Soil Formation by Termites a Study in the Kisii Area, Kenya

Soilformatio n by termites a study in the Kisii area, Kenya

0000 Promotoren: dr. ir. J. Bennema, emeritus hoogleraar in de tropische bodemkunde

dr. L. van der Plas, persoonlijk hoogleraar in de bodemkunde en geologie W. G. Wielemaker

Soil formation by termites a study in the Kisii area, Kenya

Proefschrift ter verkrijging van de graad van doctor in de landbouwwetenschappen, op gezagva n de rector magnificus, ü v , },-,r H E E K dr C C Oosterlee OKR ^NDBOUWHOGESCHOOt in het openbaar te verdedigen WAGENINGEN op woensdag 24 oktober 1984 des namiddags tevie r uur in de aula van de Landbouwhogeschool te Wageningen.

No part of thispublicatio nma y be reproduced, stored ina retrieval system,o r transmitted inan y form or byan ymeans ,mechanical , photocopying, recording, or otherwise,withou t theprio r written permission of theauthor .

Copiesma y be purchased from theDepartmen t of Soil Science and Geology, Agricultural University,P.O . Box 37,670 0A A WAGENINGEN, The Netherlands. The price ofDfl . 16,-include s postage. At)0-l ?v

STELLINGEN

1.Biologisch e processen ind e bodem zijn vanwezenlij k belang voor het voortbestaane n deontwikkelin g vankapitaal-arm e landbouwbedrijven in de tropen.

Dit proefschrift.

2.He t verdient aanbeveling meer onderzoek tedoe naa n demanipulati e van de bodemfauna.

Dit proefschrift.

3.D e biologische activiteit draagt, sterk bij tot eenverminderin g van de erodeerbaarheid vangronden . Bijd e bepaling van deK-fakto rword t hiermee over het algemeengee n rekening gehouden.

Zie bijv.Greenland , D.J. enR . Lal (eds.) 1977.Soi lconservatio n and management in thehumi d tropics.Joh n Wiley and Sons.

4.Geomorfologe ngaa ne rmeesta lva n uit,da t "stone lines" in tropische bodems het gevolg zijn vanerosi e en sedimentatie.Dez e hypothese kan leiden totverkeerd e conclusies over landschapsvorming en over paleoklimaten.

Zie bijv. Ab Saber,A.N. , 1982 in: Biological diversification in theTropics , p. 41-59 enRuhe , R.V., 1959.Soi l Sei. 87:223-231 .

5. Bijhe t opstellen van legenda's voor bodemkaarten,waarbi j moedermateriaal als indelingscriterium wordt gebruikt,houd t men vaak onvoldoende rekening met bijmenging vanaeolisch e sedimenten.

"Reconnaissance Soil Survey Reports"va n deKeny a Soil Survey.

6. Genese en eigenschappen van gronden met weinig actief aluminium maar toch veel röntgenamorfmateriaa l zijn onvoldoende onderzocht.

Wielemaker,W.G .e n T. Wakatsuki, 1984.Geoderm a 32:21-44 .

7.Vee l eigenschappen vankleideeltje s wordenmee r bepaald door de aard van het oppervlak dandoo r hun chemische samenstelling en struktuur.

BÜLÏOTHEEK

ÏANDB0UWHOGKSCHQOL WAGENINGEN 8.He tveelvuldi g voorkomenva n Planosolen opvulkanisch e as zoals o.a. be schrevenvoo r Chilie n Colombia,hang t samen met de grote mobiliteit van de klei.

Chili: ServicioAgricol a yGanadero , 1968.Estudi o agrologico delare a del embalse digua,7 2p . Colombia:Faivre ,P. , 1977.Scienc e du Sol 2:95-108 .

9. In zijnbesprekin g van de goede resultaten van"multipl e cropping systems" veronachtzaamt Sanchez tenonrecht e de invloed van debodemfaun a en de bodem flora.

Sanchez,P.A. , 1976.Propertie s and management of soils in the tropics.Joh n Wiley and Sons.

10. Gevleugelde termieten (alaten) zouden een belangrijker deelkunne n uit maken vand evoedin g van de bevolking ind e tropen dan thans het gevalis .

11. Tot de zorg voor onsmilie u behoort ook de handhaving van dewinke l op de hoek.

Proefschrift vanW.G . Wielemaker Soil formation bytermites ,a stud y in theKisi iarea ,Keny a Wageningen,2 4 oktober 1984 Abstract

Wielemaker, W.G., 1984.Soi l formation by termites,a study in theKisi i area, Kenya.Doctora l thesis,Departmen t of Soil Science and Geology, Agricultural University,Wageninge n (Xll)+ 132p. ,1 6 tables,3 9 figs, 119refs , 4 appen dices,Dutc h summary.

Mineralogicalan d chemical characteristics of samples froma number of soilswer e used todemonstrat e that soil materials fromvolcani cas han d local rock are thoroughly mixed. Themineralogy ,micromorpholog y and grain-size distributionwer e studied toestimat e the role of termites inmixin g soilmaterial s and in forming physical properties.Effect s of other processes onmixin g and on formation or decay ofphysical-propertie swer e also estimated. The role of termites informin g horizons free of gravelunderlai n by stone lines,an d in slope processes,wa s estimated from ananalysi s ofgrain - size distribution andmineralog y of soils ina catena.Measurement s of slope transport were also made. The significanceo f termites for soilconservatio n and fora sustained use of the land,i s discussed.

Keywords :termites ,mixing , soil formation,agro-ecosystems ,mineralogy , volcanic ash. Aan mijn ouders

Aan Ad, Martin, Erik en Sander Acknowledgements

I thank professorDr .P . Buringh,wh o offered me the opportunity to do my thesis work inKenya . I am greatly indebted toprofesso r Dr.J . Bennema,wh o encouraged and stimulated me inman y refreshing and friendly discussions,firs t in Kenya and afterwards inWageningen . In this context I should not forget tomentio n Mrs.Bennem a forhe r hospitality during theman y sessions whilew e discussed themanuscript . Thanks to the continuous support andhelpfu l suggestions of professorDr .L .va n derPlas , Iwa s able todevelo p the appropriate mineralogicalmethod s soessentia l for the theme of my work. TheWanjar e catenawoul d not have beens o thoroughly sampled and described if Dr.S- .Slage rha s not been there toassis tm ean d the students during fieldwork inKeny a and micromorphologicalwor k in Wageningen.

I thank theKenya nauthorities , particularly Mr.W.W .Wapakala , for the confidence entrusted in our University,whe n itapplie d for permission to start a field training programme inKenya .Tha t allowed me tokno w soman y fine people in this part of Africaan d to establish contacts withman y institutes and research organisations. Ihighl y appreciate,amon g others,th e long-lasting and friendly cooperationwit h theKeny a Soil Survey (Dr.N.N .Nyanda t andMr . F.N. Muchena), the University of Nairobi,Facult y of Agriculture (professor Dr. P.Ah n and professor Dr.S.O . Keya)an dwit h theNationa l Agricultural Research Stationa tKisi i (Mr.W . Nzabi). Ials o convey my thanks to theDepartmen t ofGeolog y andMines ,whic h offered its facilities to investigate certain chemicalan d mineralogical characteristics ofvolcani c ash. I am grateful to thefollowin g persons,wh o werewit hm e inKisii : - H.W. Boxernfo r creating thenecessar y facilities, forassistanc e with analyses and for much practicaladvice . - The students, (at present graduated)wh o Iguide d during research and who directly contributed to this thesis:P.N . Boerma,R.F . Breimer, Mrs. C.H.M.Duyker s van derLinden ,Mrs .J.M .Guikin g Lens,J . Hayma, MissM . Nachenius,H .va nReuier ,J.H.M .Scholte nan d P.A. van der Werff;als oE.M.G.C.A .Duykers ,Mis sE .va nLint ,K .Muilwij k andA.P . Oosterom,whos e work Indirectly contributed to thisthesis . - R.F.M. Onck and C.Kooyma nfo r theman y fruitful discussions about a subjectwhic h is our commoninterest ,namel y thesignificanc e of termites for soils and land-use. - D.va n der Eyk forhi s interestan d cooperation. - Mrs.E .O p deBee k andL .Wolk e for ahospitabl ehome . - Thefieldassistant sE .Maeta ,C . Faniako,T .Kananga ,M . Nyabonda, J. Nyambane, S.Nyareg a for their dedicationan d perseverance.

For datespresente d in this thesis and for preparation of the thesis I owemuc h topersonne l of the following Institutions in theNetherlands : Department of Soil Saienoe and Geology: - J.D.J,va nDoesbur g andA.J .Kuype r forX-ra y diffractometry and X-ray fluorescence work. - A. Engelsma formechanica lanalysi s andminera lseparations . - Dr.A.J .Having aan dR.M . van den Berg van Saparoea for palynology. - P.G.M.Verstee g for thedrawings . - Mrs.G .va nDinter-Deurin g and Mrs.J . Al.Sheikh-Haalboom for textprocessing. - Z.va nDruute n for photography. - Dr.P.M . Drlessen,P.J . Melitz,R .Miedem aan d Th.Pap e for helpful suggestions. The 'Stichting voor Isotopen Geologisch Onderzoek' at Amsterdam: - Dr.P.A.M .Andriesse n for fission track dating. International Soil Reference and Information Centre: - R. Bleiert forphotograph y of thinsections . The 'Stiahting Wereldvoedselvoorsiening': - H. Schouten for advice on modelling.

I thankMr .J . Dighton for checking theEnglish . Finally Ithan kAd ,m ywife ,an d my three sons that they tolerated the absent-mindedness of theirhusban d and father,whe nhi s thoughts lingered soofte nwit h termites. Itwa s greatho w you,Ad ,alway s found time to type thenumerou s drafts of thisthesis . CURRICULUM VITAE

De auteur werd geboren in 1942 teSilvolde . In 1960behaald e hij het diploma gymnasium- baa n het gemeentelijk Lyceum teDoetinchem , waarna hij aansluitend zijn studie aan deLandbouwhogeschoo l te Wageningen, begon. In 1968studeerd e hij af ind e regionale bodemkunde (tropische specialisatie)me t als keuzevakken de algemene bodemkunde en bemestingsleer en de cultuurtechniek. Van 1968 tot 1971wa s de auteur als assistentdeskundige bijd e FÀ0 werkzaam in bodemkarteringsprojekten t.b.v.irrigatie , in Chili.I n 1971 werd hij aangesteld alsmedewerke r tropische bodemkunde'bijd e vakgroep Bodemkunde en Geologie van de Landbouwhogeschool. Indi e funktie was hij o.a.werkzaa m inTurkij e (1971)e nKeni a (1973-1979),waa r hij leiding gaf aan bodemkundig onderzoek en onderwijs in projekten van deze vakgroep. InKeni a verrichtte hij tevens zijn promotieonderzoek, dat hij verder inWageninge n afrondde. Contents page

Abstract Acknowledgements CurriculumVlta e 1 INTRODUCTION 1 2 THEPHYSICA LAN DBIOLOGICA LENVIRONMEN T 5 2.1 Situation,altitud ean dpopulatio n 5 2.2 Geology 6 2.3 Geomorphology 8 2.4 Soils 9 2.5 Climatean dvegetatio n 10 2.5.1 Presentclimat e 10 2.5.2 Pastclimat ean dvegetatio n 11 2.6 Land-use 13 2.7 Soil meso and macrofauna in the area of study 15 3 SOILSSTUDIE D 17 3.1 General, i 17 3.2 Samplesite sfo rth edetectio no fsoi lmaterial sfro m volcanicas h 17 3.3 Profilesfo rth eeffect so ftermit eactivit yo ntranspor t andspatia larrangemen to fsoi lmaterial s 20 3.4 Profilefo rmeasuremen to fcree p 22 4 CHEMICALAN DMINERALOGICA LINDICATION SFO RMIXIN G 23 4.1 Introduction 23 4.2 Theanalysi so fmixin g 24 4.2.1 Mineralsa sindicator so fmixin g 24 4.2.2 Anestimat eo fsoi lmaterial sfro mvolcani cas hi n mixedprofile s 28 4.2.3 Anestimat eo fmixin gaccordin gt oth erati oo f ilmenitet ozirco n 31 4.2.4 Elementsa sindicator sfo rpresenc eo frelativel y youngvolcani cas h 33 4.2.5 Anestimat eo fmixin gaccordin gt oth emola rrati o ofA lt oF e 33 4.3 Discussionan dconclusion s 36 REQUIREMENTSO FTERMITE SAN DSOM ECHEMICA LEFFECT SO FTERMIT E ACTIVITYO NSOIL SSTUDIE D 39 5.1 Introduction 39 5.2 Therequirement so ftermite s 41 5.2.1 Food 41 5.2.2 Temperature 42 5.2.3 Moisture 44 5.2.4 Soildept hfo rnestin gan dbuildin gmaterial s 44 5.2.5 Risko ffloodin go rwaterloggin g 44 5.2.6 Risko fdisturbanc e 45 5.3 Effectso nchemica lpropertie s 45 5.3.1 Aliteratur erevie w 45 5.3.2 Resultsfro mth eare ao fstud y 47 MICROMORPHOLOGICAL STUDIESO FSPATIA LARRANGEMEN TO FSOI L MATERIALSB YTERMITE S 49 6.1 Introduction 49 6.2 Methods 50 6.3 Results 50 6.3.1 Termite-madepellet s 50 6.3.2 Voidsan dporosit y 52 6.3.3 Redistributionan dreorientatio no fskeleto ngrain s andplasm a 55 6.4 Discussionan dconclusion s 60 6.4.1 Featuresmad eb yth esoi lfaun aan dfeature s indicativeo fthei rdegradatio n 60 6.4.2 Pedoturbationan dminera ldistributio n 62 FORMATIONO FTEXTUR EPROFILE SI NRELATIO NT OVERTICA LAN D LATERALTRANSPOR T 66 7.1 Termitesan dtextur eprofile s 66 7.1.1 Introduction 66 7.1.2 Method s 67 7.1.3 Resultsan ddiscussio n 67 7.1.3.1 Textureprofile si nwel ldraine dsoil so f theWa nj ar ecaten a 67 7.1.3.2 Textureprofile si npoorl ydraine dsoil s ofth eMagen asurfac e 72 7.1.4 Conclusions 74 7.2 Lateraltranspor tprocesse s ....75 7.2.1 Introduction 75 7.2.2 Methods 76 7.2.3 Results 78 7.2.3.1 Slopetranspor tan dmedia ndiameter so f sandI nth euppe rWa nja r ecaten a 78 7.2.3.2 Slopetranspor tan dminera lcompositio n ofsan d 79 7.2.3.3 Downslopedisplacemen to fsoi lmaterial s fromth erock yuppe rpar to fth ecatena....8 1 7.2.3.4 Lateraltranspor to fsoi lmaterial si n deepwel ldraine dsoil s 82 7.2.4 Discussionan dconclusion s 84 7.3 Theformatio no ftextur eprofile swit hston eline s 86 7.3.1 Prevailingtheorie s 86 7.3.2 Othertheorie s 87 8 TERMITESAN DLAND-US E 89 8.1 Introduction 89 8.2 Termites-andnatura lecosystem s 90 8.2.1 Inan doutflo wo fenerg yan dmatte r 90 8.2.2 Suitabilityo fth ehabita tfo rtermite s 90 8.3 Termitesan dagro-ecosystem s 94 8.3.1 Effectso nsuppl yo fdea dorgani cmateria l 94 8.3.2 Effectso nplan tbiomas scove r 95 8.3.3 Effectso nsoi lcharacteristic s 95 8.3.4 Farmingan dtermit eactivit y 96 8.4 Significanceo ftermite sfo rhig han dlo winpu t agriculture 97 8.5 Researchneed s 99 SAMENVATTING 100 APPENDIX1 Profil edescription san danalytica ldat a 103 APPENDIX2 Method s 125 APPENDIX3 Plasmi cfabri ci nth eprofile sstudie d 127 APPENDIX4 Fissio ntrac kdat a 127 REFERENCES 128 1 Introduction

Thisrepor tdescribe sth eeffect so ftermit eactivit yo nsoil si na humi d tropicalare ai nsouthwester nKenya .Som eattentio nwa sgive nt omol erat san d ants,bu tthei roccurrenc ewa sles subiquitou stha ntha to ftermites .Onl yfe w earthwormswer eobserve di nth eare ao fstudy ,s othei reffec to nsoi lwa sno t studied.Microfaun aan dthei reffec to ncyclin go forgani cmatte rwer eno t evaluated. DespiteDarwin' s(1881 )earl ywor ko nth eeffect so fearthworm so nsoil s andlandscapes ,soi lfaun areceive donl yscant yattentio no fsoi lscientists . Thepage si njournal san dhandbook sdedicate dt osoi lfaun aan dit seffec to n soilar ebu tfe wi fcompare dwit hthos ededicate d tophysical ,chemica lan d mineralogicalpropertie so fsoi l(Hole ,1981) . Thisdoe sno tmea ntha tth eimportanc eo fsoi lfaun afo rth eformatio no f soilswa saltogethe rignored .I ntemperat eregions ,i twa sth eRussia nschoo l ofsoi lscience ,le d byDokuchae v (inRode ,1962) ,wh oalread yi nth epas t centuryemphasize d therol eo ffaun ai nsoi lgenesis .Dokuchae vstresse dth e concepto fsoi la sa biogeocoenos eo recosyste mo fwhic hfaun aform sa n integralpar tan dno tjus tit sinhabitants .I nthi sconcep tsoi li sa dynami c systemi nwhic hsoi lformatio nchange si fth ebioti cfacto ri saffected .Suc h viewssprea dthroug hEurop ean dcam eespeciall y toth efor ei nwor ko fGerma n (Kubiena,1948 ;Laatsch ,1957) ,Dutc h(Hoeksem aan dEdelman ,1960 ;Slager , 1966)an dothe rEuropea nsoi lscientist s(Russell ,1910 ;Lun tan dJacobson , 1944an dGhilarov ,1956) .

Fortropica lregion sthei rvie wi sunderline dwit hwor ko ntermite so f Drummond (1887),Dim o(1910 ,cite d byVolubuev ,1964) ,Kalshove n(1936 ,1941) , Pendleton(1942) ,Grass é(1950) ,Hess e(1953 ,1955) ,d eHeinzeli n(1955) ,Ny e (1955),Boye r(1955 ,1956 )an dman yother st ofollow .Studie so fothe rsoi l animalssuc ha sant s(Alvarad oe tal . 1981),worm s (Scrickandean dPathak , 1951;Wasaw oan dVisser ,1959 ;Scheltema ,1964 ;Madge ,1969 ;d eMeester ,1971 , Lee, 1983an dLavelle ,1983 )an dmole s(Abaturov ,1972 )ar eles snumerou stha n those on termites,whic h indicates that termites are generally more important in tropical regions than the other mentioned soilanimals .Th e importance of termites is further illustrated by thewor k of Zimmerman et al. (1982)wh o calculated thatarea s occupied by termites account for roughly two third of the earth's land area. Moreover, termites and particularly the fungus growing Maavotermttidae are among the fastest and most efficient decomposers of dead organic material innatur e (Wood and Sands, 1978). A large proportion of energy and nutrients locked up inorgani c matter,wil l thus be restored and used in the soil ecosystem. What isknow n about their effects on soilmaterial s and their spatial arrangement has been reviewed byLe e an Wood (1971a)an d will be discussed later in this report. If termites so thoroughly affect tropical soils it is surprising that many earth scientists (Ruhe, 1956an d Bigarella, 1964amon g others)di d not recognize their importance.Eve nMohr ,va n Barenan d vanSchuylenborg h (1972) in their "comprehensive"stud y on the genesis of tropical soils did not dis cuss theactio n of soilfauna , thus reducing soil toa lifeless entity.A s regards termites,soi l scientists commonly think of them aspest s forcrops , not being aware of their other role inecosystems ,no r realizing that only certain species are harmful for crops.Recentl y the termite's contribution to theworld' s annual production of carbon dioxide and methane was compared with that of other sources. Such studies,i f not speaking of the role of termites inecosystems ,easil y create a one-sided and therefore rather false imagery of their function inecosystems ,a s is clear from aquotatio n of the study by Zimmerman et al. (1982)i n Scientific American (1983)wher e termites were blamed for causing air pollution.

