aproveitavel por unidade de volume.Apena sn a zonae mqu es each aquantidad e consideravel demateri a organica, ossolo s argilosos- no que diz respeito a quantidade deagu a disponivel - tem vanta gem evidente sobre os solos arenosos.

Qualidadesdos solos sob influencia humana. Quand oo ssolo scaolinitico sd aPlanici ed e drenagem livre chegam sob influencia humana, vao modificando-se suas qualidades quimicase fisicas . Para ossolo susado s para agricultura itinerante apenas ha observa- c6esqualitativas .Par a ossolo sda ssavana santropogenica sexistente sh a muitotempo , contamos,porem/co mu mnumer od edado squantitativos .NS oconsiderand oa diferen - gan aquantidad e denutriente s armazenados nasvegetacoe s respetivas,podemo s dizer que, no que diz respeito as qualidades quimicas e fisicas, os solos das savanas em apreco ainda se assemelham bastante aos solos sob floresta. As principais diferencas saoa sseguintes :um a quantidade algomeno rd emateri a organicad a camada superior do solo, maior densidade de solo de topo e possivelmente menor quantidade de agua disponivel. Pequenas diferencas foram registradas nas razoes C/N, nos valores para o pH-H20 e no grau defixacao d e f6sforo. Os efeitos beneficos que a influencia humana pode ocasionar, sao ilustrados pelas propriedades quimicas da Terra Preta (Figs.3 8e 39) .A capacidad e total detroc a dos cations deste solo £ relativamente elevada, mesmo onde o teor da materia organica seja baixo. A alta capacidade de troca dos cations na Terra Preta,por£m ,va i acom- panhada degrande squantidade s defosfatos , as que parecem muitasveze s serconsi - deravelmente maiores que o poder de fixacao do solo pobre em humus. Osefeito sd eadubo squimicos ,estrum ee adubo sverdes ,apena spode mse raproxima -

291 dos em parte, por ser limitado at6 hoje o numero de quadros de ensaio e terras de agricultura permanente.

Ossolos tnais adequados. A estimativa do potenticial agricola dos solos da Planicie de drenagem livre ha de basear-se principalmente nas diferencas das qualidades dos solos tais como o teor de materia organica, a capacidade de retencao de agua e a penetrabilidade por raizes. Para projetos de colonizacao agricola em grande escala i natural que tambem desempenham um papel fatores externos,com o a topografia e o padrao dedrenage m da terra, sua acessibilidade eo grau de coerenciageo - grdfica dos varios solos. Nao considerando, porem, estes fatores e as exigencias das culturas individuals, pode-se afirmar que ascondicoe s maisfavoravei s para a agricul tura se encontram nos solos florestais de textura relativamente pesada, de subsolo nao-compacto. O Latosolo Amarelo Caolinitico (,Orto) de textura pesada e meio pesada e o Latosolo Amarelo Caolinitico,'intergrade 'par a soloPodzolic oVermelho - Amarelo saoo squ e melhors epresta m para aquela finalidade.

Manejo daterra. O exame das qualidades quimicas e fisicas dos solos sobfloresta primitiva ou sujeitos a influencia humana, da indicagoes para os meios mais adequa dos para manter a potencialidade dos solos apos ser derrubada a floresta primitiva, bem como para seu melhoramento apos uma interferencia humana adversa. A con- servacaoe - na medidad o possivel- oaument o doteo rd a materia organica dossolo s £d e suma importancia. Considerando o estado atual das investigacoes, apenas pode- rao seraproximativa sa srecomendacoe srelativa sa fertilizacao. Sera preciso tomar o sistema itinerante como ponto de partida para o fomento da agricultura. Aos poucos podera ser melhorada em muitos respeitos a aplicagao que este sistema tern hoje na Amazonia. Parece recomendavel uma ampIiaQao da area com perenes arboreos, particularmented o dende. Podera ter vantagens acombinaca o da agricultura com uma moderna explotacao das florestas.

292 Samenvatting

Het onderwerpva ndez estudi e isd e bodemgesteldheid van het Braziliaanse deelva n het Amazonegebied. De overgangszone van de gordel tropisch bos naar de savannen van Noord-oost en Centraal Braziliee ndi eva n Rio Branco-Brits Guiana zijn alsgrenze n voor het gebied aangenomen. In het bijzonder wordt aandacht geschonken aan deoostelijk e helft van het gebied.

BODEMVORMENDE FACTOREN Klimaat. Een klein gedeelte van Braziliaans Amazonie'heef t eenklimaa t geheelzon - derdrog e tijd (4/"typ ei nd eclassificati e van Koppen; Fig.2) ;i nhe t grootste deelva n het gebied valt gedurende een paar maanden van hetjaa r weinig of geen regen {Am type).

Geologie.Braziliaan s Amazonie bestaat uit een laag sedimentair gebied, d.i. de eigenlijke Amazonevallei, en gedeelten van de kristallijne schilden van Centraal Bra zilie en de Guianas (Fig. 5). De kristallijne schilden dateren van het Pre-Cambrium en zijn voornamelijk uit granieten, gneizen en glimmerschisten samengesteld. In het sedimentairegedeelt eworde nee n viertal geologische bekkens aangetroffen: het Acre- bekken,he teigenlijk e Amazone-bekken,he t Marajo- enhe t Maranhao-bekken (Fig. 6 en Appendix 8). Slechts relatief smalle banen met Paleozoische-Mesozoische afzet- tingenva n zeerverschillen d karakter komenaa n deoppervlakte ;he t grootste gedeelte van het sedimentaire gebied wordt ingenomen door Tertiaire onverharde sedimenten die bestaan uit kaolinietkleien en kwartszanden. De Pleistocene sedimenten, die een soortgelijke samenstelling alsd eTertiair e hebben,zij n dune nd eoppervlakt e ervanis , volgens de literatuurgegevens, beperkt. Ook de Holocene afzettingen beslaan slechts een geringe oppervlakte, veel minder dan de eerste onderzoekers van Amazonie heb ben vermoed.

Geonwrfologie. Ind ewaterscheidingsgebiede n aan de noord- end ezuidzijd e van het Amazone-rivierensysteem kan men de volgende geomorfologische eenheden onderscheiden (Fig. 8,9e n 17):

1. heuvelachtige terreinen met onbedekt kristallijn grondgebergte. 2. heuvelachtige terreinen met onbedekte Paleozoische, Mesozoische of Vroeg- Tertiaire afzettingen.

293 3. Twee schiervlakte-niveaus binnen het gebied dat als kristallijn op de geologische kaarten staat aangegeven. Dezeniveau s komen voor op respectievelijk 400-600m e n 250-400 m hoogte en zijn waarschijnlijk van Krijt- en Vroeg-Tertiaire ouderdom. Er bestaan echter weinig zekere gegevens over de betrokken niveaus. Het brede asgedeelte van Amazonie' bestaat uit: 4. Plio-Pleistoceen vlak plateauland, aangeduid als Amazoonse planalto 5. Niet-overstroombare Iagere terreinen, gemodelleerd tot terrassen van verschillend niveau,e nva n Pleistocene ouderdom. Dezetwe eeenhede n tezamenworde n aangeduid alsPlanicie . 6. Holoceen, overstroombaar laagland.

In hetoostelij k deelva nd evalle ii sd eplanalt o meestal 150-200m hoog,i nhe tweste - lijke gedeelte ietslager . Kenmerkend voor deplanalt o isd eaanwezighei d vanee ntie n tot twintigmete rdikk eoppervlaktelaa g bestaande uit homogene,zee rzware ,sedimen - taire klei. Deze zogenaamde Belterraklei moet afkomstig zijn van kaolinitische afzettingen in het Andesgebied dien a opheffing eertijds sterk aan erosie onderhevig waren. De kleimoe tzee rgelijkmati g engeleidelij k zijn bezonken opd evlakk ebode mva nee n uitgestrekte ondiepe binnenzee. Een groot gedeelte van Amazonie moet door zulkee n zee bedekt zijn geweest gedurende een deel van het Laatste Tertiair of het Vroegste Pleistoceen, toen dealgemen e zeestand hoog was(Calabrium) . De terreinen diegemodelleer dwerde ngedurend e het Pleistoceen beslaan in het as gedeelte van Amazonie een grote oppervlakte. Het 500k mlang etrac e van het bovenste gedeelte der onlangs opengelegde Belem-Brasilia weg bood gelegenheid een uitvoerige studie van de plaatselijke geomorfologie te maken (Appendices 4 en 5). In dit gebied blijken verscheidene terrassen voor te komen. Enkeleerva n zijn kustterras- sen, de andere rivierterrassen. Alle moeten in perioden met een hogere zeestand dan de huidige zijn gevormd en daarom van interglaciale ouderdom zijn. Door de terras- hoogten te vergelijken met interglacialewereldzeestande n blijkt zelfsvoo r deterrasse n in een gedeelte van het bestudeerde gebied een voorlopige gedetailleerde ouderdoms- bepaling mogelijk tezij n (Tabellen 1 en2) . Waarschijnlijk zijn ook de terrassen in de andere delen van de eigenlijke Amazone- vallei alleva n interglaciale ouderdom. De zeer plaatselijk voorkomende zogenaamde massape-terreinen en gedeelten van oostelijk Marajo zijn vermoedelijk van Vroeg-Holocene ouderdom. De jongere Holocene terreinen worden vaak onderverdeeld in varzeas en igapos. In deze publicatie wordt laagland dat bij tussenpozen onder water staat als varzea beschouwd en laagland dat voortdurend onder water staat als igapo. De varzeas, die het grootste deel van het laagland uitmaken, worden onderverdeeld naar het karakter van de schommeling der waterstand end esamenstellin gva n het water.