The author's interest in the effects of termite activity on soils was initially excited through,th e observation of the great number of termites in virtually every spade-full of soilan d through the observation of the following properties which seemed related to their presence: 1. The deep reddish clay soils in the area of study were so porous and stable thatn o surface run-off and virtually no erosionwer e observed even on intensively cultivated slopes of up to 20percent . 2. Soilmaterial s above the stone line were homogeneous with virtually no gravel andwit h very gradual horizon boundaries,whil e invie w of thewel l developed structure and the presence of argillans (Wielemaker and Boxem, 1982) -3- a clear horizon differentiation was to be expected. 3. Layers ofvolcani c ash,observe d inver y poorly drained soils,wer e not seen inadjacen t well drained soils. Could termites be responsible for the turbation of soilmaterials , the disappearance of layering and the creation of porous soils? Ifso ,wha t could they have contributed toweathering , slope transport and landform? The presence ofvolcani c ash with a chemistry and mineralogy so different from that of the older country rocks inKisi i (Wielemaker andWakatsuki , 1984) would certainly leave its fingerprints ifmixe d with soil from country rock; so it seemed an excellent situation to illustrate thepossibl e effects of termites on pedogenesis,i n theare a of study.

Would properties be equally affected by termites onal l types of soil in the area of study or would there be differences according to the soil's suitability as ahabita t for termites,o r due tovariatio n in land-use and vegetationand/o rkin d of soil materials? Theanswe r tosuc hquestion s seems important invie w of the present labour intensive type of agriculture practiced on thewel l drained soils.Th e impact of agricultural management practices, such as introduction of new crops,croppin g practice,fertilizers , pesticides and machinery,o n termite populations isno t known. If termites are beneficial for the productivity of soilswha t would thenhappe n to this productivity if termites were harmed so that their numbers declined?

Before the questions raised above are discussed, the reader finds in Chapter 2,a description of the present environment and its historical development.Change s in climate and vegetationwil l get particular attention, because they may have affected faunal activity in soils.Mos t important meso and macrofauna representatives in theare aunde r study are described here as well. Then,i n Chapter 3,a short description follows of the soils studied and why theywer e chosen. The actual results of studies are divided in two parts.Th e first part deals with mineralogical and chemical methods used inestimatin g degree of admixture of various soilmaterial s (Chapter 4).Thi s is background information for Chapters 5, 6an d 7,wher e the role played by termites in formation of soils,i s treated. Chapter 5describe s factors affecting their distribution and activity in soils and also some effects of their activity on

1. Mesofauna comprises animals thatar e 0.2-10m m long,macrofaun a comprises animals thatar e longer than 1c m (Wallwork, 1970). chemicalpropertie s of soil. Then,I n Chapter 6, follows anappraisa l of pedogenetlc processes responsible formixin g and of the role played by termites. Chapter 7describe s what effect termites have on formation of texture profiles and stone linesan dwha t thismean s for soilan d landscape formation. Data of studies discussed inChapter s 5,6 an d 7ar e synthesized in Chapter 8,wher e it isattempte d toconstruc t amode lexplainin g therol e of termites innatura l and agro-ecosystems and the factors that control their functioning and effectivity in those systems.Wha t termites possibly mean for a sustained use of the landan d how they interactwit hvariou s land-use practices is discussed aswel lan d some ideas are advanced onho w agriculture canprofi t from soilbiologica l processes in generalan d from termites in particular. 2 Thephysica lan d biological environment

2.1 SITUATION,ALTITUD E AND POPULATION

The region of study is situated just south of the equator in the south western part of Kenya and comprises an area,whic h extends from the Great Rift Valley in the east tonea r Lake Victoria in thewes t (Fig. 1).Th e area where most of the research was concentrated (Fig. 1)ha s analtitud e of 1300 to 2130 m a.s.l. (Fig.3) . The rolling tohill y landscape with generally deep soils is intensively cultivated by theKisi iwh o are Bantus (Ochieng, 197A). Luos,wh oar e Nilotes (Ogot, 1967), live in the lower,drie r andwester n part,whic h has a rolling topography. The latter part,wher e more soils are shallow and poorly drained than in thehighe r part,i s not so intensively cultivated (Wielemaker and Boxem, 1982).

Ï16, * LONGONOT'"*' \o

KENYA | & crater A # • township m-> Nairobi \ "\ i road "*-, r" border Situation 20 0 20 40k m N-x / y^-v river ///' of study area

Fig. 1.Ma p indicating location of themai n area of study 0_J), and of profiles and sites studied (•). -6-

2.2 GEOLOGY

Theare a is entirely composed of Precambrian intrusives and volcanic rocks,an d sediments.Thes e rocks are allmor e or less tectonically influenced and some show traces of a low temperature, low pressuremetamorphis m as for instance therhyolites . The geological studies of theare a (Huddleston, 1951)mentio n the occurrence of diorites,granites ,felsites ,basalts ,andesites ,rhyolites , conglomerates and quartzites (Figs.2 and3) .

STRATIGRAPHY ROCK TYPE STRATIGRAPHY ROCK TYPE Tertiary BJBBBBHI nephelinites, nepheline basalts Nyanzian t a Û d upper andesites Precwnbrlum Nyanzian [' !'_ j rhyolites Bukoban NSNgJ rhyolites and tuffs Nyanzian basalts Bukoban I» » *\ (eisites and andesites Intrutiv«« Bukoban ^^^^^J quartzites and cherts Post-Kaviron dia n doleriles and lamprophyres Bukoban }£££&fl basalts Post-Kavirondian F+*•,."*"+" ] Wanjaregranile s and{quartz)diorites Kavironctian ItaWVol congh omerates Post-Nyanzian C'X V,1 Kitere andOyugi s granites y *•* *» *-*i

Fig. 2. Simplified geological map of themai n area of study (Fig. 1) adapted from Huddleston (1951)wit h location of profilesan d sites (• ) , and cross-sectionA B (Fig.3) . -7-

KISII KISII KISII ' ' Magombo ' Magombo

Chepalungu

Magena

GEOLOGY quaternary deposits quartzites basalts mainly andesites I«* » I granites ;-_'

erosion surface surface remnants parent material deptho f classification AB-hor.(m) soil taxonomy

KISII gently undula a. East Kisii: 2-3 Paleudollan d ting ridge > 5mvolcani cas h Argiudol1 crests b. West Kisii: 2-6 Haplohumoxan d volcanicas h Paleudoll and local rock

MAGOMBO broad, gently mainly volcanicas h >8 Paleudollan d undulating Palehumult ridge crests

CHEPALUNGU broad, gently volcanicas h 1.5-3 Paleudollan d undulating and local rock Paleudult ridge crests

MAGENA plains Volcanicash , 1.0 Tropaqualf (dissected) local rock (abruptic) and coarse alluvium

Fig. 3. Cross-section AB (for location see Fig. 2) through main area of study with some characteristics of erosion surfaces, parent materials and soils.

Since the Tertiary and probably up to the middle of the last century (Wielemaker and Wakatsuki, 1984) volcanic ash from the Rift Valley was deposited in the area of study. It covered the region almost completely from the Rift Valley up to the eastern part of Kisii district. Farther west, ash layers were only observed in flat-bottomed valleys (Wielemaker and van Dijk, 1981), although studies on adjacent well drained soils showed that nearly all soilscontaine dmineral sfro mvolcani cas h(Wielemake ran dBoxem ,1982) .Th e volcanicashe swer epantelleriti ctrachyte sderive dfro mth eRif tValle y volcanoes(Wielemake ran dWakatsuki ,1984) .Thos etrachyte shav ea hig hmola r ratioo fF et oA l(2.3-2.65) ,a hig hmas sfractio no falkali-ion san dn ofre e silica(fo rchemica lcompositio nse eTabl e1) .

Table1 .Mas sfraction s(g/kg )o felement s(calculate da soxides )i n volcanicas hsample sfro mth eLongono t(VI6 )an dth eMenenga i(V26) .

+ SampleSi0 2 A1203 Fe203 FeO MnO MgO CaO Na20 K20 Ti02 P205 H20

V16 619 145 33 41 3 2 14 64 56 7 2 7 V26 597 145 45 44 4 6 17 51 51 11 1 23

2.3 GEOMORPHOLOGY

Thedat adiscusse dhereafte rar ebase do na stud yb yWielemake ran dva n Dijk(1981) . (a)Th eKisi iHighland sprope r(Fig .2 )a ta naltitud eo f150 0t o213 0m consisto fa rollin gan dhill ytopography .Th eare aform sa bloc ko fhig hlan d duet oth eresistan tquartzite san drhyoliti ctuffs ,whic hfor mridge srunnin g mainlyNE-S W(Figs .2 an d 3). Therathe rfla tsummit so fthes eridge sfor mth e remnantso fth eKisi io rhighes terosio nsurface ,whic hwa scorrelate dwit h theCretaceou speneplai n(Shackleton ,1951) .Summi tlevel so fth elowe rridge s (Fig.3 )for mremnant so fth eMagomb osurfac ewhic hwa sdate da searl y Tertiary.Th esurfac ei sno wassociate dwit hflat-bottome dvalleys . (b)Th eare asurroundin guni t(a )a ta naltitud eo f130 0t o150 0m ha s predominantlya gentl yundulatin gt orollin gtopography .Tw osurface sar eo f interest,calle dChepalung uan dMagena .Th efirs tprobabl yforme ddurin gth e laterTertiar yan dth elatte rdurin gth eMid-Pleistocene .Th efirs tconsist s ofridge swit hrathe rwid ecrest swithi na rollin gtopography ;th esecon di sa dissectedplain . -9-

2.4 SOILS

Thefollowin gdat aar ebase do npublication sb yWielemake ran dBoxe m (1982)an dWielemake ran dva nDij k (1981). Soilsar ewel ldraine dan dver y permeableo nsurface solde rtha nMagena .Soil so nth eMagen asurfac ean di n bottomlandsar epoorl ydraine dan dver yslowl ypermeable . Thedept ho fnon-gravell ysoi lmaterial si s2 m o rmor eo nth eChepalung u andolde rerosio nsurface s(Fig .3) .Th esoi lmantl ei sdeepes tove rth e Magombosurface ,wher ei treache sdepth so f1 0m an dmore .Becaus ethi s surfacei sth esecon doldes ti tindicate stha tsurface sexperience da differentdegre eo ferosio nsinc ethei rdevelopment .I fa ston eo rgrave l layeri spresen tthe ni ti susuall yfoun da tth etransitio nfro mnon-gravell y soilmaterial st osaprolite .Clay si nth ewel ldraine dsoil sar ekaolinitic , apartfro mth euppe rhorizo no fsoil so nth eMagomb osurfac ean dsoil smor e easto nth eKisi isurface ,whic hcontai nals osom e2: 1 typecla yminerals .Th e lastmentione dsoi lmaterial shav eles stha n6 0percen tclay ,whil eal lothe r soilshav emor etha n6 0percen tclay .Al lwel ldraine dsoil shav ea reddis ht o reddishbrow nB horizon . Inth eeaster npar to fKisi idistric tal lsoil shav ea molli cepipedon , whichi sver ydar kan d40-8 0c mdeep .The yals ohav ea nargilli chorizon , whichi sespeciall ydee pi nsoil so nth eMagomb osurface ,s otha tsoil sther e areclasse da s Paleudolls (SoilSurve yStaff ,1975) ,unles sbas esaturatio ni n theB horizo ni sles stha n35% ,whic hplace sthe mi nth e Paleudults. Volcanic ashi sth emai nparen tmaterial . Incontras twit hsoil si nth eeast ,soil si nth ewes to fth eare ao f studyo nth eKisi ian dChepalung usurfaces ,sho wa clea rrelatio nwit hth e countryrock .Intermediat et obasi cparen tmaterial s(felsit ean dbasalt ) carry Paleudolls and Palehumulte similart othos eo fth eMagomb osurface .Ove r quartzite,granit ean ddiorit emor eweathere dsoi lmaterial soccur , particularlyi nth ehighes tan dmos tleve lpositions .Thes esoil shav ea dar k coloredtopsoi lwit ha lo wbas esaturatio nan da noxi co rweakl ydevelope d argillichorizon ;the yar eclasse da s Haplohumox and Paleudults. Thesoil s overdiorit ei nslopin gposition shav ea highe rbas esaturatio nan da mor e developedargilli chorizon ;the yar eclasse da s Paleudolls. Soilso fPlain san dBottomland scarr yheav ytexture ddens esoil swit h siltybleache d toplayers (abruptic Tropaqualfs). TheBottomland sar efille d with4 m dee pdeposit so fquaternar yvolcani cas han dalluvium .Soil so fth e -10-

Plains overlie,a t 80-100c m depth,a weathere d rock. In the dense clay, rounded gravel and stones occur,usuall y at the transition to the weathered rock.

2.5 CLIMATEAN D VEGETATION

2.5.1 Present elimate

Rainfall ishighes t in theHighland s proper,whic h enjoy ahumi d high altitude tropical climate.Th eare a west of ita t lower elevation has a lower rainfall (Fig.4) .Figure s ofmonthl y rainfallan d evaporation (Fig. 5)o f

—• — provincial boundary • district boundary J700. isohyet with rainfall (mm) _>-U river

Fig. 4.Ma p of average annual rainfall of the study area. -ïi-

Morumba . Altitude 1920 n , Altitude 1768r » i. Altitude 1430m

ru 220-

1B0 n. — 140- .._ rrrl ----, 100- .,.! 60- u L__ 20-

F M A M J S O N D JFMAMJJASOND S O N D

Fig. 5.Averag e monthly rainfall (solid line), reached or exceeded in three out of four years and theaverag emonthl y potential évapotranspi ration (ET= 2/3 Eo,broke n line), for the stations Morumba,Kisi i and Asumbi (for location seeFig .4) .

stations situated inbot harea s (Fig. 4)sho wa dry season from December till March and a less pronounced dry season inJun e and July,particularl y in the lower lying area.Th emai n rainy season is fromMarc h till June. Annual rainfall fluctuates considerably fromabou t 1000 to 1800m m in the lower part near Asumbi and fromabou t 1500t o 2500m m in thehighe r part near Morumba. The range in temperatures for theKisi i area isa s follows (Braun, 1980):

altitude 1300 m 1600 m 2130 m meanmax . temperature 27.8°C 26.0°C 22.8<>C mean min. temperature 15.6°C 13.5°C 9.8°C mean temperature 21.8,,C 19.8°C 16.4°C absolute max. temperature 35.3"C 33.7"C 30.8°C absolute min. temperature 7.8"C 5.8"C 2.3"C

2.5.2 Past aliwate and. vegetation

During the Quaternary,climat e changed considerably inEas tAfrica . From around 15000- 12000B.P . the climate was so dry thatLak e Victoria was without an outlet,whil e lakewater s were alkaline (Kendall, 1969). Later than 12000B.P .Lak e Victoria overflew into the Nileagai nwit ha n interruption around 10000B.P .A wette r period occurred from 9000- 6500B.P . (Kendall, 1969). During thatperio d both Lake Naivasha and Lake Rudolf (Butzer et al. -12-

1972)ha da noutlet .Th esmalle r number offores t pollen after 6000 to300 0 B.P.compare d with theperio d before,coul d bedu et oeithe r anincrease d deforestation foragricultur e ort oa change inclimate . More generally theperio d from 60000t o 12500B.P .i sconsidere d drywit h more humid periods particularly around 40000-21000B.P .an da reallydr y period from21000-1200 0B.P .(Butze re tal . 1972;Livingstone ,1975 ;UNESCO , UNEP,FAO ,1978 ;Stree tan dGrove , 1976). Studies onth eRuwenzor imountain s learned that during themor e arid periods glaciers extended farbelo w their present level.A taroun d 14750B.P . they were founda slo wa s300 0m ,whic h is240 0m below their present level (Livingstone, 1962).Accordin g toOsmaston , cited byLivingston e (1975) temperatures were then4-6° Clowe ro nEas tAfrica nmountains .Thi sestimat e is based onth elowerin g of thefer n line,whic h extended to63 0m belowit s present level.Thi s loweringwa sno tonl y duet oa dro p intemperatur ea s Osmaston assumed,bu tals o toa greate r aridity,s otha t this estimate ofth e drop intemperatur e isprobabl y conservative.Th edro p intemperatur e might thenb ea smuc ha sth e6-7

A pollenanalysi swa scarrie d outb yv.d .Ber gva nSaparoe a in

1. R.M.va nd e Berg vanSaparoea .Departmen t of Soil Sciencean d Geology,Wageningen , theNetherlands . -13- 2 Cooperation withHamilto n ,wh o helped to determine pollen types ona profile situated ina poorly drained flat-bottomed valley near Profile S3(Fig .1) . The profile had a depth of 350c mdow n to solid rock and consisted of peat from 0-110 cm depth,o f peaty clay from 110-340c man d of loamy sand with peat from 340-350cm . C14 ageswer eapproximatel y 5800B.P .a t 340c m depth, 4510 B.P.a t 230c m depth,285 0B.P .a t 140c m depth and 1615B.P .a t 100c m depth (Wielemaker and vanDijk , 1981). Tentative results indicated an openwoodlan d vegetation since at least 4000B.P . This conclusion isbase d on thehig h percentage (50)o f pollen from Gramineae in the pollenspectra above a depth of 220c m and on the occurrence of scattered trees in the present day landscape. Below a depth of 220c m pollen froma Macarang a type,a treewhic h probably grew under poorly drained conditions,outnumbere d pollen of Gramineae (pollen of sedges and spores of fernswer e excluded from the sum of pollen,becaus e they probably originated from the direct surroundings of the site). Unidentified pollenwer e 20-30%o f thepollensum . The percentage of pollen from Gramineaewa s about 25nea r the surface of the profile. The small number of pollen of this type in the surface horizon, comparedwit h deeper layers could be due toa n increase in the intensity of grazing during thepas t millennium. Pollen from crops could not be identified apart frommaiz e of which pollen grains were observed in theuppe r 3c m of theprofile .

2.6 LAND-USE

Although agriculturewa s probably introduced inwester nKeny a as early as 250-300A.D. ,linke d with thearriva l of dimple based pottery (Soper, 1969 cited byMorriso n and Hamilton, 1974), i twa s only recently that itwa s greatly intensified. Several mzees (old men)livin g inKisii district recounted the situation that during their youth hillswer ewidel y covered with forest.Nowadays ,natura l forest virtually disappeared from thebette r drained areas, except on some rocky hilltops and along some very steep sided valleys (Wielemaker and Boxern,1982) . The intensification of agriculture only dated from the last three or four decades when the population doubled innumbe r

A.C. Hamilton. The NewUniversit y of Ulster, Coleraine Co. Londonderry,Norther n Ireland. BT5 2 ISA. -14-

Fig. 6. Intensively cultivated land inKisi i district at an elevation of almost 2000m .

(Ministry of Agriculture, 1975). Cropsgrow n can be subdivided into perennial crops such as coffee, tea, bananas and sugar-cane,whic h give a rather permanent cover to the soil,an d annual crops such asmaize ,beans ,potatoes ,pyrethru m and tobacco,whic h cover thesoi l temporarily (Fig. 6).Accordin g as theare a per farm declined, soils weremor e permanently cultivated. To raise production perha , fertilizers and pesticides were commonly applied, butus e ofheav y machinery wasmainl y restricted to thecultivatio n of sugar-cane (Wielemaker and Boxem, 1982). The grazing of cattlewa s traditionally important,bu t thenumbe r of cattle were declining,becaus e people needed more land for foodcrops. The flat-bottomed valleys and the Plainswit h imperfectly drained soils unsuitable for agriculture,wer euse d entirely for grazing. -15-

2.7 SOILMES OAN DMACROFAUN AI NTH EARE AO FSTUD Y

Termiteswer eth emos tnumerou san dwidesprea drepresentative so fsoi l mesoan dmacrofaun a(Chapte r 1).Ho wthe yliv ean dwha tdetermine sthei r activitywil lb etreate di nChapte r5 . Belowfollow ssom einformatio nregardin gth edistributio nan dactivit yo f someothe rsoi lmacrofaun agroups .The yrarel yaffecte dsoi lmaterial ss o thoroughlya stermite sdid ,s otha tthey'l lreceiv elittl eattentio ni nth e subsequentchapters . TheAfrica nmol era t Taehyovyotes sp.(Romer ,1966 )feed so nlivin g plantso fwhic hi teat smainl yth eroots .It ssiz ean dfor mar esimila rt oth e mole, Talpa euvopaea (Grzimek,1970) .I tals omake stunnel swhic hwer emainl y observedi nth euppe r3 0c mo fth esoi lprofil ean dles sfrequentl ydow nt oa deptho f12 0cm .Exces ssoi li sthrow na sheap so nth esurfac ewit ha nai rdr y weightpe rhea po f0. 5 to4 kg .Heap swer eobserve deverywher eexcep to nver y shallowan drock ysoil san di npoorl ydraine dsoils .O nwel ldraine ddee p soilsmol era tactivit yvarie dfro mon efiel dt oth enex tbu tals owit h season.The ywer eespeciall yfrequen ti nlan duse dfo rgrazing ,wher ethe y havea grea timpac to nslop etranspor t (seeSectio n7.2) . Ants,particularl y Myrmieavia eumenoides (Werff,va nder ,1978) ,o ri n theKisi ilanguag e Chimonyo Chimbavivi, built15-4 0c mhig hheap swit ha basa l diametero f60-10 0cm .The yfee do ndea dplan tremains ,dea dinsect san dals o onlivin gtermites .The ytolerate da wid erang ei ndrainag econditions ;onl y onver ypoorl ydraine dsoils ,i nwhic hals otopsoi lwa susuall ywaterlogged , noChimony oant swer eobserved .Man ynest soccurre do nth eimperfectl yt o poorlydraine dabruptl cTropaqualf so fvalle ybottom si nEas tKisii ,whic h wereunde rnatura lpasture .Thei ractivit yi nthes esoil swa scorrelate dwit h thefluctuatio ni ngroundwate rlevel .A risin gwatertabl econsiderabl y increasedthei ractivit ynea rth esurface ,probabl ydu et olac ko fnes tspace . Alsoth etemperatur eaffecte d theirdiggin gactivit yfro mn oactivit ya ta surfacetemperatur eo f6.5° Ct oa maximu mactivit ya t25° C_n da minimu magai n at35-40° C(Werff ,va nder ,1978) . Manyothe ran tspecie swer eobserved ,bu tthei rdiggin gaetivit ywa sles s conspicuouscompare dt oth eone sdiscussed . Aardvark Ovysteropus afev (Dorstan dDandelot ,1970 )fee dmainl yo n termites.Searchin gfo rtermites ,the yexcavat emound san dsubterranea nnests . Becausefungu schamber sar ecommonl y founddow nt oa dept ho f1 m ,thei rhole s -16- mayb ea sdee pa sthat .Th ehole so faardvar kwer eobserve despeciall yI n uncultivatedareas ;n orecord sar eavailabl eregardin gthei reffec to n microrelief. Dungbeetle s Saarabaeoidea (Brussaard,1983 )mak eball so fdun gwhic hi n theare ao fstud ywer eobserve da tdepth sdow nt o12 0cm ;thei roccurrenc ewa s linkedwit hcattl egrazing .Th ehole swer e 1c mwide .Th esoi lexcavate dfro m thehol ewa sdeposite do nth esoi lsurface .Th eoveral leffec to fth edun g beetleo nsoi ltranspor tan dmicrorelie fwa sapparentl ysmal lcompare dwit h thato fant san dmol erats . 3 Soils studied

3.1 GENERAL

Figure 1show s locations of sample and observation sites.Site s were chosen for different reasons;som e were selected for only one purpose,other s served several purposes (Table 2).Profil e descriptions and detailed analytical information are given in Appendix 1.Hereafte r follow brief descriptions of the studied profiles and sites,accordin g to the purpose for which they were selected. The claymineralogica l data given in this chapter are fromWielemake r and Boxem, (1982).

3.2 SAMPLE SITES FOR THE DETECTION OF SOILMATERIAL S FROMVOLCANI C ASH

Profiles Exc. 8,S3 ,21 0an dWa n 9o f group I inTabl e 2occup y nearly flat positions on ridge crests within the main area of study (Fig. 1)an d in a sequence at increasing distance from theRif t Valley fromwher e volcanic ash originated. Members of this sequence,situate d nearer to the Rift Valley, are described inWielemake r andWakatsuk i (1984). Profiles east ofExc .8 inclusive of Exc.8 ar e developed from volcanic ash only (Wielemaker and Boxem, 1982;Wielemake r and Wakatsuki, 1984), but profilesmor e to the west are from volcanic ash and local country rock (Table 2).Al l profiles are well drained.

Members of theWanjar e catena, used for detection of soilmaterial s from volcanic ash inmixe d profiles (indicated with 1an d 2i n Table 2)ar e described in Section 3.3. Stage of soil development of profiles on ridge crests is related with distance to the Rift Valley as this determined the amount of recent additions ofvolcani c ash (Wielemaker andWakatsuki , 1984). So Profiles Exc.8 and S3 have amolli c epipedon and an arglllic horizon,indicativ e of a relatively young stage of soil development,whil e Profiles 210an dWa n 9hav e an oxic horizon and an umbric epipedon,whic h are indicative of anadvance d stage of -18-

Table 2. Some data of the studied profiles.Grou p 1wa s selected for detection of soil materials derived from volcanic ash and an estimate of the amount in soil; Group 2, for an estimate of effects of termites on soilmateria l and soil transport and Group 3 for an estimate of the rate of creep, n.a. =no t applicable.

Profiles studied

V26 133/1 118/4 132/2 131/4 Exc.8 S3 210

Group 1111 1 1 1 1.2

Altitude(m) 2200 3000 3000 2700 2090 2145 1880 1833

Land-use grazing arable bambu wood pasture pyrethrum pasture pasture land/ forest land pasture

Erosion- surface Kisii Magombo Kisii

Slope % 1-2 0.5 0.5 2.0

Depth of AB- horizon (cm) 40 >150 >200 >200 >160 140 750+ 180

Depth of (cm) saprolite >400 >300 >300 >300 >300 >300 1000+ 10

Underlying volcanic volcanic volcanic volcanic volcanic volcanic andesite quartzite rock ash ash ash ash ash ash

A 3 3 4 3 3 3 Color B-hor. 10YR /4 - 5YR /4 5YR /4 5YR /6 2.5YR /4 5YR /4 2.5YR /6

Clay %B . 16.6 54.9 43.9 55.9 33 68 75.6

CEC of B-hor.in mmol/kg soil at pH 7.0 201 205 165 89 136 22

Base saturation in B-horizon 15 56 29 100 30

Drainage somewhat good good good good good good good excessive

Eplpedon ochric mollic umbric mollic mollic mollic mollic umbric

Subsurface horizon cambic cambic argillic argillic argillic oxic -19-

Wan 1 Wan 2 Wan 3 Wan 9 Wan 4 Wan 5 Wan 7 Opapo NARS

1,2 1,2 1,2 1,2 1,2 1,2 2 2 3

1647 1634 1567 1571 1528 1523 1516 1345 1695

shrub- shrub- bananas pasture maize malze pasture pasture pasture land land maize grazing grazing

Chepa- n.a. n.a. n.a. Magena n. lungu

12 0 10 10 12

20-35 135 345 310 320 275 90 85 300

30 400+ 40 35 200+ granite granite dlorlte diorite dlorlte diorlte diorlte granite basalt

3 4 4 4 5 3 2 3 2.5YR /6 2.5YR /5 2.5YR /6 2.5YR /4 7.5YR /2 10YR /1 10YR /2 2.5YR /&

20.1 63 61 66 36 55 73 n.d.

61 120 63 101 106 260 240/290 n.d

18 66 26+ 63 76 79 56 n.d. somewhat good good good good somewhat poor poor good excessive imperf. umbric umbric mollic umbric mollic umbric ochrlc umbric mollic

argillic argillic oxic argillic argillic argillic argillic argillic -20- soildevelopment .Profil e S3o n theMagomb o surface isver y deep,compare d with Profiles Exc. 8 and 210o n the older Kisii erosion surface; this anomaly was ascribed to a different degree of erosion experienced bybot h surfaces (Wielemaker and vanDijk , 1981), but it isals o likely that the poor weatherability of the quartzite rock played a role in the case of Profile 210.

3.3 PROFILES FORTH E EFFECTSO F TERMITE ACTIVITY ON TRANSPORT AND SPATIALARRANGEMEN T OF SOIL MATERIALS

Profiles selected,differe d indepth ,drainage , claymineralogy , fertility and slopepercentage . So theeffect s of termites on soil materials and transport of soilmaterial s could be studied in relation to these variables. Most research was concentrated on profiles of theWanjar e catena (Fig. 7). Catena isa term coined by Milne (1935). Hedefine s thisa s "agroupin g of soilswhich ,whil e they fallwid e apart ina natural system of classification onaccoun t of fundamental genetic and morphological differences,ar e vet linked in their occurrence by conditions of topography and are repeated in the same relationship to each other wherever the same conditions are met with".

Wan 1 y Wan la 1645 rNy Wan 2 1640 J! 2 3 4 Wan 3 1620 -*-$;!'^j&^ 8

Altitude (m)

T Wan 9 200 m

P.* ! granite ~-i™- contour line , _,. . altitude (ft) aa diorite

1500

Fig. 7. Cross-section of upper (a) and lower (b) Wanjare catena with situation of profiles ( f ) and deep augerings ( T) . For situation of map see Fig. 1. -21-

Upper profiles of that catena are shallow (Wan 1)t omoderatel y deep (Wan 2), which points at considerable slope transport,an d because the rock underlying these profiles isgranite ,downslop e soil transport will certainly leave its fingerprints in the mineralogy and particle size distribution of soils downslope,whic h overlie diorite.T o study the effects of slope transport versus in situ formation of profiles more thoroughly,dee p augerings were made (Fig. 7);the y were only sampled for determination of particle size distribution,san d mineralogy and profile depth. To study the relation between termite activity and the fertility levelo f soils,Profil e 210 (group I,Tabl e 2)wa s included,becaus e itha s the lowest fertility of all soils studied in thearea ,a s indicated by its low CECan d base saturation. Itals o has very few weatherable minerals and a kaolinitic claymineralog y just asProfile s Wan 9an dWa n 3.Wa n 3has ,however , amuc h higher base saturation and a finer and stronger structured A-horizon than Wan 9 and 210,whil e otherwise theonl y observable differences are an argillic horizon,a deep saprolite layer and a 12%slop e in the case of Wan 3,versu s noargilli c horizon,a shallow saprolite layeran d no slope in thecas e of Profiles Wan 9an d210 . Profiles Wan 4an dWa n 5a t the lower part of the catena (Fig. 7)ar e lessweathere d thanProfile s 210,Wa n 3an d Wan 9a s reflected by the high percentage ofweatherabl e minerals in the fine sand fraction (App.l), the presence of some 2:1 clays and thehig h CECan d base saturation. The lower profiles also have strongly developed argillic horizons and some signs of impeded drainage, butn o abrupt textural change.A nabrup t textural change is present in the poorly drained ProfileWa n 7whic h is representative of a strip of land in a small,flat-bottome d valley at the bottom of the catena. Profiles with a similarmorpholog y and drainage as Wan 7occup y extensive flat remnants of theMagen a erosion surface (Fig.3) . Several profiles froma cross-section through a termitemoun d on the Magena surface near Opapo (Fig. 1)wer e sampled and described tostud y the effects of termites on these soilswit h very slowly permeable and compact subsoils,consistin g mainly of 2:1 clays over aweathere d granite rock at80 - 100c mdept h (abruptic Tropaqualfs according toSoi l Taxonomy, Soil Survey Staff, 1975an d Planosols according toFAO/UNESCO , 1974,Appendi x1) . -22-

3.4 PROFILE FOR MEASUREMENT OF CREEP

The profile at theNationa lAgricultura l Research Station (Fig. 1an d Table 2)i s representative of many deep well drained reddish clayey soils. Also the long (400m )an d straight slopes of about 12%ar e representative of many slopes in the area of study. Classificationan d slopemorpholog y are similar toWa n 3,bu t land-use and type of country rock differ. The profile was used tomeasur e the rate of creep. 4 Chemicalan d mineralogical Indications for mixing

4.1 INTRODUCTION

This chapter treats chemicalan d mineralogical characteristics of soil materials studied, sofa r as they reflect theeffec t ofmixing . It provides background information for thesucceedin g chapters,wher e the impact of termiteactivit y on soil properties and processes will be discussed. Mixing of the soilmaterial s in the region under study was evident from the presence of glass fragments ofvolcani c ash in soilmaterial s derived from underlying rocks (Wielemaker and Boxern, 1982); because termites were abundantly present,the y were held responsible for the process ofmixing . Whether the termites were solely responsible for this process has still to be proved. Strong slope transport due to changes inclimate ,coul d have contributed to themixin g of soilmaterial s and may even be entirely responsible. If soilmateria l from various parentmaterial s ismixed , the chemistry and mineralogy of the mixed material may still reflect this. However, some of the characteristics may fade through theweatherin g of certain minerals or through leaching of certain elements. Such properties are particularly useful to show the presence of recently added soilo r parentmaterial ,bu t they can not beuse d toestimat e the proportion of added soil material.T o estimate this,mineral s resistant toweatherin g and elements little susceptible to leaching could be useful.Th e degree of mixing and theaccumulatio n of soil fromvariou s parentmaterial s over a long period of time could then be estimated,unde r theassumptio n that these minerals and elementsar e typical of one of the constituent parentmaterials . Inaddition , the ratios of minerals and of chemical elements typical of a certain type of parent material in composite soil samples could also be used toanalys e the distribution and amount of various parent materials in such mixed soils. This approach seemed very promising for an estimate of the degree of -24- mixing because theperalkalin e volcanic ash from the RiftValle y has a composition,whic h isver y different from that of the local rocks in the area under study (Section 2.2 and Wielemaker and Wakatsuki, 1984). Inadditio n to mixing by termites, theeffec t of slope transport could beestimate d in profiles of theWanjar e catena (Fig. 7).Th euppe r part of this catena is over granite and the lower part over diorite. If lateral transport was important then the soils of the lower partwil l reflect this in their mineral composition.

4.2 THEANALYSI SO F MIXING

Before describing themethod s used for theanalysi s ofvariou s aspects of mixing, thisparagrap h gives first an inventory of minerals typical of each of theconstituen t soilmaterial s inmixe d profiles.Method s for theanalysi s of soils and rocks are given inAppendi x 2.

4.2.1 Minerals as indicators of mixing

Minerals typical ofvolcani c ash.Th e chemical composition of the peralkaline ash from the Rift Valley is given in Table 1an d the mineral composition ofver y fine sandwit h a density of >2.89 of soils formed in this type ofvolcani c ash is given in Wielemaker and Wakatsuki (1984). The composition of themagneti c very fine sand fractionswit h a density of >2.89 is given in Table 3an d the composition of thenon-magneti c fraction with a density of >2.89 isgive n inAppendi x 1. Volcanic glass,sanidine ,apatite ,aegirin e augite,barkevikit e and a certainmagnetit e species (M-II)ar e typical for this type ofvolcani c ash and because they dono t occur in the local rocks of theare a under study they could serve aspossibl e indicators ofmixing .

Themas s fraction of volcapic glass in very fine sand of fresh volcanic ash isabou t 80 percent.Glas s hasmainl y a density of <2.5,bu t a small fractionha s a density between 2.5 and 2.89. The last fraction weathers very easily as follows from itsabsenc e in samples ofmor eweathere d volcanic ash in soils farther from the Rift Valley (Wielemaker and Wakatsuki, 1984). Of the transparent crystallineminerals ,sanidin e ismos t important. It -25-

•4-1 O « ai ti,,, >> 4J -. o + I + + (0 Ei V4 C M M 01 « O I 13 h h h ii h |4 I"! tU ao 4-1 U 4-1 4J 4J U U + Il D} U 4J !-! CO O 1-1 B ' a to co *•> o -o S M Ci C M C , oo cd > td m a s (U ai S! # * M n l + + + + + CO , i u •H /-S .-H M-l 8 • a • M >, C -H V4 co co 00 (U o H > O O g iiiïïi •H S*. I a_i • ""> o •ri «ai H H H o C _ i • îi s-^ 00 ro »o ai to > 01 co g c- - • i *J 01 + cd CM (3 M-l i-i r~ r~ oo r>. -cr oo rs CM O -H o CU fl H iO H N 0\ CO vO —I A CBi C3 -* O O O O -H 00 r-4 co s^r "a M • ••••• 4J C J3 o. • cd •H (0 4-1 o, o o o o o o CO o > co ai »o 't rs co »o rs CO o r- •o I cd cd co N CN N N H 1Û >3- CO ai S o e co cu Cl Ps N \û CO »O -^ m 'V • u CO • vy 3 »H • ••••• ó o ^^ - + A m 4-1 to 01 tu H NN M HH t-H co m co co J3 cd CO 05 -TJ O ~H -a- o •SA OoOOOO • (3 • • O) 01 -• C a -H ..t d "i M-l S K o * - - CO I •Htwco-^ oooooo O TH O O o u 0) .. - (3 cu CO ^-\ i-I r-l + M •H U o c a 4-1 •« O (3 >s O • 4J td _ ai •H CO CO U B «HA 4-1 w 2°: * co o un IH a m \o m CM co o ü co o f-i •H U -H CO Ji -H -* oo c^ CM \0 td o eu Q M-l 05 > ^ UTIH 1^8 tu 4-1 ai -- a o o o o vO « e o 4J -- o 4-1 /-» co m m »n fi vo C0 00 >J iH -- tl CM I I I I | I td CO CX 4-1 -f- CH P. a I O O O O CM O 01 u O -3- -Cf -* -3- CM (3 B •• CO Q w • >t -H 01 CS .M-I cd —I -* CM -Cf 00 •H -H O O o ^^~^ —. • U 0 J3 u CM a Cd U 1 *0 CO 00 CM H ü O CO td 4-1 H CU - '- ^ CM CO —< CO CO X •H M H 10 A H Wc C o co > -< -" -I ^ W •O 60 o > cd -26- ismor e resistant toweatherin g thanvolcani c glass,whic h caused an increase in its mass fraction with distance from the source (Wielemakeran d Wakatsuki, 1984). Aegirlne augite,a green to bluish-green transparentweatherabl emineral , isno t so common inmixe d profiles as volcanic glassan d sanidine. Barkevikite,whic h isa ver y dark brown almost opaque hornblende, is observed inmixe d profiles,bu t sometimes confused with opaqueminerals . Apatite is only common in the samples of fresh to little weathered volcanic ash (V26 in Fig. 1).Smithso n (1941)an d Pettijohn (1941)cite d by Brewer (1976)conside r apatite a stablemineral .Th e absence of apatite in the soils from volcanic ash indicates,however , that itweathere d very easily. Magnetite-IIwit h d-values of 2.56, 1.50 and 2.99 (powder diffraction file of JCPDScar d No.10-319 )prove d typical of volcanic ash. Iti s the commonestmagnetit e species inX-ra y diffraction patterns from this type of volcanic ash (Table 3).Th e less common species in theas h (magnetite-I)wit h d-values of 2.53, 1.49 and 2.97 is also the species occurring in the strongly magnetic fraction from local rocks in theare a of study (Table 3).Magnetite - IIha s d-values similar to those of jacobsite.Th eMn-conten t ofmagnetite-I I is,however , too low for jacobsite and does not differ significantly from that ofmagnetite-I .Th e only difference is ahighe r content ofTi in the case of magnetite-II. Substitution of Ti for Fe in the crystal lattice of magnetite-II canhav e caused d-values intermediate betweenulvospine l and magnetite-I and similar to those found for jacobsite.