Niveausvan fossiele plinthiet. Langs het eerdergenoemde bovendeel van de Belem- Brasilia weg werden uitvoerige waarnemingen omtrent het voorkomen van fossiele plinthietniveaus gedaan (Appendices 4e n 5). De fossiele harde plinthiet in dit gebied

294 blijkt in verscheidene typen te kunnen worden ingedeeld, en elk van deze typen heeft een eigen positie en gebied van voorkomen. Het grootste gedeelte van de fossiele plinthiet is blijkbaar ter plaatse gevormd. Het moet ontstaan zijn gedurende tenmin- ste twee perioden van het Laat-Tertiair, toen in dit gebied vlakke en onvolledig gedraineerde landoppervlakten aanwezig waren. Gedurende een deel van het Pleisto- ceen moet ook plinthietvorming,zi j het in geringe mate, hebben plaatsgehad. Waarschijnlijk isoo k defossiel e plinthiet ind eander edele n van Amazonie vooree n belangrijk gedeelte gedurende het Tertiair en het Pleistoceen ontstaan. Vegetatie.Ui t de geomorfologische opbouw van het Planiciegedeelte van Amazonie en het karakter van de aanwezige sedimenten, wordt opgemaakt dat er zowel tijdens het Tertiair alstijden s het Pleistoceen verscheidene tijden geweest moeten zijn met een relatief droog klimaat (interpluvialen), gedurende welke grote gedeelten van Amazonie geen bosbedekking hadden. Het huidige plantendek wordt onderverdeeld in verschillende bostypen enerzijds en diverse savannen en savannebossen anderzijds (Fig. 12).

DE BELANGRIJKSTE AMAZOONSE LATOSOL De begrippen 'latosol', 'lateriet' en 'plinthiet' hebben reeds een historische ontwikkeling doorgemaakt. De nationale Braziliaanse Bodem Commissie heeft onlangs het begrip latosol, op basisva n de in de Verenigde Staten gebruikte opvatting, uitvoerig beschreven. Daarbij is aandacht besteed aan de wijze waarop Latosolen van andere rode en gele gronden in detrope n en subtropen onderscheiden kunnen worden. Met name de kenmerken van de latosol-B en van de textuur-B (de karakteriserende hori- zontenva nLatosole nrespectievelij k RedYello wPodzolic se nander egronden )worde n gedetailleerd tegenover elkaar gesteld (Tabel 5). De diverse Latosolen in de wereld zijn met vele namen benoemd. Op het ogenblik wordt in Brazilieee n systeem van onderverdeling van de Latosolen ontwikkeld, waar- bijd esamenstellin g vand ekleifracti e - deonderling everhoudin gva nd ehoeveelhede n silikaatkleimineralen (kaoliniet),ijzerkleiminerale n(goethiet ,hematiet )e naluminium - kleimineralen (gibbsiet)- alsd ebelangrijkst emaatsta f wordt gehanteerd.Dez everhou - dingenworde ngescha tme tbehul pva nd emoleculairverhoudinge n Si02:A1 203:Fe 203. Hetmerendee l van degoe d gedraineerdegronde n vanhe t Planiciedeelva n Amazonie blijkt eenlatosol- B tebezitte n zoalsdi edoo r de Braziliaanse Bodem Commissiegede - finieerd is. De moleculairverhoudingen duiden opee n zeer groot percentage kaoliniet in de kleifractie, hetgeen bevestigd wordt door enkele Rontgen-diffractielijnen van karakteristieke grondmonsters (Fig. 14a-e,e n deTabelle n 7e n 8).D e gronden beho- rendaaro mto td egroe pva nd ekaolinitisch eLatosole ne nworde nverde ral sKaoliniti c Yellow Latosol aangeduid. Voor dehierme et e vergelijken zandige gronden, namelijk diewelk ei nhu n B-horizont minder dan 15% lutumbevatten ,word t denaa m Kaolini tic Latosolic Sand gebruikt. De bovenbeschreven Amazoonse gronden komen overeen met verscheidene van de gronden in tropisch Afrika, die gerangschikt zijn volgens andere classificatiestelsels dan in deze publicatie gebruikt zijn.

295 AMAZOONSE PLINTHIETGRONDEN De gronden die plinthiet in enigerlei vorm bevatten worden afzonderlijk besproken. Plinthiet vervangt hier, in overeenstemmingme t de Seventh Approximation, het oude woord lateriet. Plinthiet wordt onderverdeeld in harde plinthiet (laterietgruis en -stenen) en zachte plinthiet (Buchanan's lateriet; 'mottled clay').