Minevals not typieal of volaanio ash. Thefollowin gmineral swere ,i nsom e situations, typical for one of the constituent soilmaterials . Opal phytolithwit h a density of <2.5 is asecondar y amorphous mineral consisting of silica. Its form isusuall y tubular and itha s thestructur e of plant cells (Fig. 8).Th e distribution of opal phytolith isassociate d with the distribution ofvolcani c glass inwel l drained profiles. Formation and decay of opalphytolit h are probably related to the concentration of silica present in the soil solution,whic h may explain its association withglass , because glass is probably themos t important source of silica and so determines its concentration in thesoi l solution. Epidotewa s rarely observed in profiles fromvolcani c ash because it is a metamorphic mineral. Itis ,however ,commo n inprofil e S3,whic h at first was thought tob ederive d fromvolcani c ash only.Becaus e theunderlyin g parent Fig. 8.Electro n micrograph (A)showin g some forms of opal phytolith. Themos t common type (B)seem s to be derived from grass (personal communication of J. Hayma). material of profile S3 contains a relatively high mass fraction of epidote, its presence in the solum proves that Profile S3 isa mixe d profile. All country rocks in the area of study and also volcanic ash contain zirconswit h awel l developed typical crystal morphology. Fission track dating showed that zircons from volcanic ashwit h anag e of 700 to90 0 thousand years (App. 4)stil lhav e their original form. Inprofile s of the Wanjare catena developed from granite and diorite, zirconmorpholog y ranges fromwel l formed idiomorphic crystals to zircons with corroded edges and rounded zircons.I n the latter profiles itwa s nearly impossible todistinguis h zircons from volcanic ash,fro m zircons derived from the country rock;moreove r zircons in ash are rare in comparisonwit h zircons from the granitic or dioritic rock.

The situation in Profile S3 is different,wit h soilmainl y derived from volcanic ash.Thi s profile is underlain by a parentmateria lwhic h besides many epidote minerals contains few zircons,whic h were all rounded. The -28- rounded zircons contained an uncountable number of fissions,whic h indicates that they are very much older than thewel l formed crystals fromvolcani c ash. The presence of epidotean d also the presence of rounded zircons throughout the profile indicates that Profile S3 (Fig. 1),whic h seemed to be derived from volcanic ash only,i s in fact amixe d profile.

4.2.2 An estimate of soil materials from volaania ash in mixed profiles

Amount of volcanic glass and ofmagnetite-I Ii nmixe d profiles are used as indicators for the distribution of relatively young and relatively old soil materials fromvolcani c ash inmixe d profiles.The y are among themos t common minerals in this type ofvolcani c ash and have the advantage of being easy to determine. Volcanic glassweather s easily, so that its presence represents relatively young soilmateria l from volcanic ash.Magnetite ,althoug h much more resistant toweatherin g than glass,particularl y in the oxidizing environment of thewel l drained profiles under study, is,however ,no t considered by all authors tob ea ver y stablemineral .Pettijoh n (1941) considers it less stable than zircon,ilmenit e and apatite and Weyl, 1952 (cited by Brewer, 1976)place s it in the stability group of zircon,rutile , tourmaline and sphene. Ifmagnetite-I I isver y resistant toweatherin g itsmas s fraction in mixed profiles could serve as anestimat e of the soil materials derived from ash.T o judge its stability,mas s fractions ofmagnetite-I I inTabl e 4wer e compared with mass fractions of ilmenite and zircon. Itappear s that fractions of ilmenite and zircon increase in samples farther from theRif t Valley,whil e the fraction of magnetite-II remains constant or decreases slightly. Infact , themas s ratio ofmagnetite-I I to ilmenite decreases from 10o rmor e in samples near the RiftValle y to0.3-0. 6i n samples far from the Rift Valley such as in samples from Profile Exc. 8. It seems,however , unlikely that the difference indensit y or the difference in form of the twomineral s could have caused the drastic decrease in the fraction ofmagnetite-I I compared with ilmenitean d zircon,wit h the exception of the possible effect of winnowing described later.Becaus e the ratio ofmagnetite-I I to ilmenite is rather similar inol d and young deposits of little weathered volcanic ash, the decrease in the content ofmagnetite-I I compared with ilmenitean d zircon seems due to its lower resistance to weathering. -29-

Table 4.Mas s fractions inver y fine sand (50-105 pm) x 10 ofmagnetite-I I= Mil, ilmenite = Ilm, zircon = Zan d weatherable minerals (amphibole, pyroxene) = W,estimate d from X-ray fluorescence analysis (Table 3)an d from the number of transparent minerals in thenon-magneti c fraction (Appendix 1)o f profiles onvolcani c ash (Vole.P.)an d inmixe d profiles (Mixed P.).Fo r sample depth and distance from the Rift Valley, see Table 3.

Vole.P. Mil Ilm Z W Mil/ Ilm Ilm/Z

V26 7.1 -7.9 0.3- 0.8 0.00 6.1 9 -23 - 133/1 1.6 -3.2 0.3- 0.8 0.04 13.7-15.2 2 - 5 7 - 19 118/4 1.6 -3.5 1.0- 2.9 . 9.2-15.3 0.6 - 1.6 . 132/2 3.3 -7.5 1.2- 3.7 0.09 10.2-15.3 0.9 - 2.8 12 - 37 131/4 1.0 -1.6 4.0- 10.4 0.28 7.8-11.8 0.1 - 0.3 14 - 36 Exc.8 1.2 -1.8 3.1- 6.3 0.21 3.1- 6.1 0.2 - 0.4 15 - 30 Mixed P. S 3 0.5 -1.0 8.7- 9.7 0.08 0.5- 1.0 0.05 -0.06 112 - 125 Wan 3 0.19-0.58 90.4-102.5 36-46 0 0.002 1.9- 2.9

A reliable estimate of the amount of volcanic ash inmixe d profiles according to itsmagnetite-I I contentwa s further hampered by two other effects: (i)Winnowing . Themas s fraction ofmagnetite-I I to soilmaterial s derived from volcanic ash should first beestimate d in soils from pure volcanic ash. Such profiles are,however ,a t aminimu m distance of 40k m (Profile Exc. 8 in Fig. 1)fro m themixe d profiles of the Wanjare catena (Fig.7) .Thi s distance would certainly affect the ratio because,whe n distance to the source of ash increases, the fraction of heavymineral s in soils will decrease (winnowing effect). It follows then that the ratio inProfil eExc . 8 is not representative of soil materials from volcanic ash inmixe d profiles of the Wanjare catena. (ii) Stage ofweatherin g of soilmaterials .Thi swoul d affect the ratio of magnetite-II to total soil materials from volcanic ash inmixe d profiles.T o avoid this effect, the ratio ofmagnetite-I I to totalA l in soil materials from volcanic ash was used,becaus e in the soils studied Alseeme d not susceptible to leaching and therefore independent of stage of weathering. These effects prevented an exact estimate of the fraction of Al derived fromvolcani c ash inmixe d profiles.Therefor e only an estimate of the relative distribution of that fraction inmixe d profiles was possible. -30- Distribution of volaania glass in mixed profiles. Indee pprofile so fth e Wanjare catena (forsituatio n seeFig .7) ,glas s isespeciall y observed inth e upper 100cm ,wher e itsmas s fraction decreases from thesurfac e down toabou t 100cm .Belo w that depth theamoun t ofglas s islow ,compare d with theuppe r part,bu tremain s nearly constant down toth egrave l linea tdepth s of30 0t o 350c m (Profiles Wan3 an dWa n9 i nTabl e 5)o rth ebotto m of thepi t (ProfilesWa n4 an dWa n5 i nTabl e 5).Belo w thegrave l line some glass fragments arestil l detected inProfil eWa n3 i nth epar twit h soil structure down toth ebotto m ofth epi ta ta dept h of 7m (Table5) .

Distribution of magnetite-II and Al from voloanie ash in mixed profiles. Fig. 9 shows thetren d inmas s fractions ofA lfro mvolcani c ashi nProfil e Wan3 according toit smagnetite-I Icontent ,assumin g amas s fraction ofA l fromas h in topsoilo f0.6 3 (thisvalu e isbase d onth erati o ofmagnetite-I Ii nver y fine sand totota lAl ,i nsample s offin e earth from Profile S3.Th e calculation procedure isexplaine d inApp .2) .Mas s fractions inth euppe r 160 cmar eslightl y higher thani nth elowe r part of this profile. Below 330c m depth,n omagnetite-I Iwa s detected.

Table5 .Numbe ro fvolcani cglas s fragmentspe r10 0grain s(Appendi x 1)i nth ever y finesan d fraction (50-105(im ) recalculateda smas sfraction so ffin eeart hi n profileso fth eWanjar ecaten a (x1 0 ).

Profile Wanl Profile Wan3 Profile Wan9 Profile Wan4 Profile Wan5 Depth glass Depth glass Depth glass Depth glass Depth glass (cm) (cm) (cm) (cm) (cm)

0-3 5 0.42 10-2 5 0.8 0-1 8 0.35 20-3 6 0.025 8-2 3 0.27 40-5 5 0.026 18-3 8 0.33 50-6 5 0.022 38-5 3 0.07 70-8 5 0.059 38-8 0 0.21 95-110 0.013 70-8 5 0.04 Profile Wan2 85-100 0.23 80-130 0.025 170-185 traces 110-125 0.016 Depth glass 145-160 0 130-210 0.035 211-226 traces 166-182 0.013 (cm) 190-205 tr. 210-286 0.008 260-275 traces 198-214 0.017 0-1 6 0.13 240-255 0 287-300 0.05 320-335 - 280-295 0.019 5-3 0 0.04 255-270 0.074 55-165 traces 315 traces

750 -31-

mass fraction of Al

0 25 0.50 0.75 1. n u AI • A3 •

BI •

100- • BZIt

• B22t

200- •

- • B23t •

300- Fig.9 .A lfro mvolcani c ash (• ) a s mass B24t • fraction of totalA l in samples of profile - -gravel 1 ine Wan 3.Amoun t of Al from volcanic ash is B3 estimated from the content ofmagnetite-I I 700nn-Jl in thever y sand fraction,a s explained Depth(cm ) inAppendi x 2.

4.2.3 An estimate of mixing according to the ratio of ilmenite to zircon

Themas s ratio of ilmenite and zircon in samples ofver y fine sand of soils from volcanic ash ranges from 12-40,whic h is muchhighe r than in samples fromsoil s overlying theWanjar e dioritewhic hha s ratios lower than 3 (Table 4).Becaus e ratios of ilmenite to zirconar e so characteristic for each parentmaterial , trends of ratios inmixe d profilesma y give information on the degree of mixing of soil materials from volcanic ash and the rock underlying the profiles of theWanjar e catena. Because ilmenite and zirconhav e a similar density of about 4.7 and also a similar surface area, the ratiowil l not changewit h distance to theRif t Valley due towinnowing . Furthermore,weatherin g seems tohav e little effect on the ratio as a result of the similarity in ratiovalue s in older and younger soils from volcanic ash.Th erati o could not be determined for soils nearest to theRif t Valley because no zirconswer e detected in the heavy fraction. Its value must be lower than that of Profile 133/1 because the ilmenite content of both samples is similar,whil e the zircon content of the -32- sample from ProfileV2 6 Isapparentl y smaller than that of Profile 133/1.

Fig. 10show s the trend in ratios of ilmenitean d zircon inver y fine sand of Profile Wan 3.A rati o respresentative for soilmaterial s from diorite was taken tob e 2.1.Thi s ratiowa s found in the partwit hweathere d rock just below the solum,wher e stage of weathering may differ least from soil materials above it (To compare ratios from different parts of theprofile , stage ofweatherin g should differ as littlea s possible,becaus e itma y affect distribution ofmineral s over different size classes (Brewer, 1976)). Addition ofvolcani c ash withhig h ratios of ilmenite to zircon would raise the ratio.Thi s is indeed observed, but in contrast with the distribution of glass and magnetite-II,ratio s of ilmenite to zircon would indicate thatmor evolcani c ashwa s added to thepar t below 150c m depth, than to thepar t above it.Thi sanomal y might be due to slope transport because soilmaterial s inprofile s upslope had low ratios of 0.3 or less, so that addition of those materials to theuppe r part of Profile Wan 3coul d have caused the decrease in ratios.Mor e evidence for this hypothesis will be given

mass ratio of ilmenite to zircon 12 3 4 0 A1

100 I _ B21t

B22t

200

B23t

300 Fig. 10. Mass ratios of ilmenite B24t gravel line to zircon in the 50-100 pm frac tion of samples from Profile Wan 3, B3 estimated from X-ray fluorescence data (App. 1) and from countings of zircons in the non-magnetic 400 J fraction (App. 1). Depth (cm) -33-

in Sections 5.4 and 6.3,wher e itwil l be shown that slope transport affected themineralog y of theuppe r part of Profile Wan 3,particularl y itsuppe r 30 cms.

4.2.4 Zinc and potassium oxide as indicators for présence of relatively young volcanic ash

Themas s fraction of zincwa s (0.2-0.3).10~ 3 and that of potassium oxide was 0.5-0.6i n the little weathered samples ofvolcani cas h from the Rift Valley.Thos emas s fractions are very high formos t soilmaterials .Zin c and potassium are susceptible to leaching, so thatmas s fractions of zinc and potassium oxide are a good indication for theamoun t and distribution of relatively little weathered soilmaterial s fromvolcani c ash. Fig. 11 shows how Zn-content of topsoil decreases from east towes t in the Kisiiarea . The figure also shows that Zn-content of topsoil is always higher than that of subsoil,excep t in themos twesterl y soils,wher e the lowest Znconcentration s occur. Concentration of Zn isapparentl y linked with mass fractions of young soilmaterial s fromvolcani c ash in soils,resultin g from thedecreas e inconcentratio n with distance from source and with depth. The trend in theconten t of potassium oxide insoil s from east towes t Kenya is similar to the trend inZn-contents .A compariso n of the deep soil Profile S3 from EastKisi iwit h thedee p Profiles Wan 3an dWa n 9 fromWes t

Kisii (Fig. 1)show s that Profile S3ha s a mass fraction of 0.017 K20i n topsoil decreasing gradually to0.00 8a t 7m depth (App. 1).I n contrast,Wa n 3 and Wan 9hav e mass fractions in theuppe r 3m , ranging from0.00 5t o 0.009. Because S3 isa deep soil, the high content of potassium oxide in the topsoil of Profile S3ca n beascribe d to the presence of volcanic ash.

4.2.5 An estimate of mixing according to the molar ratio of Al to Fe

Toestimat e mass fractions of soilmaterial s from volcanic ash inmixe d profiles, themola r ratio of Al to Fe (III)seeme d promising. Both elements are nearly immobile underwel l drained conditions,s o that the ratio will hardly beaffecte d by leaching. There were,moreover ,n o indications that Fe and Al cheluviated with organicmatter . The ratio inwel l drained profiles fromvolcani c ash ranges from 2.3-2.5 -34-

Menengai

f Nakuru

Longonot

Suswa

LoitaPlain s

(Wielemaker andWakatsuki , 1984 and Table 6,Profil e Exc. 8),whic h is low in comparison with a ratio of 3.46 in diorite and 4.3 in granite measured in rocks, underlying profiles of theWanjar e catena. Despite the fact that Profile S3prove d to be a mixed profile, its soil materials in thewel l drained part seemed mainly derived from volcanic ash; this followed from the ratios of Al to Fe,whic h were within the range of soil materials derived from volcanic ash. If itwer e possible tokno w the ratio ofA l toF e in soilmaterial s from diorite or granite present in themixe d profiles,the n the mass fractions of soil derived from volcanic ash and local rock could be reliably estimated. This is however not possible,becaus e ratios in the zonewit h saprolite and rock are affected by alternating reduction and oxidation of iron (see e.g. -35-

Table 6.Mola r ratios ofA l toF e (III)i n fine earth of profiles Exc. 8 and S 3 (for situation see Fig.1) .

Depth Exc. 8 Depth S 3 (cm) Ratio (cm) Ratio

0- 22 2.33 0- 50 2.29 22- 66 2.35 79-270 2.56 90-105 2.41 340-345 2.32 150-165 2.23 740-750 2.02 170-200 2.12 200-250 2.03 C-hor. 2.86

Profile Exc.8 a t a depth of >150c m inTabl e 6).Suc h conditions lowered the ratio in theuppe r part of the reduced zonean d raised it in the lower part, where reduced conditions weremor e permanent; this had occurred in the weathered rock sample of Profile Wan 2wit h a ratio of 9.9, compared with 4.3 in thehar d rock.Latera l transport of iron could causea decrease of the iron content particularly in freely drained profiles on ridge crests so that an increase, rather thana decrease,o f the ratio is to be expected. Due to reasons mentioned above a reliable estimate of themola r ratio of Al toF e for soilmaterial s derived from local rocks isno t possible. The trend in ratios in thewel l drained part of the profile may,however , still reflect how soilmaterial s fromvolcani c ash and local rocks are mixed, assuming that soil materials derived from the local rocks in thewel l drained part of theprofil e have a uniform ratio due tohomogenization .

Fig. 12show s a decreasing fraction of Alderive d fromas h with increasing depth of soil down toabou t 300-350cm ,wher e the zone of alternating reductionan d oxidation starts.Th e trend inratio s of Profiles Wan 3an d Wan 9 isver y similar,althoug h ProfileWa n 9 is situated ona level ridge top and Profile Wan 3o n a slope. So it seems that slope transport did not cause a situation leading to serious overestimatlon of the fraction of Al derived from ash in Profile Wan 3,whic h could have occurred if an important fraction of soilmaterial s from upslopewit h high ratios of Al toF e (see Profile Wan 2i n App. 1)ha d beenadded . Effects of slope transport will be further discussed inSectio n 7.2. -36-

Depth (cm) fractiono f Al 0 0.5 1.0 • o

o • • o

o 100 • •

° •

0 200-

B2t o • o

Fig. 12. Fractions ofA l from volcanic ash in 300 • fine earth of profile Wan 3 (o)an d Wan 9 (• ) , line red o gravel calculated withmola r ratios ofA l toF e (III) oxyd of 2.3 for volcanic ash and of 3.5 for diorite. Mass percentages of oxides ofA l and Fe in fine B3 earth of profiles Wan 3an d Wan 9ar e given in o Appendix 1.

4.3 DISCUSSION AND CONCLUSIONS (TABLE 7)

Concentration of zinc and potassium oxidewa s useful todetec t addition or admixture of relatively recent volcanic ash,especiall y indee p weathered soils. Chemical elements were,however ,no t specific for a particular parent material,s o that detection of a smallmas s fraction of allochthonous material was not possible.Fo r this purpose,volcani c glass fragments proved very useful. Their numbers decrease from the surface down toabou t 1m and then remainnearl y constant down to thegrave l line.Belo w this line,glas s fragments are present in the part with soil structure down to the bottom of the profiles investigated (7m inProfil eWa n 3an d 10m inProfil eS3) . According to the small fraction of glass at a depth ofmor e than 1m , only a small but consistent fraction of relatively young soilmaterial s from volcanic ash is present.Method s indicating the total fraction of soil materials derived from volcanic ash inmixe d profiles,showe d that those fractions did not differ much with depth above the gravel line. X-ray fluorescence analysis could not detectver y small fractions ofmagnetite-II , -37-

Table 7.Summar y ofmethod s used in estimating presence of soilmaterial s from various sources in soilsan d their applicability,xx ,reliabl eresults ;x (x) , reliable If ratios differ significantly and If enough minerals are present;x , reliable only if content ishig h enough,n o reliable quantitative estimate possible;8 ,onl y insom e profiles applicable;- ,no tapplicable .

Method Presence of soil materials Semi-quantitative Difficulties from volcanic ash estimate young old country rock

Zinc X - - - leaching potassium X - - - leaching volcanic glass XX - - - weathering aegirine augite XX X not common, demanding method zircon - - a - demanding epidote - - a - weathering magnetite-II/I X XX X X winnowing and weathering ilmenite/zircon - X x (x) X size frequency Al/Fe X XX X X leaching or accumulation due to reduction and oxidation of Fe indee p subsoil

that may have occurred in the part below the gravel line or in relatively young profiles such asWa n 4an dWa n 5.Neithe r did thismetho d work in samples from profiles derived from granite,havin g a large fraction of normal magnetite (Profile Wan2) . The ratio of ilmenite to zircon shows that theuppe r 1.5m ofProfil e Wan 3 was derived from three parent materials:diorite ,granit ean d volcanic ash. Soilmateria l from granitewa s apparently added to this part of the profile by slope transport. Detection of epidote and rounded zircons inProfil e S3,which , according to its ratio of Al to Fe seemed mainly derived from volcanic ash,show s that minerals from the underlying rock at a depth of >10m had beenmixe d throughout the profile. Thatmineral s from recent and old volcanic ashwer e present throughout the profiles shows that the process of mixing was active over a long period of time and is still active.Th e results illustrate what impact vertical and -38- sometimes lateral transport processes have on chemicalan d mineralogical properties of the profiles studied. Why soilmaterial s became so intimately mixed will be explained in the succeeding chapters,dealin g with the effect of termites onsoi l properties.

Fig. 13.Aeria l photograph of the Magena surface at Opapo; dark spots aremound s coveredwit h shrubs and trees; grass grows between the mounds. In foreground Luohomesteads . 5 Requirementso ftermite san dsom echemica leffect so ftermit eactivit y onsoil sstudie d

5.1 INTRODUCTION

Thefollowin gchapter swil ldemonstrat eth eeffect so ftermit eactivit y onsoi lproperties .Wh ytermite sd ocertai nthing si nsoil san dho wthi s relatest othei rrequirement san dt oprimar ylan dconditions ,i sdiscusse di n section5. 2o fthi schapter . Differencesi nrequirement so ftermite sar epartl yrelate dt oth etyp eo f nestthe ymake .Thre enes ttype sar eknown :mounds ,subterranea nnest san d nestsi ntrees .Nest si ntree swer eno tobserve di nth eare aunde rstudy . Moundswer eespeciall yobserve do nshallo wsoil san do nimperfectl yt opoorl y drainedsoils .Th emos tcommo nmoun dtype swere : (i)Bar emound so fabou t5 0c mhig han d 100c mwid ea tth ebase ,mad eb y Cubitermes spp..Thei rmound swer eno tobserve di ntille dfields .Moun d densitywa s25 0h a onth eshallo wsoil snea rWa n1 an d50-7 0o nth epoorl y drainedsoil snea rOpap o(fo rlocatio nse eFig . 1). (ii)Mound so fabou t1- 3m hig han d3-1 2m diamete ra tth ebas ewer emad eb y Pseudacanthotemee spp.an d Maavotevmes spp.(identifie db yC .Kooyman , personalcommunication) .Th elarges tmound swer eobserve do nth epoorl y drainedsoil so fth eMagen asurface ,wher eshrub san dtree sgre wo nth emound s andthe ynumbere dabou t5- 8ha - (Fig. 13). (iii)Smal lmound so f Pseudaaanthotermes (identifiedb yC .Kooyman )wit ha heighto f30-6 0c mwer eobserve di nwel ldraine dsoil swes to fKisi iTow na t analtitud eo fles stha n160 0m .The ywer econspicuou salon groad san d footpaths.O nth eshallo wsoil snea rWa n1 the ynumbere dabou t50 .h a . Subterraneannests ,mad eb y Odontotermes spp.an d Miorotermes spp.wer e common.Thei rnes tcentr ewa sdifficul tt olocat ei nth ewel ldraine ddee p soilsparticularl y ifthes esoil swer eporou swit ha lo wbul kdensit y(11 0kg . m )a swa sth ecas ewit hth emajorit yo fsoil si nth eare ao fstudy .Fungu s chambersseeme ddistribute drandoml y throughth esoil .Mos tchamber swit ha -40

Fig. 14.A .Youn g nest of Odontotermes in poorly drained soil on theMagen a surface near Opapo. B. Fungus combs and empty chambers of nest shown inA . C.Thi n section of fungus comb,buil t of faecal pellets consisting of macerated organicmaterial .

diameter of 10-25 cmoccurre d at a depth of 50-100cm , but some small fungus combs (2-5c m 0)wer e observed at depths of over 200 cm. In the imperfectly drained soils of Valley Bottoms in the eastern part of theare a of study,onl y nests of Odontotermes (Fig. 14)occurred . Nest sites were distinct due toa vegetation of shrubs and trees,absen t on soils between mounds,whic h carried a vegetation ofgrass . Most work in the area of study was done todescrib e and estimate the effect of termites on transport and spatial arrangement of soilmaterials . Less emphasis was placed on the effects of termite activity on chemical properties,becaus e these are extensively treated in the literature. A summary of these effects and some results from the area of study follow inSectio n 5.3 of this chapter. -41-

5.2 THE REQUIREMENTS OF TERMITES

The presence of termites, their numbers andwha t they do in soils is related to their requirements and to the degree the environment meets these requirements or can be changed tomee t them. Termites are very demanding as regards the climate in theirnest s and galleries,althoug h they occur inwidel y differing climatic regions from the humid rainforest to the semi-desert and from the equator to the mediterranean area,wher e night frostma y occur.The y occur inwel l drained soils but also in poorly drained and compact clay soils and inshallo w soils. It demonstrates their fantastic ability tomodif y primary land conditions to suit theirneeds .

Their needs and abilities toadap t primary land conditions,wil l be discussed below.Chapter s 6an d 7wil l showwha t this constant modification means for soil properties and how difficult it is tomodif y soilproperties . The requirements of termites concern the following items: sufficient food,a sufficient high temperature,a sufficiently highhumidity , sufficient soildept h fornesting , presence of suitable building materials,a low risk of flooding or waterlogging and a low risk of disturbance.

5.2.1 Food

Fungus growing termites in theare a of study make foraging galleries to the placewher e they collect the organic detritus (including wood). Inside the soil, walls ofmai n channels are plastered with salivated pellets of earth. Tunnelsabov e ground are entirely constructed of those pelletsan d radiate to the placeswher e they collect food (Fig. 15).Th e construction of foraging channels and chambers for fungus combs are important activities through which soil ismodified . Combs are built from faecal pellets ofmacerate d organic material onwhic h termites cultivate certain species of fungi (Fig.14) .Th e fungi effectively decompose organic material including lignin and cellulose so that itbecome s digestible for termites.Th ewhit e spherules of fungal hyphae are a particularly high quality food resource for them (Section5.3) , providing the queen and her off-springwit h food rich innitrogen . No figures are available about the number of subterranean termites in soils of Kisiian d their consumption of organic debris.Fo r comparison data cited byWoo d and Sands (1978)ar e shown inTabl e 8.A sLe e andWoo d (1971) rightly observe the total respiration by termites per unit of surface is of -42-

Fig. 15.Th e termite species inKisii , which forage above ground, cover the food they eat,wit h a layer of soil consisting of salivated pellets of earth.

the same order ofmagnitud e as that of large herbivores grazing on the savanna of East Africa. Most authors agree that termites,particularl y the fungus growers,ar e by far themos t efficient cellulose and lignin decomposers innatur e (Matsumoto, 1976; Lee andWood , 1971a;Wood , 1978;Woo d and Sands,197 8an dCollins , 1982).

5.2.2 Tempe rature

Termites tolerate only a narrow range in temperature, particularly in their nest.Dail y fluctuations in temparature within theirnes t are controlled throughmetaboli c activity and ventilation.Mor e seasonal fluctuations in temperature are controlled through variation of the depth atwhic h they construct their nest chambers.Accordin g toSand s (1965)termite smigrat e to deeper layerswhe n surface temperatures become toohig h and theymov e to upper layerswhe n surface temperatures decrease.Becaus e termites are very susceptible to desiccation,vertica l migration could moreover be due to a humidity effect (Bouillon, 1970 and Ferrar, 1982). -43-

Table8 .Biomass ,consumption ,productio nan drespiratio no ftermite si nthre e savannahecosystem si nWes tAfric aan drainfores ti nMalaysi a(afte r Matsumoto,1976 )fro mWoo dan dSands ,1978 .

Live Annualconsumption " Annual Annual weight production* respiration Ecosystem (gpe r kJpe rm 2 (kJpe rm ) (kJpe rm )

Sahelsavanna h Soil-feeders 0 0 0 0 0 Macrotermitinae 0.,7 2 15.8 298 24 31 Others 0,,2 4 2.6 49 4 10

Total 0.96 18.4 347 28 41

S.Guine asavanna h Soil-feeders 0.66 - 136 11 28 Macrotermitinae 6.39 139.9 2635 209 272 Others 3.54 38.7 729 58 143

Total 10.59 178.6 3500 278 443

Derivedsavanna h Soil-feeders 0.16 - 33 3 6 Macrotermitinae 0.64 14.0 264 21 25 Others 0.93 10.2 192 15 39

Total 1.73 24.2 489 39 70

Rainforest Soil-feeders" 1.62 - - - - Macrotermitinaea 1.62 35.5 669 53 72 Others" 1.83 20.0 377 30 72

Total 3.55 55.5 1046 83 144

Soil-feedersno tstudie d butman y species present;on especie so f Macrotermitinae andthre e species ofnon-Macrotermitina estudie dbu tman y other species present (total species= 56 ;Ab e& Matsumoto , personal communication). Consumption: calculated onbasi so fcalorifi cvalu eo fplan t tissue= 18. 8 kJpe rg ;consumptio n byMacrotermitina e= 6 0m gpe rg pe rda yan db y non-Macrotermitinae (i.e. 'others')= 3 0m gpe rg pe rday ;fo r'soil - feeders' consumption based onsam e calorific requirements asothe rnon - Macrotermitinae =20 6k J perg pe ryear . Production:base d onliv e weight/dry weight= 4. 0an dproduction :biomas s ratiofo rMacrotermitina e =6:1an dfo rsoil-feeder s andothe r non-Macrotermitinae =3:1 . -44-

Termites,occur throughout theare ao fstudy ,s otha t theyar eabl et o control the temperaturesi nthei rnest sa tambien t averageannua l temperatures of 15*Ca t 2100m an do f22" Ca t 1500m . Theyar e capableo fdoin g this even atmuc h lowerambien t temperatures,becaus e theywer e observed ata n elevation of 2400m i ngrasslan d at Nyahururui nKenya ,wher eabsolut e lowest and highest air temperatures reachvalue so f0.6° Can d 21.2°C respectively (calculated accordingt oBraun , 1980).

5.2.3 Moisture

Termites needmoistur e for (1)maintenanc e ofa hig h humidity leveli n their fungus gardens,(2 )th e production of saliva and (3)t opreven t desiccation.Par to fthei rwate rnee d isme tb ythei rmetabolis m but during periodso fdrough t thisi sinsufficien t and thentermite sar e reportedt o descendt oth egroundwate r table evena tdepth so f2 0m o rmor e (Watson,1974 ; Leean d Wood, 1971;Clou de tal . 1980). InWa n3 termit e burrowswer e found downt oth ebotto m of thepi ta t7m i nth e zoneo ffluctuatin g groundwater.

5.2.4 Soil depth for nesting and building materials

Subterranean termitesnee da certai n depth ofnon-indurate d soilfo r construction of theirnests ,s otha t theywer eno t observedi nth e shallow soils nearWa n 1.Mos t termitesals onee d some clay for theconstructio no f nestsan d channels andwher e theyar eno tabl e tofin d enoughclay ,a sit e willbecom eunsuitabl e for them.Th e occurrenceo ftermit eheap so nsand y soilsi nAruam a Stateo fRi od eJaneir oi nBrasi lillustrate s this:wher e the clayey subsoilwa sa tadept h of less than 125cm ,heap swer e present,bu t not where clay was deeper than 125c m(J . Bennema,persona lcommunication) .

5.2.5 Bisk of flooding or waterlogging

Termites are reported nott ob esusceptibl et olo w concentrationso f oxygeni nsoi lai r (Leean d Wood, 1971a). Onlyhig h carbon dioxide concentrationso f30-40 %ha d effecto nlowe r termites (alltermite s except the Family of the Termitidae);a tthes e concentrations the termites became motionless,eve ni nth epresenc eo f20 %oxygen .Coato n (1958)report s that

Hodotermes mossambiaus was ablet osurviv ei nterrain s flooded for periodso f -45-

3months ;the y apparently lived from the oxygenan d food reserves In their waterproof nests. Walls ofmound s observed in seasonally waterlogged areas in theare a of study were impervious towater ,becaus e no water entered laterally even though adjacent soilswer e completely waterlogged. Nomound swer e observed, however, inalmos t permanently waterlogged areas invalle y bottoms in East Kisii.

5.2.6 Risk of disturbance

Disturbance of nests and galleries may be caused by animals aswel l as man. Antbears when searching for termites thoroughly disturb thenest . Other predators, such as ants caused no damage to termite structures. Man disturbs nests when ploughing his fields,particularl y nests of the mound building species.Therefor e mounds of humus feederswer e not observed in cultivated fields. Because effects of cultivation on termites isa subject of further study, itwil lno t be discussed here.

5.3 EFFECTS ON CHEMICAL PROPERTIES

5.3.1 A literature review

Fungus growing termites feedmainl y onwoo d and leaf litterwhic h have very lownitroge n contents (0.05%, page 173,L a Fagean d Nutting, 1978). By contrast the termite's body has ahig h N content.Matsumot o (1976)studie d the flowan d accumulation of nitrogen during food preparation. He found theN content of combmateria l to be only slightly higher than that of its Source materials, but thewhit e spherules of fungalhypha e showed high Ncontent s (7.3% of dry weight).Matsumoto' sdat a indicate, that the nitrogenous substances are concentrated step by step during the production of food by termites. Thewhit e spherules finally provide theappropriat e nitrogenous food for termites to build their bodies,whic hhav e a Nconten t of 9-12% (dry weight)an d a C/N ratio of approximately 4 to 6.Matsumot o did not investigate the role of nitrogen fixing bacteria,bu t Boyer (1955,1956a )an d Pathak and Lehri (1959), cited byLe e andWoo d (1971a)foun d nitrogen fixing bacteria in combs of Macrotermitinae.N o data are available toestimat e theeffec t of -46-

nitrogen fixing bacteria inaccumulatin g nitrogen during food production. Studies regarding chemical effects of termites on soilmaterial s are usually based on a comparison of chemical characteristics of termite mounds with those of adjacent soils (Hesse, 1953,1955 ;Boyer , 1956b;Le e andWood , 1971b;Cornab yan d Krebs, 1975;Pomeroy , 1976a;Watson , 1976;Miedem a and van Vuure, 1977;D eWit , 1978an d Arshad, 1981, 1982). Thehig h level of bases inmounds ,particularl y of Ca,Mg ,K an d sometimesN a comparedwit h theadjacen t soil, isexplaine d as follows: - Large amounts of cellulose and lignin containing materials are accumulated in termite nests and decomposed. During this process carbon is efficiently transformed to CO2leavin g residual material rich inash . Inth e process organic matter becomes saturated with bases and looses its acid character. - Leaching is reduced due to the impermeablewall s of themound s (Umbrella effect). - Ifnecessary ,nes t chambers are moistened with water orwit hmois t soil material derived from thehumi d subsoil.Furthermore ,channel s made by termites improve permeability for air, so thathumi d air from the subsoil may passmor e rapidly towards the dry topsoil. Botheffect s accelerate accumulation of salts and minerals inmounds ,particularl y in dry regions. - Inpoorl ydraine d soils,capillar y rise from the saturated zone may be important. Environmental conditionsma y thus have an important effect on the amount of bases accumulated in termitemound s (Watson, 1976).

Studies regarding the chemical effects of subterranean termites with their diffuse and inconspicuous nest systems are relatively scarce. Robinson (1958)compare d the characteristics of soil transported by termites to cover mulch,wit h the characteristics of topsoilan d subsoil.Althoug h the color of the transported soilwa smor e like that of subsoil, its organic matter content and its CECwer e nearly as high as that of topsoil.H e found a systematic 1 1 1 I higher pHan d saturationwit h Ca andM g ,especiall y with respect to subsoil, butals o with respect to topsoil. The possible effect of saliva in this respect is emphasized by Harris (cited by Robinson, 1958). Thatadditio n of saliva to soil increases the exchangeable Caconten t is alsomentione d by Miedema and vanVuur e (1977); itma y at the same time raise the organic matter content äs suggested byPomero y (1976b)an d Wood (1976) . -47- 5.3.2 Results from the area of study

Chemical effects of the activity of termites on poorly drained soils of the Magena surface were studied near Opapo (Fig. 1),wher e a transect was dug through a termite mound. The difference in chemicalpropertie s of soils from amoun d of Pseudacanthotermes and theadjacen t soilwa s conspicuous (Fig. 16an d Table 9).

Depth ,,„, MIA M1B MIC MID M1E i I III

Fig. 16.Cross-sectio n through a termitemoun dwit h locations of 250 profiles and depths of sampling. 280

Table 9.Analytica l data of profiles (Pro.)o f a cross-section through a termite mound. For situation of profiles and samples see Fig. 16.CE C =catio n exchange capacity byNH.OA c at pH7 inmol/k g dry soilan d B.S.= bas e saturation.