Gronden met recenteplinthiet. De huidige vorming van plinthiet in Amazonie' vindt bijna uitsluitend plaats op vlakke terreinen met een bij tussenpozen onvolledigedrai nage. Op zulke plaatsen ontstaat op een geringe diepte zachte plinthiet, namelijk in de grondlaag waarover de grondwaterstand schommelt, een waarneming die volledig in overeenstemming isme td eklassiek ebevindinge n van MARBUT(1932) .I nee n enkel gevallig td ebetrokke n grondlaag tedie po mno gva n invloed tezij n ophe t solumva n het bodemprofiel (Profiel 1).Meesta lechte r bevindt delaa gzic hz odich t bij de oppervlakte,da t defijnste mineral edeeltje s end esesquioxide n vand ebovenlaa g uitspoelen. Er ontwikkelt zichda n eenonvolledi ggedraineer d profiel metee n uitgeloogde A-horizont, en een B-horizont die bestaat uit zachte plinthiet: het Ground Water Laterite profiel. Er is een grote variatie in de kenmerken van deze bodemeenheid (Profielen 2t/ m 18 en Appendix 9),waaraa nverscheiden efactore n tengrondsla gliggen .Al seind - toestand in de Ground Water Laterite-ontwikkeling kan men het profiel beschouwen dat bestaat uit een gebleekte en zandige, dikke A-horizont en een kleiige en slecht- doorlatende,dikk eB-horizon t vanzacht eplinthiet , meti nd eovergangslaa g enigreed s verhard plinthietmateriaal (in deze publikatie aangeduid als Ground Water Laterite soil, Lowphase) .

Gronden metfossiele plinthiet. Dikwijl s komen in Amazonie op goed gedraineerde plaatsen, waar de grondwaterstand pas op vele meters diep te vinden is, lagen van harde of zachte plinthiet voor. Zulke plinthiet wordt, gezien de beschreven recente plinthietvorming,al sfossie l beschouwd, hetgeen in overeenstemming isme t een groot deelva n dejonger e literatuur over dit onderwerp. Lagen van fossiele harde plinthiet, rustend op een laag fossiele zachte plinthiet, zijn doorgaans ter plaatse ontstaan. De lagen vormen tezamen het overblijfsel van een Ground Water Laterite profiel, ontstaan op een vroegere landoppervlakte met onvol ledige drainage. Het betreffende profiel moet, nadat de drainagebelemmeringen weg- vielen, zodanig aan erosie onderhevig zijn geweest, dat de gehele zandige en losse A-horizont enwellich t een gedeelte van de kleiige plinthiet-Bverwijder d werden. Ter- wijl het diepste deelva n deno gaanwezi g gebleven plinthiet-Bzach t isgebleven , ishe t bovenste deel ervan tot harde plinthiet veranderd. Afzonderlijke lagen fossiele harde plithiet zijn gewoonlijk van alluviale-colluviale oorsprong. Er zijn verscheidene aanvullende kenmerken die behulpzaam zijn bij het vaststellen of plinthiet recent dan wel fossiel ise n ter plaatse isgevorm d of van allu viale-colluviale oorsprong is. Wanneer de fossile plinthiet zich slechts beneden het solum bevindt, is het niet van belangvoo r declassificati e van de grond. Wanneer de fossiele plinthiet deel uitmaakt

296 van het solum, wordt de classificatie van de grond bepaald door de verweringsgraad van de plinthiet, diei n dit gevalhe t moedeimateriaal vormt. In Amazonie bestaat het uiteindelijke verweringsprodukt bijna altijd uit een mengsel van harde concreties en losse kruimelige aarde waarvan de kleifractie een kaolinitisch karakter heeft. Wanneer de verwering van de plinthietlaag slechts ondiep is, wordt de betreffende grond geclassificeerd als Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase. Wanneer de laag diep verweerd is, wordt de grond Kaolinitic Yellow Latosol, Concretionary phase genoemd. De laatste benaming is steedsva ntoepassin gwannee rd efossiel e plinthiet vanalluviale-colluvial e oorsprongi s (Proflelen 19 t/m 23 en Appendix9) . Fig. 15 geeft eenoverzich t van deverschillend egronde n diebi jd evorming ,respectie - velijk de onthoofding en de bedekking van plinthiet in Amazonie onderscheiden kunnen worden.