Pro. Depth Org. C. PH Exchangeable bases CEC- B.S. Si02/Al203 (cm) (g/kg) (mol/kg so il) in clay H20 KCl iCa ±Mg K Na % (mole %)

MIA 20 17 0 170 51 4 2 280 81 4.2 170-180 14 6.7 5.7 248 46 4 6 301 100+ 4.4 220-230 11 7.6 6.1 250 35 4 6 317 93 4.8 270-280 1 n.d. n.d. 265 39 3 11 303 100+ 4.2 fungus 51 6.5 5.4 305 54 5 8 358 100+ 4.8 material

MID 0-10 15 5.8 4.0 72 18 2 2 148 64 4.4 10-20 10 5.8 3.8 61 9 1 6 122 63 4.8 20-30 6 5.9 3.6 66 7 2 7 129 64 4.0 30-40 8 5.5 3.5 78 8 2 12 173 58 3.8 50-60 9 5.6 3.5 101 11 2 18 229 58 3.4 60-70 8 5.6 3.5 95 10 2 17 205 60 3.5

MlE 10-20 9 5.8 3.7 39 2 1 4 91 51 4.5 30-40 7 5.9 3.8 32 1 1 5 79 49 3.9 60-90 12 6.0 3.7 133 3 3 36 378 46 2.9 -48-

The base saturation ranged fromapproximatel y 1 in the centre of themoun d to 0.5 in theadjacen t soil. The pH (KCl)decrease d accordingly from 5.7 to 6.1 in themoun d to 3.5 to 3.8 in theadjacen t soil. The organic matter content in themoun d differed little from that of theadjacen t soil. The results are in linewit h similar studies discussed earlier. When considering SiC^/A^Oo molar ratios (Table 9),value s werehighe r in themoun d (4.2-A.8) than in the adjacent soil (2.9-4.5). These higher ratios could point at a different mineralogy of clay in themoun d and the adjacent soil. Bachelier (1978)quote s several authorswh o found higher ratios in mounds. Some of them found montmorillonite inmounds ,wherea s adjacent soils contained kaolinite and gibbslte only.Condition s of a high pH, in the presence of much Ca,M gan d Si in the soil solution as commonly occur in mounds,woul d favour formation of 2:1 typecla yminerals . Concentration of silica in the soil solution of soils in theare a under study was probably rather high.Thi s follows from the presence of poorly crystalline clay constituents with ahig hmola r ratio of Si02 to A120,. These high ratiosma y be the result of theweatherin g of thehighl y siliceous peralkaline volcanic ash (Wielemaker and Wakatsuki, 1984an d Wielemaker, 1984) fromwhic h part of thefine r soilmaterial s were derived. 6 Micromorphologicalstudie s of spatialarrangemen t of soilmaterial s by termites

6.1 INTRODUCTION

Chapter 4 showed that soilmaterial s from various parentmaterial s were intimately mixed. Mixing is the result of transport and rearrangement of soil 'als.Whic h pedogenetic processes are responsible formixin g and how they follows from an interpretation of micromorphological features,describe d s chapter. When not considering thepossibl e role of erosionan d sedimentation, which will be treated in the next chapter,rearrangemen t and transport of soil material are the result of two sets of interacting pedogenetic processes.T o the first set of processes belong: (i) Transport of soilmaterial s by soil fauna. (ii) Transport due to physical processes sucha s shrinking and swelling, when soilmaterial s become alternately dry andwet . During thefirs t set of processes,pore s and cracksar e being formed allowing the second set of processes to take place.T o the second seto f processes belong: (i) Infilling of pores and cracks due to dispersion and translocation of clay. (ii) Collapse of pores and cracks due to instability of structure, (iii) Compression of pores and cracks due toa n overburden, (iv) Compaction of pores and cracks due to farmmachiner y or trampling by cattle and game. (v) Structural decay due topractice s such as ploughing, (vi) Surface sealing due to insufficient coverage byvegetatio n orcrops . Weight of overburden, Theological characteristics of soilmaterials , moisture content and changes inmoistur e content are important rate determining characteristics for processes i, ii and iiian d partly ivan dvi . Management practices are especially important for processes iv,v an d vi. The -50-

second set of processes leads to complete compaction of soil materials if collapsed structures and pores are not replaced or restored. Restoration of structure and pores due tophysica l processes is especially important insoil swit h a large fraction of 2:1 clays,bu t the majority of soils in the tropics has clays with a lower ratio of Si to Al. Particularly in the latter soils,spee d of transport of soilmaterial s is thus regulated by: (i) theactivit y of soil fauna and (ii)th e ratea t which voids and structure elements decay. These two processes are interacting. It is in fact ahomeostati c control mechanism,whic h will function as long as theabilit y or potential of soil animals for restoration isno t exceeded.

6.2 METHODS

The definitions used in themicro-morphologica l analysis are those according to Brewer (1976) . Area fraction of argillans and papules was determined in the solid phase by point counting; theare a fraction of other features was determined by a ranking of thin sections for each feature.pF : mass fractiono fmoistur e insaturate d soilan d soilafte r equilibration with a sandbox at pF0.4 , 1.0, 1.5 and 2.0,a kaoli n box (for pF 2.3 and 2.7)an d pressure equipment forp F 3.0, 3.7 and 4.2 (Stakman et al., 1969).

6.3 RESULTS

6.3.1 Termite-made pellets

Two types of pellets,mad e by termites,wer e observed in thinsections ; the smallest with a diameter of 40-80 pmwer e nicely round with a smooth surface (Fig. 17).Th e bigger ones with a diameter of 200-800 \m had an irregular"bu tsmoot h surface. Both types occurred as backfillmateria l in channels and chambers down to and between the weathered rock or asmatri x material in theuppe r horizons of some profiles such asWa n 1,2 an d 3.Th e depth atwhic h they were observed and also thearche d pattern of some of the backfillings suggested that the -51-

Fig. 17.Ora l (A)an d faecal (B)pel Fig. 18.Arche d pattern inchannel , lets in backfill material,observe d backfilled by termites. ina thin section, takenfro ma depth of 300-315 cm in profile Wan 3.

pellets were made by termites (Fig.18) . The bigger pelletswer eapparentl y formed by themandible s of the termites,bu t the smaller pellets had a form and size similar to the pellets fromwhic h fungus combs are built andwhic h were faecal in origin (Fig.14C) . In thewel l drained deep soils,th eare a fraction of termite-made pellets decreased strongly from the surface down toa depth of 100-200cm ; it then remained nearly constant or decreased slightly down to thegrave l line. Below the gravel line theare a fraction decreased with depth; thedecreas e coincided with an increase in theare a fraction of weathered rock (Table10) . Most termite-made pelletswer e observed in thin sections of Wan 1,2 and 3,followe d by Wan 9an d 210.Leas t pellets were observed in Wan 4,Wa n 5an d 2 M/4, indecreasin g order. -52-

Table 10.Are a fractiono fth esoli d phase composedo fora lan dfaeca l pellets inth estudie d profiles,estimate d byrankin g ofthi n sections (+++,>0.5;++ , 0.5-0.2;+ ,0.2-0.05 ; (+),<0.05-0. 0an d- ,no tdetectable) .x Partwit h soil structure;x x partwit h rock structure.

Wan1 Wan3 Wan9 Wan4

Depth ranking Depth ranking Depth ranking Depth ranking (cm) (cm) (cm) (cm)

5-1 3 +-H- 0-4 0 +++ 40-6 0 ++ 0-214 + 40-115 ++ + 260-275 (+) 115-315 + 305-325 (+) 320-335 - 315-700 (+) 330-350x + 330-350xx

Wan2 210 M1E Wan5

Depth ranking Depth ranking Depth ranking Depth ranking (cm) (cm) (cm) (cm)

0-3 0 +++ 5-2 0 -H- 0-5 0 + 0-125 + 55-100 ++ 40-150 + 50-110 - 166-295 - 120-165 +

6.3.2 Voids and porosity

The totalare a occupied byvoid s (pores)wa sbigges t inth euppe r parto f the soilsdow nt oa dept h of100-20 0cm ,particularl y inWa n1 ,Wa n3 ,Wa n4 andWa n5 (Fig .19) .Thi s trendwa sals o observed inFig .20 ,assumin ga relation between theare a occupied bylarge r voidsan dwate r held between pF0 andp F3 .Th eamoun t decreased down toabou t 100c man dremaine d relatively constant below that depth. Typeo fvoid san dthei r relative distribution differed among soils andwit h deptha sshow ni nFig .19 .

Voids between pellets and grains. Wheretermite-mad epellet sdominate dth e matrix,th eare a occupied bypore swa salmos t continuous between pellets (compound packing voids inFig .1 9an d21A) .Deepe r inth eprofil e termite- made pellets seemed tohav emerge d together,s otha t voids were less continuous (interconnected vughs)o rmerel y consisted ofisolate d openings (vughs)(Fig .21A) .Th enumerou s transitions which existed between loosely -53-

M1E WAN 5 WAN 4 WAN 9 WAN 3 WAN 2 WAN 1

j qj •4 \ INI iiiil 50 mi Mil iiii mil mi 100 Hin I ! I inn in I mi iiiil

IN i inn in M II 200 - m i inn ill m inn m II 300-

fcgravel line inn I I

I area fraction of1. 0 400- I areafractio n of0. 6 etc.

| areafractio n> 0.0 5

p.v.=packing void IIII ch.=chamber

600

700- Depth (cm) Fig. 19.Rankin g of theare a fraction of voids in thin sections, p.v.= packing void; ch= chamber. -54-

Depth (cm)

U i A • • • 50- T * D •TD A 100 - TD D • 150 • profiles T D A Man 2 • Wan 3 D » 210 200 • D NARS D

PRO- P , Fig. 20.Volum e fraction ofwate r held between 0 0.2 0.4 0.6 0.8 pFOan d pF3a t different depths inwel l drained volumefractio no fwate r soils. (pFO- pF3 )

packed termite-made pellets and amatri x inwhic h individual pellets were no longer recognizable suggested, however,a faunal origin,eve n for the isolated openings (vughs), though much degraded comparedwit h their original form. Similar vughs were not degraded,wher e they occurred in densely packed walls of chambers and channels made by termites (Fig. 2IB). The vughs between large-sized quartz grains in thin sections of profile Wan 2 (simple packings voids inFig . 19)wer e not faunal inorigin .

.Channels, chambers and cavities. Channels had a diameter of 0.8-5 mm of which the larger sizedwer e usually linedwit h pellets.Th e form of the channels and also their distribution throughout the profile indicated their termite-made origin. The area fraction of channels varied littlewit h depth and soil type except for the imperfectly to poorly drained subsoils of Wan 5an d M1Ean d the horizon below the gravel line in thewel l drained profiles (Fig.19) . The macro-morphology of fungus chambers was described in chapter 5. In thin sections ofwel l drained soils only backfilled chambers were observed, easily recognizable from their densely and regularly constructed walls (Fig. 21B)an d from the backfillmateria l composed of loosely packed termite-made pellets (see for instance Fig.17) . Irregular shaped backfilled cavities were also observed amid weathered rock in theweathere d rock zone,whic h were probably developed as a result of the termite's search forwate r and/or wet clay (Bachelier, 1978;Le e andWood , -55-

Fig. 21.A . Transitions between amatri x with compound packing voids (1), and a matrixwit h interconnected vughs (2)an d simple packing voids (3)sugges t that pellets loosely packed by termites gradually merge together as the matrix grows older. B. Simple packing voids if occurring in a regular pattern inwall s of chambers and channels,ar e made bytermites .

1971a).Th e imperfectly drained subsoil of Wan 5showe d 3-10m m wide cavities, partly infilled with very fine sand, silt and clay (Fig.22) .Thoug h few of these cavitieswer e backfilled by termites,thei r persistence at that depth could hardly be imagined without the action of termites. In the subsoil (100- 110cm )o f ProfileMID , all channels were backfilled by termites.

Cracks .Crack s (crazean d skew planes)cause d by shrinking and swelling were common in thin sections of the subsoil of Wan 4,Wa n 5an d most important in the subsoil of M1E.

6.3.3 Redistribution and reorientation of skeleton grains and plasma

Illuviated soil materials and their redistribution. Illuviationo fcla ywit h or without silt and sand is a decay process,whic h leads to texture -56-

Fig. 22.Composit e cutans consist of parallel layers of illuviated fine sand, silt and oriented clay (thin section from profile MIDa ta depth of about 60cm) .

differentiation and compaction of the illuviated horizon,unles s counteracted by redistribution, due tophysica l and/or biological processes (see Section 6.1). Illuviated clay occurs as argillans (oriented clay)o r ferri-argillans if stained by iron (Brewer, 1976) . Argillans were observed inpart s of the profiles studied except inWa n 1. In the subsoils of Wan 4,Wa n 5an d MID, ferri-argillans were commonly associated with parallel bands of illuviated silt and fine sand (Fig.22) .Thes e composite illuviation features are called composite cutans.Th e term matri-ferriargillana s defined byVa n Schuylenborgh et al. (1970)doe s not apply here because the term refers to illuviation of a mixture of clay,sil t (sand)an d organic matter. Thickness of theargillans , depth of occurrence and also their associationwit h different fabrics were recorded inorde r toestimat e past or present strength of illuviation processes.

On theothe r hand papules (pieces of oriented clay)wer e signs of redistribution. They were often observed within termite-made pellets,bu t also ina matri x inwhic h individual pellets were less clear. In the subsoils of Wan 5,MID ,M1E ,bi g (500 |jm)an d small (20 jjm)irregularl y shaped papules were observed. Sizean d shape excluded a faunalorigi n of the bigger papules. Illuviation phenomenawer e weak,o r absent in theuppe r 2-3m of Profiles 210 (Table 11)an d Wan 9, common and distinct in Profiles Wan 2an d Wan 3 -57-

Table 11.Are afraction so foriente dcla y (ferri-argillansan dpapules )determine do nth e solidphase ,an dth edisplace d fraction (displ.)o foriente dcla y (papules)fo rthre e profileso fth eWanjar ecaten aan dfo rprofil e210 .Papule swit hasteris kwer eno to f faunalorigin .

Wan2 Wan3 Wan5 210 depth total displ. depth total displ. depth total displ. depth total displ. (cm) (cm) (cm) (cm)

0-3 0 - - 0-1 5 <0.003 # 8-2 3 <0.001 0.50 5-2 0 - _ 55-7 0 0.017 0.30 40-5 5 0.017 0.24 38-5 3 0.003 0.48 40-5 5 - - 83-9 8 0.023 0.23 85-100 0.023 0.26 70-8 5 0.016 0.33 70-8 5 - - 120-135 0.014 0.21 130-145 0.016 0.44 110-125 0.047 0.38 100-115 0.002 tr 160-285 <0.003 n.d 166-182 0.081 0.06* 135-150 0.009 tr 285-300 0.014 0.29 232-247 0.036 0.60* 315-330 0.022 0.45 280-295 0.090 0.11* 345-350 0.062 0.13 0.056 0.20* 600-700 0.055 0.05

(Table 11)particularl y above adept h of16 0c man d frequent and strongly developed inProfile sWa n5 (Tabl e 11),Wa n4 an dMID .(Fig .22) .Alon g cracks and incavitie s inth esubsoi l ofth elas t three profiles,bi g (100-1000 (jm) composite cutans were observed (Fig.22) .Thi san d thepresenc e ofbi gpapule s (Table 11)pointe d atinstabilit y of thesoi lfabric . Incontras t with Profile MID,whic hwa ssituate d onth eslop e ofa termitemound , theadjacen t ProfileM1 Eshowe d only fewcomposit e cutans.Th e subsoil of this profilewas ,however ,ver y compact;n ocavitie s andonl yfe w crackswer e observed, probably duet oth eabsenc e of termite activity. Moreover,th eeluvia l horizonha d lostmos t ofit sclay ,s otha t the process of eluviationha dcom e toa nend .Unde r such circumstances llluviationca n hardly take place,whil e existing features maydisappea r duet oshrinkin gan d swelling (Appendix3) .

Reorientation due to shrinking and swelling. Accordingt oth eplasmi cfabri c (Appendix 3)shrinkin g and swelling phenomena arewea k orabsen t inth ewel l drained profiles,moderat e inth esomewha t imperfectly drained subsoilso f profiles Wan4 an dWa n5 an dstron g inth esubsoi lo fth epoorl y drained profileM1E .

Redistribution due to the soil fauna. Thezon eo fweathere droc ki nWa n2 (depth 1.60 m), Wan3 (4-> 7m )an dWa n9 (> 3m )wa sillustrativ e ofwha t transport by soil fauna canachieve . Soila t those depths consisted oftw o contrasting fabrics.On efabri c seemed olda si twa sassociate d with thick -58- welldevelope d ferri-argillanswithi n a rather dense fabric.Th e other fabric consisted of loosely packed pellets,whic h had not merged so that the fabric seemed young.Th e absence of papules in theyoun g fabric indicated that the pellets ditno t originate from the surrounding matrix. The proportion of plant opalan d glassmineral s observed in thesepellet s (Fig.23 )an d also their form and color were similar to that observed in pellets of theuppe r 70c m of theprofile ,i n contrast with the "old"fabri cwher e suchmineral s were not observed. The pellets were found as backfill material in cavities and also as linings of larger sized channels (0 1cm) .

Contrasting colors inA an d Bhorizon s indicated transport of soil material from theA to theB horizo n and from theB to theA horizon.Th e dark colored material in theuppe r part of the Bhorizo n was usually found in backfilled chambers and channels and in the deeper part only in large channels and small chambers.Th epatter n of the backfillings was usually arched (Fig. 18)whic h left little doubt about the faunalorigi n of those backfillings. Similar arched patterns at 15m depth in soils of Zambia were attributed to termites (Cloud et al. 1980). Visible relicts of transport by soil fauna were most clearly expressed in Profiles Wan 3,Wa n 2an d Wan 9 (Table 12),bu t pellets observed inProfil e 210di d not contrastwit h the surrounding matrix.

Fig. 23.Th e presence of opal phytoliths (bar-shaped minerals)an d glass fragments (not observable), and themorpholog y of the pellets with dark finely dividedhumus , observed in a backfilled cavity at a depth of 350c m inProfil e Wan 3,indicat e that pellets originated from the upper horizon of theprofile . -59-

Table12 .Rankin go fbiologica l(Biol. )an ddegradationa l (Degr.)propertie si nth estudie dprofiles , basedo nTable s1 0an d1 1an ddescription si n6.3.3 .Rankin go fbiologica lpropertie swa sbase do na n estimateo fth eare afractio no fV (voids) :compoun dpackin gvoids ,channels ,chamber san dcavitie s from- ,no tdetectabl et oMil , verybig ;S (structure) :termite-mad epellet sfro m- ,no tdetectabl e to+++ ,big ;P (papules) ,mad eb yth esoi lfaun afro m- ,no tdetectabl et o++ ,fairl ybi gan dR (relicts):soi loriginatin g fromothe rhorizon san dtransporte db ytermite sfro m- ,no tdetectabl e to-H-+ ,big .Rankin go fdegradationa lpropertie swa sbase do na nestimat eo fth eare afractio no fV (voids):ske wan dcraz eplane san dmasepi cfabric ,fro m- ,no tdetectabl et o000 ,big ;A (argillans) : from- ,no tdetectabl et o000 ,bi gan dP (papules) :no to fbiologica lorigi nfro m- ,no tdetectabl e to000 ,big .Reductio nphenomena ,i fpresent ,ar eindicate dwith :gl ,sligh treduction ;g2 ,moderat e reduction;g3 ,moderatel y strongreduction ;g 4 ,stron greductio nan dg5 ,ver ystron greduction ,tr , traces;n.d .no tdetermined .Dept hi ncm .

Profiles Wan9 210 Wan3 Wan2 Wan4 Wan5 MjE MjD

Depth 0-50 0-50 0-50 0-50 0-50 0-50 0-50 0-50g 3 Biol.V +(+ ) n.d 1 II 1 ++ + ++ (+) n.d S +(+ ) +(+ ) +++ +++ + + + n.d P tr tr + + ++ +(+) (+) n.d. R + (+) +++ ++ + + - n.d. Degr.V ------A 0 0 00 0(0) 000 P _ : _

Depth 50-100/115 50-100 50-110 50-100 50-120 50-130g l 50-100g 5 50-100g 3 Biol.V +(+) n.d ++(+) +(+) + ++ - n.d S + + ++ ++ + + - n.d P tr tr + + ++ +(+) - n.d R ++ + - - - n.d Degr.V - - - - (0) (0) 000 n.d A - - 0 0 00 0(0) 0 000 P ------(0) - Depth 100/115-315 100-150 110-315/345 100-170 120-250g l 160-300g 3 Biol.V +(+) n.d +(+) +(+) + (+) S (+) + + + + - P tr tr (+) (+) + - R + + - - Degr.V - - - - (0) 0 A - - (0)(to160cm ) 0 00 000 P - - - - 0 000 Depth >315-350+ 345-700+ 250-335g 2 Biol.V + + + S + + - P tr tr (+) R + + - Degr.V - - (0) A 00 00 00 P tr tr 0

Redistribution due to mechanical transport. Presenceo fcomposit ecutan si n the subsoils of Wan 4,Wa n 5an d MIDsuggeste d backfilling of large cavities by sedimentation fromwater .Thi s and the presence of big papules were suggestive of a process of internal erosionan d collapse of structure, which would continue to operate as long as termites kept excavating new holes and cavities (Section6.1) . -60-

Thisproces sma yals ohel pt oexplai ncertai nfeature so fgley ,whic h appearedi nthi nsection sa sdistinc tan dwel ldevelope ddar kferri cnodules , butals oa sdens edar kreddis hbrow nmatri xdomains .Sometime sth edomain s resembled theweathere droc k(botto mpar to fProfil eWa n5 an dWa n 9),bu t theyals oresemble da noxidize dreddis hbrow nmatri xo ra mor ereduce dgrayis h brownmatri x(Profil eWa n5 ,70-8 5cm) .Thei rrounde dform san ddistinc t boundariescoul dwel lb eth eresul to fabrasio ni nvie wo fth econtinuou s displacemento fsoi lmaterials .Disintegratin gnodule san ddomain swer eals o observed,whic hsuggeste dthei rincorporatio ni nth esurroundin gmatrix .

Presence of organic matter. Theorgani cmatte ri nth eplasm ao ffabric s discussedi nAppendi x3 wa sdar kcolore dan dintimatel ymixe dwit hth eminera l fraction.I twa sno tacidi cs otha tth eorgani cmatte rconform st oth e definitiono fmul l(Kubiena ,1955) .Piece so fdar khumus ,i fpresent ,wer e especiallyobservabl ei nloosel ypacke dpellet s(Fig .23 )bu thumu swa s usuallypresen tlik ea dy egivin gth eped sa nopaqu edar kcolor .

6.4 DISCUSSIONAN DCONCLUSION S

6.4.1 features made by the soil fauna and features indicative of their degradation

Table1 2show stha tth eare afractio no ffeature sindicativ eo f degradationi nth euppe rmetr eo fth esoi lprofil eincrease sfro mWa n9 t oM1 E andMID .

Impeded drained profiles. InProfile sWa n4 ,Wa n5 an dM1 Ethes epropertie s arecorrelate dwit hdrainag econdition sbecaus ethei rare afractio nan d intensityincreas efro mProfil eWa n4 t oProfil eM1 Ean dMID .Bu tpropertie s indicativeo fdegradatio nsuc ha sargillan ssee mt odisappea rdu et oshrinkin g andswellin gi fther ei sn ofaun at orestor eth epor esyste ms otha tcla yca n illuviate(Sectio n6.1) . Whendeca yprocesse sar es ostron gtha tclay ,sil tan dfin esan dactivel y illuviate,termit eactivit ymus tals ob estron gt opreven tformatio no fa compacthorizo nan da nabrup ttextura lchang esuc ha sdevelope di nProfil e M1E.Thi smean stha tProfil eWa n5 ,whic hha spropertie sindicativ eo fstron g -61-

degradation,woul d have beena degraded soil,wer e itno t for thever y high biological acitivity,whic h apparently prevented texture differentiation and compaction of the subsoilwit h its unstable structure. Furthermore, the presence of composite cutans around cavities in the reduced and therefore most unstable part of thematrix , reflect processes operating atpresent . In the somewhat better drained Profile Wan 4,degradationa l processes were apparently less strong than inWa n 5,bu t stronger than in thewel l drained Profiles Wan 9, 210,Wa n 3an d Wan 2(Tabl e 12),a s follows from the area fraction of argillans in these soils.Th e presence ofman y papules of biological originan d theabsenc e of a sharp texture change indicate a high biological activity even though therewer e not soman y biological voids and structure elements as in Profile Wan 3.

Well drained profiles. Properties indicative of degradationwer e least developed in Profile Wan 9an d Profile 210,whic h follows from the virtual absence of argillans in theuppe r part of the soil.Argillan s were only observed inan d below the gravel line,i n the "old"dens e fabric,whic h strongly contrasted with theapparentl y more recent fabricmad e up of loosely packed pellets. Because the younger fabric did not showan yargillans , illuviation of clay seemed not to be anactiv e process. When properties indicative of degradation and of faunal activity in Profiles Wan 9an d 210ar e soweak , it follows that present activity of the soil fauna is much smaller than in Profiles Wan 3an d Wan 2,whic h had more properties indicative of degradation and of faunal activity.Despit e the few signs of degradation in Profiles Wan 9an d 210,the y were most susceptible to sealing, structure decay and erosion (Wielemaker and Boxem, 1982)o f all profiles presented in Table 12 (except for M1Ean d MID). This apparently contradicts with theabsenc e of properties indicative of degradation (Table 12), but can be explained by assuming that old termite-made pellets are less stable than fresh ones. The stability of fresh pellets could be due to thepresenc e ofmicrobia l slime probably composed of muco-polysaccharides (Hayes,1983 )an d also to the high saturationwit h Ca ions. Products ofmicrobia l slime are easily decomposed (VanDijk , 1961)an d Ca ions may gradually leach,s o that stability declines as pellets grow older. This process of "ageing"woul d also result inmorphologica l changes such as disintegration ormergin g together of individual pellets (Bal, 1973), -62- unless the soilwer e consumed again before this occurred.Agein g of the soil fabric and a consequent loss of stability could then only occur in soils or soil layerswit h a low biological activity such asWa n 9an d 210an d this could explainwh y these profiles are so susceptible tosealin g and surface erosion. Besides structure stability, termites also controlled other physical soil characteristics such as porosity and permeability (Table 12).Al l pores in the well drained profiles down to the groundwater tablewer e eithermad e by the soil fauna or derived from pores made by the soil fauna.Onl y themos t poorly drained soils had mainly pores of physical origin (Fig. 19,ske w and craze planes).Porosit y and consequently permeability inal l soils studied were, therefore,determine d by theactivit y of soilfaun a and ranged from aver y rapid permeability and ahig h porosity in thewel l drained soils to a slow permeability and aver y low porosity in themos t poorly drained soils (M1E).

Formationan d disappearance of features of gley seemed also controlled by termite activity.Th e numerous transitions betweenweatherin g rock, concretions,oxidize d soil matrix and reduced soil matrix observed in imperfectly drained soilhorizons ,suggeste d that pieces of soilfabri c were continuously translocated and this propably caused their transformation, disappearance and new formation.

6.4.2 Pedoturbation and mineral distribution

Themineralogica l properties as described inChapte r 4, showed particularly that parent material from volcanic ashan d local country rock were intimately mixed. How the turbation process works andwha t effect itha s on mineral distribution in soils follows froma n appraisal of the results of Chapter 4an d this chapter.

Distribution of volcanic glass and plant opal .I nwel ldraine ddee psoil s volcanic glass,a n easily weatherable mineral,wa s especially observed in the upper 40 cm of theprofile s studied; it then decreased irregularly toa low butmeasurabl e mass fraction ata depth of 100-120cm ;belo w that depth it remained practically constant down to thegrave l line (Table 5).Belo w the gravel line it decreased with depth due toa n increase in the area fraction of soilwit h rock structure (Section 4.2,Profil e Wan 3).Thre ekind s of mechanisms could have played a role in the formation of this distribution: -63-

(i) Active upward and downward transport of minerals related to the construction of foraging channels on the soil surface and to the construction and backfilling of chambers and channels. (ii) Downward transport ofmineral s by gravity. (iii)A difference in the degree of weathering of glass minerals from upper and lower horizons due toa difference inage . The observation of termite-made pellets containing glass and plant opal minerals in cavities in the zone ofweathere d rock and along main channels in the zoneabov e it (Section 6.3) suggests anactiv e transport of soil particles from theA horizon even down to thezon e of fluctuating groundwater. Mechanism (ii)seem s irrelevant because no signs of mass transport bywate r were observed apart from some illuviated clay inProfil eWa n 2an d theuppe r part of Profile Wan 3.Als omechanis m (iii)seem s unlikely because pellets and glass grains in topsoll and subsoil are similar. Although signs of mass transport bywate r were not observed, the deposition of soil on the surface and the regular passage of animals through a soilwit h a loosely packed matrix may increase the chance,tha t glass minerals sink in theprofile .Thi s process,i f operative,wil l only be important above 100c m depthwher e big pores were most common.Mechanis m (ii)would ,belo w this depth result ina decreasing number of glass grainswit h increasing depth,whic h isdenie d by the constant mass fraction of grains observed there.

Presence of composite cutans in the imperfectly topoorl y drained subsoils of Wan 4,Wa n 5an d M1Ewit h unstable structures are a signo f gravity transport bywater . Plant opalmineral s were commonly observed in the composite cutans and their distribution is probably related to this downward transport. Incontras t with thewel ]draine d soils,mechanis m (ii)seem s an important process especially in horizons with poor drainage conditions.

Distribution of epidote .Th emas s fraction of ash-derived soilmateria l in Profile S3wa s quite uniform even to a depth ofmor e than 7m (Section 4.2.5), but themas s fraction of epidoteminerals ,decrease d from the surface down to a depth ofabou t 10m wher e epidote strongly increased (Appendix1) . It shows that soilmateria l from a depth of at least 10m is transported to theuppe r soilhorizo n (slope transport can be discarded as a source of epidoteminerals ,becaus e Profile S3 is situated on thefla t part of a ridge). That termites add soil from a depth of at least 7m to theuppe r soil horizon -64-

was already observed when studying glass distribution. The epidote distribution tells us that it ispreferentiall y added to the top 50-100c m of the soil profile,s o that thishorizo n gets enriched with fresh epidote minerals. Because also thedeepe r horizons consist ofmixe d soilmaterial is t seems that the lower percentage of epidote in thosehorizon s is duet o weathering. The continued addition of ash to the surface horizonan d its mixing with soil containing little orn o ash,wil l result in very deep mixed soils,a s isindee d reflected by themineralog y of the soils (Chapter4) .

Distribution of magnetite-IImMagnetite-II ,a rathe r stable mineral, was only detected in relatively old and deep soils (Chapter 4). Incontras t with glass, itwa s rather evenly distributed down to the gravel line at about3 m (Fig . 9). Its distribution in soils is the result ofmixin g overa ver y long period of time,s o that evena lo w transport activity of termites belowa dept oh f 1m (bu t above the stone line if present), resulted ina thoroug h mixing and accumulation of stable minerals from volcanic ash.

Depth (cm) 0• percentagemineral s inver yfin esan d A1

100-

200

300- ooooooo< ooooooc ooooooobooooooooooo gravellin e B3 III

Fig. 24.A .Widt h of the arrows indicates schematically,downwar d and upward transport activity asa functio n of depth in deep well drained soils. I, zone of nesting and foraging; II,zon e of refuge and passage to groundwater; III,zon e ofmoistur e supply in time of drought below thegrave l line (oooo). B. Schematic distribution of rather recent (glass,soli d line)an d old (magnetite II,broke n line)mineral s from volcanic ash in well drained profiles. -65-

Transportactivity andmineral distributionm Results of studies concerning transportactivit y of termites and itseffec t onminera l distribution in the deepwel ldraine d soilsar esummarize d in themode lpresente d inFig .24 .Fig . 24Ashow s that termitesar emos t active in theuppe r metre of the profile which is thezon e ofnestin g and foraging.Activit y in thezon ebelo w1 m i s much lower,bu t decreases only slightly with depth downt o theston e or gravel line. Below that lineactivit y did not decrease inth epart swit h soil structure; the overallactivit y did decrease,however ,du e toa n increasie n thevolum eo f saprolite. Theglas s distribution represents theresul t of transport activity overa relatively short period of time,whil e themagnetite-I Idistributio n results froma muc h longer period ofmixin g (Fig.24B) . The results show that termiteswit h subterraneannest swer e active in the deep subsoil,whic h contradicts Bachelier's (1978)statemen t that termites with subterraneannest s would derive soilmaterial s fromsurfac e layers only. 7 Formation of texture profiles in relation tovertica lan d lateral transport

7.1 TERMITES AND TEXTURE PROFILES

7.1.1 Introduction

Most soils in theKisi i areahav e a relatively gravel and stone free upper horizon over a layer rich ingrave l and stones.Thi s layer usually overlies weathered rock (Fig. 25)o r ironstone,bu t it occurs also on top of a B2t horizon. Such texture profiles are common inman y soils of the tropics, but there isconsiderabl e debate regarding their formation (Section7.3) .

Fig. 25.Th e layer consisting of quartz gravel separates an upper horizon free of gravel from aB C horizon inwhic h géodes are still in situ.Géode s above thegravel - layer were the source of thegravel , that accumulated in the gravel layer. -67-

Studies described here,ar e all done ina rather humid environment.Th e important role termites play in forming texture profiles with stone lines will first be demonstrated on thewel l drained soils of theWanjar e catena. It is also demonstrated on the poorly drained soils of the flat Magena surface.

7.1.2 Methods

Cumulative curves ofmas s percentages of the logarithm of different size fractions of sand and gravel (50-4000 um)ar e shown in Fig.2 9 (page 70) . Size fractions were:50-10 0 um,100-25 0 um,500-100 0 um, 1000-2000 uman d 2000- 4000 um. The quartilemeasure s and medianvalue swer e determined on graphs according toKrumbei n and PettiJohn (1938).Q l is the grain-size diameter with 25 percent of the grain-sizes smaller thanQ l and Q3 is the grain-size diameter with 75percen t of the grain-size smaller thanQ3 .Th emedia n is the grain-size diameter separating equalweight s (50%)o f larger and smaller grain-sizes. Log S is used as ameasur e for the degree of sorting, it is equivalent to (log Q3-log QD/2.

7.1.3 Results and discussion

7.1.3.1 Texture profiles inwel l drained soils of theWanjar e catena

Selective transport by termites. Fig.2 6 shows the size of pellets transported by termites and used for the construction of their nest opening. It indicates that theymak e pellets of a size,relate d to the opening between their mandibles. When comparing grain-size distribution of termite-made channels with that of the adjacent soilwit h rock structure in the zone of weathered rock of ProfileWa n 3,i t shows that channelwall s and their infillings have more grains larger than 2m m than the adjacent soil (Table 13).A s explained in Section 6.3 termites transport materials from theweatherin g rock upwards,s o that they caused the increase in grains larger than 2mm . Fig. 27 illustrates the same effect ona profile scale.Fin e grained material is removed from the deeper part of the profile and transported -68-

Fig.26 .A .Ventilatio n shaft of a nest of Odontotermes showing the size of pelletsmad e by theworke r termites. B. Pellets are fixed with a glue of saliva.

Table 13.Grainsiz e distribution of sand and gravel (0.05 - 4mm)fro m weathered rock,an d from soil material in channels at a depth of 6.5 m in profile Wan 3.

Fractions (mm) and mass percentages

2-4 1-2 0.42-1 0.21-0.42 0.105-0.210 0.05-0.105

Soil 25.4 38.4 21.0 7.0 5.1 3.1 Rock 10.2 43.8 26.3 10.3 6.8 2.6

upwards,wher e it forms ahorizo n composed of fine grained material. In the lower part of the profile where the fine grained material is removed, the proportion of coarse grained material increased. Termite activity decreases, however,wit h depth so that the grain-size distribution of soil materials deeper down gradually equals that of the original saprolite.

Sorting by termites. Fig.28 aan d 28bsho w that grain-size distribution of a termite mound depends on that of the corresponding substratum. In themoun d near Wan 1onl y the mass fraction of grains larger than 1m m decreased compared with the grain-size distribution of the substratum, but in themoun d near Wan 4,whic h is over a finer textured substratum, also the mass fraction of grains larger than0. 5 mm decreased. Themoun d of Wan 4ha d the best sorted sand (log So= 0.236, from log So= (log Q3 -lo g Ql)/2,Q 3an d Ql from Fig. 29), while sand in themoun d of Profile Wan 1wa s least sorted (log So= -69-

Profile WAN 3 Profile WAN 9

20 40 60 80 100' 330 415 1- 625'

515

560

925

850

1960

• 1500 Depth (m) Md (pin)

f o) Fl)F^ F3^ F4 ) F5^

Depth (m) Md (pm)

Fig. 27. Grain-size distribution and median diameter (Md) values in profile Wan 3 over diorite and profile Wan 9 over quartz diorite. Size fractions are: gravel, 2000-4000 Mm; Fl, 1000-2000 \m; F2, 500-1000 pm; F3, 250-500 urn; F4, 100-250 pm; F5, 50-100 |jm and F6 < 50 \m. The Md value is of sand and gravel only. -70-

Füass(% )

40- 1

a L _ ' " I I i u -\1 T 1 l r 1 1 1

4000 2000 1000 500 250 100 50 Logscal eo fdiameter s(firn ) ofsan dan dgrave l

Fig. 28.Relatio n between mass percentages of different size fractions in sand and gravel (50-4000 \m) of a termite mound ( )an d its corresponding substratum (—) for profile Wan 1 A, 14-40 cm (a) and profile Wan 4, 95-110 cm (b).

0.449). From a comparison of median diameter values of sand in mounds and of sand in the adjacent soils, it follows that termites prefer finer sand (<1000 |om)tha n coarser sand (>1000 (jm)s o that continued termite activity will

-50M d

25Q 3

4000 2000 1000 500 250 Logscal eo fdiameter s(pm )o fsan dan dgrave l

Fig. 29. Cumulative curves of mass percentages of sand and gravel (50- 4000 |im)o f profile MIA, depth 14-40 cm (——) and its corresponding mound (---); of profile Wan 4, depth 95-110 cm ( ) and its corresponding mound ( ).Quartil e (Ql and Q3) and median diameter (Md) values are indicated with dots on the respective curves. -71-

shiftmedia n diameters of sand in topsoil to lower values until the distribution is to the "termites'taste " (Fig.30) . Such a grain-size distribution in sand of topsoils will only develop, when the fine grained material brought up by termites isno t removed by surface erosion. Itca n thenaccumulat e and become increasingly fine as termite activity continues.I f instead, fine grained material of themoun d is removed by surface erosion, thena coarse textured substratumwil l result, composed of grains which termites canno t transport,a s is probably the case inprofile s Wan 1an d Wan 2. Sand in such profiles iswel l sorted (log So= 0.15) because termites selectively removed the finerfractions . Although termites with subterranean nests make nomounds ,thei r effect on soil texturewil l be essentially the same as that ofmoun d building species, namely ane t transport of fine textured material touppe r horizons. This materialwil l become increasingly fine-textured unless coarse-san d isadde d to it from deeper layers or fromupslope . What slope transport means for composition of soils will be demonstrated in Section 7.2 on soils of theuppe r Wanjare catena.