ONDERSCHEIDEN BODEMEENHEDEN EN HUN GEOGRAFISCHVOORKOME N Alle bekende Amazoonse bodemeenheden (zie samenvatting in Tabel 9) worden systematisch beschreven door middel van een algemeen concept, de mogelijke variatie in kenmerken en eenrepresentatie f profiel met bijbehorende laboratoriumgegevens. Debodemkundig e overzichtkaarten vantwe ekarteringsgebiede n (Appendices 1 en2 ) gevenee nindru k vanhe truimtelijk evoorkome n vand everschillend e bodemeenheden. Het eerste karteringsgebied, de Guama-Imperatriz area, kan men als voorbeeld beschouwen voor het Planiciedeel van Amazonie, met zijn onverharde sedimenten bestaande uit kaolinietkleien enkwartszanden . DeAraguai a Mahogany area isee nvoor beeld voor het niet-Planiciedeelva n Amazonie, met zijn groteverscheidenhei d ingeo - logische ouderdom en gesteenten. Een voorlopige bodemkaart voor geheel Brazi- liaans Amazonie geeft Fig. 17. In de heuvelachtige terreinen met onbedekte kristallijn grondgebergte worden als waarschijnlijk belangrijkste gronden Red Yellow Latosols aangetroffen. Red Yellow Podzolic soils en Lithosols komen echter ook voor. Een grote verscheidenheid van gronden komt voor in deheuvelachtig e terreinen met onbedekte Paleozoische, Mesozoische of Vroeg-Tertiaire afzettingen. Lithosols zijn waarschijnlijk het veelvuldigst. Omtrent de gronden van de schiervlakte-niveaus in het gebied van het kristallijne grondgebergte zijn zeer weinig gegevens beschikbaar. Het grootste deel van de bodemkundige kennis omtrent Amazonie heeft betrekking opd e Planiciee no p deHolocen e laaglanden. Degronde n van deme t Belterrakleibe - dekte Amazoonse planalto vertonen over grote afstand een gelijke profielontwikkeling. Verreweg het meest komt zeer zware Kaolinitic Yellow Latosol (,Ortho) voor. Bepaaldestukke n planalto inhe tzuidwestelij k deelva nAmazoni ehebbe nee nonvolle - dige drainage, zodat de grond daar Ground Water Laterite soil of Kaolinitic Yellow Latosol, intergrade to Ground Water Laterite soil is. Devoornaamst e gronden van de Pleistocene terrassen in het oostelijk deelva n Ama zonie zijn matig lichte, matig zware of zware Kaolinitic Yellow Latosol (,Ortho) en

297 het lichte of zeer lichteequivalen t hiervan de Kaolinitic Latosolic Sand. Verscheidene aspecten van de profielontwikkeling van deze gronden vertonen een verband met de textuur vanhe tmoedermateriaal .Oo k de terrashoogte speelt blijkbaar een rol, waar- door een schatting van de invloed van de factor tijd bij de bodemvorming mogelijk is (Tabel 10).Ander e goed gedraineerde gronden zijn Kaolinitic Yellow Latosol, intergrade to Red Yellow Podzolic soil; Kaolinitic Yellow Latosol, Compact phase en Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol; alle onderverdeeld naar de zwaarte van de B-horizont. Dan zijn er de plinthietgronden Kaolinitic Yellow Latosol, Concretionary phase en Red Yellow Podzolic soil, intergrade to Kaolinitic Yellow Latosol, Concretionary phase. Ground Water Lateritesoil se nGroun d Water (Humus)Podzol szij n de belangrijkste onvolledig gedraineerde gronden. Opd ePleistocen eterrasse n vanoostelij k Amazonie'worden plaatselijk gronden aangetroffen die opvallend donker en vruchtbaar zijn en bekend staan als Terra Preta. Dit zijn 'man-made' soils, ontwikkeld op de voor-Columbiaanse woonplaatsen van Indianenstammen. De gronden van de Pleistocene terrassen in het westelijk dcel van Amazonie zijn, afgaande o.a. opd egegeven sva n MARBUTan d MANIFOLD over dit gebied, grotendeels gelijk aan dieva n oostelijk Amazonie. In devarzea szij n Low HumicGle ye n HumicGle ysoil she t meestverbreid , en ind e igapos Bog en Half Bog soils. Ground Water Laterite soils komen ook voor. In het kustgebied liggen zoute en alkali-gronden. Voor beschrijving van tot nu toe onbekende bosgebieden, waarbij het gedetailleerde geografische patroon van de verschillende gronden aandacht kan krijgen, wordt het begrip land-unit ingevoerd. Als zodanig wordt opgevat een gebied waarbinnen de geologische, topografische en hydrografische omstandigheden, het klimaat, het pa troon van devegetati e en het patroon van de gronden gelijk zijn.

SAMENHANGEN TCSSEN BODEM EN PLANTENDEK Voor bepaling van de samenhangen tussen bodemgesteldheid en plantengroei in Amazonie zijn de beschikbare vegetatiegegevens slechts van beperkte waarde, aan- gezien onderzoekingen van zuiver ecologische aard nog niet op enige schaal uitge- voerd konden worden.Wegen s het nagenoeg ontbreken van gepubliceerde gegevens over Amazoonse bodem-plant verhoudingen, isee n bespreking niettemin wenselijk. Vand ebostyperend eelemente n worden inhe tbijzonde r hetbrut o houtvolumee nhe t voorkomen van afzonderlijke boomsoorten besproken, aangezienjuis t hierover gege vens beschikbaar zijn via een reeks bosinventarisaties.

De invloed van Miet-edafischefacloren op verschillen in bosgroei.D e hoogste hout- voluminava n de Planicie magme n daar verwachten waar detotal cjaarlijks e regenval niet alt e hoog ise n een kort droog seizoen voorkomt. Een gcdeelteva n de verschillen invoorkome n van afzonderlijke boomsoorten isnaa rall ewaarschijnlijkhei d ook door klimaatsverschillen bepaald.

298 Antropogene factoren kunnen giote gebieden met ogenschijnlijk oorspronkelijk bos beinvloed hcbben. Een voorbceld daarvan is het selectievekappe n van bepaalde zeer waardcvolle houtsoorten. Een eigenaardig bostype, dat voomamelijk uit koit- levendeGuadua-soorte n bestaat (tabocal;bamboe) enlang see ngedeelt eva nd eBelem - Brasilia weg voorkomt, is zeer waarschijnlijk veroorzaakt door indiaanse brand- gewoonten.

Edafisch bepaalde verschillen inbosgroei. Bijn a alle bosinventarisatiegebieden liggen in de Planicie, waarvan de vrij drainerende gronden op het toegepast classificatie- niveau ingrot cmat egelij k zijn.Vastgesteld everschille n inboskenmerken ,voorkomen d binnen een gebied met gclijke klimaatsomstandigheden en ontbreken of gelijkmatig voorkomen van antropogene invloeden, moeten daarom bekeken worden op hun eventueel verband met verschillen in grondeigenschappen van een lagere orde, zoals het vochthoudend vermogen, de totaal beschikbare hoeveelheid van de verschillende plantenvoedingsstoffen en de bewortelingsmogelijkheden. Deze eigenschappen wor den bepaald door factoren zoals de zwaarte van de grond, dedichthei d van de ondergrond end eaanwezighei d van plinthietmateriaal.Verschille n inboskenmerke n blijken dikwijls samen te vallen met verschillen in deze factoren. Welkefactore n precies deze waargenomen co-incidenties veroorzaken, isechte r meestal nietvas t testellen . Er bestaat een samenhang tussen zwaarte van degron d en bruto houtvolume. In het bijzonder de zware en matig zware gronden kunnen bos met een hoog houtvolume dragen. Zeer zware Planiciegronden echter - de gronden ontwikkeld op de Belterra- klei- zijn op verscheidene plaatsen bedekt met bosva n een laaghoutvolume ,doorda t het geheel of voor een belangrijk gedeelte uit lianen bestaat (cipoal ofcipoali c forest). Deze bostypen komen voor op het volledig vlakke centrum van planalto-terreinen en op de oudste Pleistocene terrassen voorzover die bedekt zijn met opnieuw afgezette Belterraklei.Cipoa l encipoali c forest lijken opdi e Belterrakleigronden tegroeie n die, in vergelijking met gronden van de nabije omgeving, een dichte ondergrond en een dunne Ai-horizont bezitten. Ook eldeis in Amazonie lijken lianenrijke bossen voornamelijk voor te komen op ondiep doorwortelde gronden, zowel die met een Iageal s die met een hoge natuurlijke vruchtbaarheid. Aan de hand van enkele voorbeelden wordt het verband nagegaan dat kan bestaan tussen bodemgesteldheid en afzonderlijke boomsoorten met een ongelijkmatige ver- spreiding. DeAcapt i (Vouacapoua americana) bijvoorbeeld wordt alleen gevonden in sommige gedeelten van het gebied, waar de soort dan groepsgewijze groeit. Deze groepen komen echter voor op een aantal verschillende gronden. DeAngeli m pedra (Hymenolobiumexcelsum) is een van de soorten die voomamelijk op zeer zware gronden wordtaangetroflen . De sooit Pauamarelo (Euxylophoraparaensis) heeft in het Guama-Impcratriz gebied, waar het voorkomen onderzocht werd, zijn gunstigste groeivoorwaarden op gronden met fossiele plinthietconcreties van bepaalde hoogtee n oudcrdom. Een afzonderlijk onderzoek werd ingesteld naar de groeiplaatsen van Mahonie(SH/Wem' amacrophylla) di e in een gedeelteva n dezuidoostelijk e overgangszone van de gordel tropisch bos voorkomt. Het blijkt dat in dit gebied de soort bijna

299 uitsluitend groeit op terreinen met een onvolledigedrainag e en steik ontwikkel d hydromorfe gronden, en wel speciaal die, welke in chemisch opzicht relatief rijk zijn (Fig. 25e n Appendix 6). Het voorkomen van palmsoorten in de Amazonebossen lijkt vaak meer verband te houden met verschillen in klimatologische en bodemgesteldheid dan het voorkomen van dicotyle boomsoorten.

Niet-edafische savannen. Enkele savannen binnen het bosgebied zijn ontstaan voornamelijk door ongunstige niet-edafische factoren, waarbij antropogene invloeden meestal hebben overheerst. Deze savannen liggen in het noordoostelijke deel van Amazonie: Amapa-territoiium, het zuidoostelijke deelva n het eiland Marajo, en de noordzijde van de beneden-Amazone.

In oorsprong edafischesavannen ensavannebossen. De andere savannen van de Plani- cie en de savannebossen zijn edafisch van oorsprong. Deze savannen en de meeste van de savannebossen komen voor op terreinen met een onvolledige drainage. Het bodemtype van zulke terreinen isofwe l een Ground Water Laterite soil - en dan vaak de zogenaamde Low phase hiervan - ofwel een Ground Water (Humus) Podzol. Uitgestrekte savannen en savannebossen die deze omringen, aangetroffen op vlakke waterscheidingsteneinen (de campos; een gedeelte van de campina-ranas), komen voornamelijk op Ground Water Laterite soils voor. Pleksgewijze voorkomende savannen en savannebossen, op smalle banen laag en zandig terrein langs rivieren of op verzande rivierbeddingen (de campinas; de caatingas amazonicas; een gedeelte van de campina-ranas), staan voornamelijk op Ground Water Podzols. Sommige plekjes savannebos (een deel van de campina-ranas) komen voor op relatief hoge, vrij drainerende terreinen met White Sand Regosols. Over de savannen en savannebossen buiten het Planfciegebied bestaan slechts weinig gegevens. Waarschijnlijk zijn zij edafisch van oorsprong en staan ze voornamelijk op Lithosolen en hydromorfe gionden. Tabel 13 geeft een schema van de groeiomstandigheden der verschillende besproken savannen en savannebossen.

Bodem en plantendek van de overstroombare laaglanden. De onderlinge verschillen tussen laaglandbossen zijn vaak zeer duidelijk en zij kunnen veelal gemakkelijk in ver band gebracht worden met het karakter der overstroming. De savannen en savanne bossen die op de laaglanden van oostelijk Amazonie voorkomen zijn van edafische oorspiong.

GESCHIKTHEID VAN- DE BELANGRUKSTE AMAZOONSE GRONDEN VOOR LANDBOUW- DOELEINDEN Een waardebepaling van de belangrijkste-Amazoonse gronden met betrekking tot hun ingebruikname voor landbouw is van onmiddellijk belang. Aan gedeelten van Ama zonie wordt namelijk gedacht als uitwijkgebieden voor de dichte landelijke bevolking

300 van het Noordoosten van Brazilie, waar periodiek calamiteuze droogten en over- stromingen optreden. Wat de laaglandgrondcn betreft dient benadrukt te worden dat hun gebruikswaarde sterk afhang tva n hettyp evarzea . Dechemisch eeigenschappe n van degronde n enhe t karakter van de overstroming zullen bij de bodemgeschiktheidsbepaling de belangrijkste criteria moeten zijn. Denoodzakelijk e inpoldering zou veelonderzoek , organisatie en kapitaalsinvestering vereisen. Degronde n van het hogere land buiten de Planicie zijn zeer verschillend. In deze af- gelegene n moeilijk toegankelijke gebieden isd enatuurlijk e chemische vruchtbaarheid van de grond ongetwijfeld het belangrijkse criterium voor de bodemgeschiktheids bepaling.

De vrijdrainerende kaolinitische Planicie-gronden. De Planicie bestaat grotendeels uit vlakke of zwak golvende terreinen met een vrije drainage. Deze terreinen liggen dicht bij debelangrijkst e waterlopen en deenkel ebestaand e wegen.He t hieropgroei - ende bos heeft een relatief hoog houtvolume en het huidige Amazoonse landbouw- areaal iso p dezeterreine n geconcentreerd. Daarom biedt het Planicie-deel van Ama- zonie, hoewel de natuurlijke chemische vruchtbaarheid van de gronden laag is,voor - uitzichten voor landbouwontwikkelingsprojecten van grote omvang. Het is dan ook gewenst bijzondere aandacht te besteden aan de chemische en fysische eigenschappen vandi ePlanicie-gronde n welke een goede externe drainage bezitten, oftewel: de vrij drainerende kaolinitische Planicie-gronden.

Degronden onder hun natuurlijke begroeiing. Voor het benaderen van de chemische eigenschappen van de vrij drainerende kaolinitische Planicie-gronden onder hun na tuurlijke bosgroei staan laboratoriumgegevensva n3 5 profielen ter beschikking. Deze worden gerangschikt eno phu n samenhangbekeke n(Fig .2 7t/ m 34).He t blijkt duidelijk dat het vermogeno mee nbete r milieu tescheppe n voor deinstandhoudin g van de relatief zeer actieve organische stof het belangrijkste voordeel van de kleiigebove n de zandige gronden is. De diverse vrij drainerende Planicie-gronden verschillen ook dikwijls in hun fysische eigenschappen. Het vochthoudend vermogen van de gron den isva n belangme t het oog op het feit dat een groot deel van Amazonie een droog seizoenva nenig ebetekeni sheeft . Uitvochtequivalenten ,enkel e vochtspanningscurven en andere laboratoriumgegevens (Fig. 35 t/m 37 en Tabel 16) wordt afgeleid dat de gronden onder bosbedekking in hun geheel gezien slechts kleine hoeveelheden op- neembaar vocht per volume-eenheid bezitten. Slechts voor de laag waarin een bedui- dende hoeveelheid organische stof aanwezig is,zij n dekleiig e gronden wat dehoeveel - heid beschikbaar vocht betreft duidelijk in het voordeel boven de zandige gronden.

Degronden ondermenselijke invloed. Wanneer de vrij drainerende kaolinitische Pla nicie-gronden onder menselijke invloed komen, ondergaan de chemische en fysische eigenschappen verschillendewijzigingen . Omtrent degronde n diei n gebruik zijn voor shifting cultivation bestaan slechts kwalitatieve waarnemingen. Over de gronden van

301 delangbestaand eantropogen e savannen zijn echter eenaanta l kwantitatieve gegevens bcschikbaar.Indie nme nhe tverschi li nd ehoeveelhei d voedingsstofTen diei nd eplanten - groei is opgeslagen buiten beschouwing laat, kan gesteld worden dat qua chemische enfysisch e eigenschappen debetrokkensavannegronde n nogtamelij k dicht bijd ebos - grondenIiggen . Alsvoornaamst everschille n kunnen genoemd wordenee niet skleiner e hoeveelheidorganisch esto fi nd ebovengrond.ee ngroter edichthei d vand ebovengron d en vermoedelijk een kleinere hoeveelheid beschikbaar bodemvocht. Geringe verschil lentrede n op metbetrekkin g tot deC/ N verhouding, depH-H 20 waardee nd e fosfor- fixatie. Dat menselijke invloed ook een duurzamegunstig e uitwerking opd egron d kanheb - ben, wordt gei'Ilustreerd door de chemische eigenschappen van de Terra Preta (Fig. 38 en 39). Het kationenuitwisselingsvermogen van deze grond is relatief hoog, ook dan wanneer het organische stofgehalte slechts gering is.We igaa t het hoge kationen uitwisselingsvermogen gepaard metgrot ehoeveelhede n fosfaten ind eTerr a Preta; het fosfaatgehalte isdikwijl s aanzienlijk groterda nhe tfixatievermogen va nd ehumusarm e grond. Het effect van kunstmest, stalmest en groenbemesters kan slechts enigermate bena- derd wordenomda t het aantal proefveldjes enplaatse n met blijvende landbouw tot nu toebeperk tis .

Demeest geschikte gronden. D ebepalin gva n debodemkundig e gebruikswaarde van devri jdrainerend ePlanfcie-gronde n voor landbouwdoeleinden dient voornamelijk te steunen op de verschillen in grondeigenschappen als de organische stofhuishouding, het vochthoudend vermogen en de doorwortelbaarheid. Bij landbouwontwikkelingsprojecten van grote omvang spelen uiteraard externe factoren zoals de topografie en het drainagepatroon van het land, de bereikbaarheid en de mate van geografische samenhangva n deverschillend egronde n ook eenrol .Doc hafgezie n vandez e factoren end eeise nva n afzonderlijke gewassen,ka n gesteld wordenda t degunstigst e mogelijk- heden voor landbouw te vinden zijn op de relatief zware bosgronden met een niet- verdichte ondergrond. De zware en matig zware Kaolinitic Yellow Latosol (,Ortho) en Kaolinitic Yellow Latosol, intergrade to Red Yellow Podzolic soil zijn in dit opzicht het meest geschikt.

Wijze van grondgebruik. Hetonderzoe k naard echemisch ee nfysisch e eigenschappen van degronde n onder oerbos of onder menselijke invloed, geeft aanwijzingen voor de manieren waarop de gronden het best op peil gehouden kunnen worden na verwijde- ring van het oerbos en kunnen worden verbeterd na schadelijke menselijke beinvloe- ding. Het instandhouden en zo mogelijk verhogen van het organische stofgehalte van degron d ishierbi j van overwegend belang. Bemestingsadviezen kunnen bij dehuidig e stand van het onderzoek slechts globaal zijn. Voor deverbeterin g en uitbreiding van debevolkingslandbou w zal het shifting culti vation systeem als uitgangspunt genomen moeten worden. De wijze waarop dit systeem in Amazonie wordt toegepast, kan op verscheidene manieren geleidelijk ver-

302 beterd worden. Een vergroting van het areaal met boomgewassen, in het bijzonder oliepalm, lijkt wenselijk. Combinatie van de landbouw met een moderne bosexploi- tatie biedt gunstige vooruitzichten.

303