600Hm(M)

400

OD D **

200

400 1000 1200}jm(S )

Fig.30 .Relatio n between the median diameter value of sand and gravel ina moun d (M)an d its corresponding substratum (S).Wa n 1( )wit h mound of Cubitermes;Wa n 1A (o)wit hmoun d of Pseudacanthotermes;Wa n 4 ( )an dWa n 5 ( )wit hmound s of Pseudacanthotermes and Odontotermes and trial field of D.va nde r Eijk (*) with a nest of Odontotermes. -72-

7.1.3.2 Texture profiles in poorly drained soils of theMagen a surface

Soils on the flat surface remnants are strongly differentiated (Planosols according toFAO-Unesco ,197 4an d abruptic Tropaqualfs according to Soil Taxonomy, Soil Survey Staff 1975). The profiles betweenmound s consist of a rather sandy Aan d A2horizo n overlying a dense clayey Bhorizon . In the B horizon a stone line occurs,consistin g of rounded alluvial gravelan d stones, not derived from,th eunderlyin g rock. The very good correlation of theminera l composition of sandwit h the underlying rock indicates,however , that the light textured material of the A andA 2 horizon ismostl y derived from the country rock.Ho w could this sand, which is derived from the local rock,occu r on top of an allochthonous B horizon and if pedoturbationwa s responsible for this situation,ho w could soils then be so strongly differentiated. The processes responsible for formation of this texture sequencewer e studied ina cross section through a termitemoun d and in adjacent soils.

Pedoturbation by termites. When comparing grain-size distribution of the soil in the centre of themoun d with that of theadjacen t profile (Fig.31) ,i t follows that termites thoroughly mixed material fromA , Ban d Chorizon ,

- H1B- MIC - H10 - • M1E - H3 T Y ! 1 HI clay EDsiU |: : ;I sand Depth (cm) m Depth (cm) 40- Up 80 111: > 20- wd

0 100 mass (%)

Fig. 31.Mas s fractions of sand, silt and clay in profiles of across - section through a termite mound of the Magena surface.MI A represents the centre of themoun d andM1 E the adjacent soil (M1B,MI C andMI D are intermediate in position betweenMI Aan d M1E).M 3 is situated in the plain betweenmounds . -73-

exceptmateria l coarser than 2mm ,whic h formed a layer at the transition to theweathere d rock. Horizon differentiation increases from the centre of the mound outwards (Fig. 31 Profile M1E). This phenomenon can heascribe d to the great structural instability and to themobilit y of thecla y in these soils, as follows from the followingobservations : (i) Inpit s dug in these Planosols,thic k clay flowswer e observed on the walls of the pit after only one shower,whil e grey sandy tosilt y material remained,wher e clay was leached. (ii) In termite-made cavities ina thin section of soilMI Dwel l developed composite cutans were observed (Fig.22) .Eve n ina chamber with live fungus combs, illuviated clay occurred. The formation of the texture sequence in thesePlanosol s can thus be explained asfollows : - Pedoturbation by termites enriches soilswit h sand, clay and some silt derived from theweathere d granite. - Horizon differentiationi s caused by illuviation of clay, so that theA and A2horizon s become enriched relatively with siltan d sand and the Bhorizo n with clay. The soils grade, therefore,fro ma homogeneou s soil in the centre of the mound,wher e bioturbation most strongly counteracts illuviation, toa soil where bioturbation is not strong enough tocounterac t illuviation.

Pedoturbation due to vertio properties. Thefollowin gthre ephenomen ai n Planosols were,however ,no t explained by bioturbation or clay illuviation: - The stone line in soils between the termitemound s undulatedwit h a position between the top of theweathere d rock and the top of the Bhorizon ,wherea s termiteswoul d leave it at apositio n justabov e theweathere d rock. - Material too coarse for termites to transport was sometimes found on top of the Bhorizon . - Thehorizon s overlying theB horizo n fluctuated indepth . Field studies showed that slickensides were well developed inon e direction.The ywer e parallelwit h theuppe r part of theB horizon ,wher e they connected à relatively lowwit h a relatively high position. Theston e line in theB2 t followed the same direction as theslickenside s andwa s sharply interrupted,wher e it reached the surface of the Bhorizo n (Fig.32) . The slickensides and the position of theston e line could be the result of differential pressure which developed after illuviation of clay in cracks -74-

Fig. 32.Schemati c cross-section through a pedon of aPlanoso l showing the effect of differential pressure and clay illuviation on the situa tion of slickensides (broken line), horizon boundaries (solid line)an d the stone line (000).

and successive wetting and expansion of the clay.Du e to this pressure,cla y with stones could have been pushed upwards.The y did not reach the surface, because claywa s probably washed downverticall y incrack s or laterally to the depressions in the Bhorizo n (Fig.32) .S o theupwar d movement of soil with stones and the successive leaching of clay,reasonabl y explains the presence of coarse textured material,includin g stones,directl y on top of theB horizon.

Bioturbation and vertlc turbation contributed topresen t soil characteristics.Erosio n of abandoned mounds and the start of newmound s at different locationswil lhav e occurred repeatedly.Thi s also follows from the observation of some young mounds during the field survey. Soth ewhol e surface willhav e beenaffecte d by bioturbation since its formation some 50000 0year s ago (Wielemakeran d vanDijk , 1981).

7.1.4 Conclusions

Results showed that termites in theprofile s studied, have the following effects ongrain-siz e distribution inwel l drained and poorly drained soils: (i)Remova l of fine grained material from deeper soilhorizon s and weathered rock,causin g a concentration of grains,whic h termites are not able to transport. This process results informatio n of a gravel or stone layer. -75-

(ii)Termite s have a preference for a certain grain-size,s o that their continued activity shifts themedia nvalue s of sand to that preferred value and, in so doing,san d becomes increasingly sorted. The thickness of the well sorted and gravel-free layer increases as termites keep on adding material from below the gravel layer. Incontras t with transport by termites,verti cmovemen t can bring material coarser than 2m m touppe r soilhorizons ,s o that itma y disturb the work of termites informin g horizonswithou t gravel.Thi s explains why coarse material (>2m m 0)i n large termitemound s on theMagen a surface was only found at the transition to theweathere d rock and insoil s betweenmounds , also in the B2thorizo n and on top of it.

7.2 LATERAL TRANSPORT PROCESSES

7.2.1 Introduation

In the preceding part of this chapter itwa s explained how termites form texture profiles. Soilmovemen t along the slopewa s mentioned, but its effects on formation of the texture profiles and the role of soilanimal s in the lateral displacement of soilmateria lwa s not explained. The situation on theWanjar e catena,wher e theuppe r part consists of granite and the lower part of diorite,i s ideal to estimate the effect of lateral displacement of soilmateria l onminera l composition of soils downslope. Atwha t rate lateralmovemen t took place can,however ,no t be deduced from theminera l composition of soils.Therefore , some measurements were done on both parts of the catena where different types of lateral transport occur.O n the shallow rocky upper part exposed soilmaterial s are moved by surface run-off. Termite heaps seem an important source of soil material exposed toerosion .Ther e isals oa smallare a fraction of bare soil around stones exposed toerosion . On the straight lateral slope of theWanjar e catena,wher e soils are well drained,permeabl ean d deep, the following types of transport were recognized: (i)Downslop e displacement of soil,whe n soilanimal s excavate soilan d deposit itmor e downslope thanupslop e from the opening. Atpresent , cultivation may also cause downslope displacement of soilmaterial , (ii)Splas h of exposed soil materials.Unde r a permanent vegetation, this -76- process only moved soilmaterial s exposed by soilanimal s or falling trees, but since cultivation of soils strongly intensified during this century, the process becamemuc h more important. (iii)Cree p (the result of small disturbances in the soilunde r the influence of shear stress acting upona slope)(Carso nan d Kirkby, 1972) . In the previous chapter itwa s shown thatmos t disturbances in thewel l drained deep soils are caused by the soil fauna,s o that theiractivit y will mainly determine the rate of creep. (iv)Run-of f ofwate r overver y short distances,i n the range of a fewmm s or cms. (v) Slumpsan d landslides were not seen inth eWanjar e catena,bu t were sometimes observed where slopes were undercut by springs at the side of valley-bottoms.

Particular items (i)an d (iii)wer e and stillar e important processes of transport on the deep well drained soils,s o that an estimate of these was made. Unfortunately no figures are available regarding theamoun t of soil raised to the surface byant s and subterranean termites.T ohav e an idea about the possible impact of this surface soil transport,mol e rat activity was estimated (see Section 2.7).Literatur e data show thatmole s can raise enormous quantities of soilt o the surface.Thi s amounted to 20 t.ha .a in the Luxembourg Ardennes (Imeson, 1976 )an d to5 5 t.ha .a in Ivory Coast (Abaturov, 1972).

7.2 Methods

Mineral composition of very fine sand in soils of the Wanjare catena. Detailed methods ofanalysi s are given inAppendi x 2.

Soil erosion from mounds. This requires anestimat e of the totalweigh t of soilaccumulate d inmounds .Tw o types ofmound s were observed: themoun d of Cubitermeswit h a semi-circular form and themoun d of Pseudacanthotermes with the form of a cone (Fig.33) .Th emas s of soil of aCubiterme s mound was calculated as follows:2/3i tr ^x (s.m.). With r= 0.28 m and s.m. = 900kg . m ,th emas s of soil permoun d amounts toA1. 4kg . Themas s of soilo f a Pseudacanthotermes mound was calculated as follows: 1/3 hnr x s.m. Ash is taken theaverag e height of two cones.Wit h r= 1.5 m, average h= 0.9 and -77-

— 0.28m 1.5m -

Fig.33 .Cross-sectio nthroug hheap so fCubiterme s(a )an dPseudacantho - termes(fo rexplanatio nse etext) .

s.m.= 100 0kg .m ,mas so fsoi lpe rmoun damount st o1/ 3x 0. 9m x %x (1. 5 2 —3 m) x100 0kg .m =212 0kg .Actua lerosio nfro ma Cubiterme smoun dwa s estimatedwit hstee lneedle sinserte da tsevera lplace sint oa mound .Lengt h ofth eneedl eabov eth esurfac eo fth emoun dwa smeasure dafte rinsertio nan d againafte ron eyear .Th edifferenc ei nlengt ho fth eneedl eabov eth esurfac e ofth emoun dwa suse da sa nestimat efo rth eannua lerosio nfro mit .

Surface erosion measurements. Thecontac to fsoi lwit hroc koutcrop so r boulderswa spainte da tsevera lplace si norde rt oestimat esoi lremova l aroundrock sdurin ga on eyea rmeasurin gperiod .Afte ron eyea rth ecleare d parto fth eroc kbelo wth epain twa smeasured .Additiona lrecorde ddat awere : slopea tth esit eo fobservation ,degre eo fsoi lcoverage ,are afractio no f stonesan dboulders .

Downslope displacements of soil by mole rats. Molera tactivit ywa smeasure d ongrazin glan dnea rMagomb odurin ga 5 da yperio di nth emont ho fMay .Th e masso fsoi ldeposite do nth esoi lsurface ,wa sair-drie dan dweighed .Th e fractionso fth etota lmas sdeposite ddownslop ewa sestimate dwit hth e followingformula : xslop e assumingtha tth emol ehea pha sth evolum eo f 180" halfa globe .Th eaverag edownslop edisplacemen to fsoi lfro mmol eheap swa s calculateda sfollow s(Fig .34) :OD E= \ surfaceo fOCA .G Hi srectangula rt o OFan dcut sD Ei nhalf ,O Hi sth eaverag edownslop edisplacemen to fth e fractiono fsoi l(OCA )o nslop etha tmove sdownslop edu et omol eactivity .I t doesno tinclud eerosio nb ysplash .

Creep. Steelneedle swer einserte dint oa vertica lwal lo fth esoi lpi ta t NARS(Chapte r 3). Theywer eplace di na vertica lro wa tregula rdistance sfro m theto po fth epi tdow nt oth eston elin ea tabou t3 m .Positio no fth eto po f -78-

Fig. 34.Downslop e displacement of soil due tomol e rats.

theneedle s was recordedwit h respect toa plummet;position swer e recorded againafte r two years.Th e difference between both measurements was a measure for soil creep.Th e bottom needlewa s used as a reference (no displacement).

7.2.3 Results

7.2.3.1 Slope transport andmedia n diameters of sand in theuppe rWanjar e catena

Fig. 35 shows how thedept h of soilwit hmedia n diameters of less than 100 (jm,increase s from 0a t deep augering DA2 tomor e than 3m at Profile Wan 3. Simultaneously themedia ndiameter s in theuppe r soilhorizon s decrease from nearly 100 \m at Da2 to less than 400 uma t DA7.Th ewel l sorted and coarse textured gravel layer is found at increasing depth below the surface downslope of profileWa n 2.It sdeepes t occurrence isa tWa n 3 justabov eDA2 . DA2 isapparentl y overharde r rock. As discussed inSectio n 7.1 finemateria l brought up by termites is removed byerosio n from Profiles Wan 1an d Wan 2an d moved downslope. During its transport downslope sand becomes increasingly fine (Fig. 35).I t indicates, that termites had sufficient time to sort soil materials during their transport downslope, so that lateralmovemen t was probably slow. It isals o possible that long ago lateral transport was much more important'an dtha t termites gradually sorted soilmaterial s after their deposition on top of the gravel layer.Whic h processes were at work,wil l be discussed below. -79-

1640

1620

1613 m Altitude (m)

Wan 1 Wan 2 DA1 DA2 DA3 DA6 DA5 DA4 Wan 3 DA7 DA8 T T T T T ï 1 1 J JUL

Fig. 35.Ma p of iso-mediandiamete r values of sand fractions(50 - 2000 um)o f theuppe r Wanjare catena (For situation of profiles,se e Fig. 7).Samplin g depths are indicated with dots.

7.2.3.2 Slope transport and mineral composition of sand

Because the rock upslope is granite and downslope diorite,th eminera l composition of soil materials downslope may indicate the importance of vertical and lateral transport processes.Th eminera l composition of soil materials (Table 14an d Fig.36 )give su s the following information about lateral transport of soilmaterial s fromgranite : (i)Onl y samples from theuppe r metre of profiles downslope contained a detectable number ofmineral s derived from granite. Itmean s that slope Han 1 ! -80- 1645 \wan 1a 1640 • \ Wan 2 oVJ DA1 50i 100: DA2 • 30~~- ^l^ DA3 200 "^ 2.5 ^ ï^_ DA6 N 26—^i DA5 1620 • DA4 400- \ ' T^ r—-^» Wan 3 0.4 DA7 \ 0.4 ^J^^ Al tit jde (m) 0.3 0.7 0.2 0.4 0.5 \ 0.6 50 0.4 0.3 v 0.1 100 ^3 0.4 0.4 S 0.2 S 0.2 0.4 / 0.5 I 0.7 100m

400 (cm)

Fig. 36.Mas s percentage of strongly magneticmineral s (mainly magnetite)i n sand (50-420 |im)wit h a density > 2.89 in samples from the upperWanjar e catena.DA = deep augering. Scale of altitude is 0.25 x depth scale and 4x horizonta l scale.

Table 14.Sem iquantitativ e estimates ofmas sfraction s ofheav ymineral si n sand (50-420um )o fprofile san ddee paugering sfro m theuppe rWanjar ecaten a (Fig.) .Mas sfractions : II I H >0.9;-H-++ ,0. 6 -0.9 ;+++ ,0. 3 -0.6 ;++ ,0.1 5 - 0.3;+ ,0.0 5 -0.15 ;tr ,<0.0 5an d- ,no tdetectable .

Profile Depth Ilmenite Zircon Rutile Anatase Magnetite Haematite Amphibole (cm)

Wan1 0- 35 + + + Wan2 0- 16 + + + DA1 0- 15 + + + + 60- 65 + + + + DA2 0- 15 -H-H- + + + tr 60- 65 + + tr DA3 0- 15 +++ + + + 60- 65 + tr tr tr + DA6 0-1 5 tr tr tr + 165-170 tr tr 245-250 tr tr 340-355 tr tr tr DA5 0-1 5 tr tr 120-125 tr tr 245-250 tr tr 440-445 tr tr DA4 0-1 5 + tr tr 120-125 + tr tr 245-250 + tr tr 390-395 ++ tr tr Wan3 10-2 5 tr tr 315-330 tr tr saprolite tr DA7 0-1 5 + tr tr 160-175 + tr -81-

transport occurs and occurred especially in theuppe rmetr e of the profile, (ii)Th e number of stable minerals from granite such as rutile and anatase decreased less inprofile s downslope than the less stablemineral s magnetite and haematite (cf.Chapte r 4).Weatherabl emineral s such as hornblende were not present inprofile s further downslope,bu t therati o of ilmenite to zircon in Profile Wan 3showe d that soilmateria l from granite was present in the uppermetr e of Profile Wan 3 (cf.Chapte r 4). The selective weathering of heavymineral s (Table 14)durin g transport downslope shows that lateral transport was slow. (iii)Th e decreasing number of stablemineral s from granite in theuppe r metre of profiles downslope indicates that soil materials from granite are gradually mixed with othermaterial s during transport downslope.Result s of studies discussed in Chapter 6showe d that termites transport soilmateria l from the uppermete r to thedee p subsoilan d vice-versa.Thi smechanis m may explain the gradual decrease of soil materials from granite due tomixin g during their transport downslope.

7.2.3.3 Downslope displacement of soilmaterial s from the rocky upper part of the catena

Mass of soil accumulated in mounds. Ona surfac eo f40 0m 2,1 0mound so f Cubitermes and 2mound s of Pseudacanthotermes were counted. Mass of soil per mound is 41.4 and 2120k g respectively (see Section 7.2.2), so that themas s of soilaccumulate d inmound s equals 10344kg .h a for Cubitermes and 106.000 kg.h a for Pseudacanthotermes, totalling 116.3 t.h a .

Annual erosion from mounds. According to measurements with needles,1. 5 cm of soilmateria l was eroded from the lateral side of aCubiterme s mound in one year. This would mean that a mass fraction of 0.15 was eroded from the whole mound. This seemed an overestimation, because part of themoun d had a less steep slope;0. 1 seemed therefore amor e reasonable estimate for the mass fraction of soilremove d from themoun d during oneyear . Lee and Wood (1971b)amon g others,estimate d that dead mounds eroded in 5 to 10years .The y observed that dead mounds usually numbermor e than 1ou t of 5. Arati o of 1dea d mound in 9liv emound s in theare a of study seemed therefore a safe estimate.Assumin g that eroded material from livemound s is reused by termites and that theannua l eroded mass fraction from dead mounds -82- is 0.1 times theirmaximu mweight , then the dwonslope erosionwil l amount to 0.01 x 116 300kg .h a .a = 1163kg .h a .a .I na n equilibrium situation thiswoul d probably be theminimu m amount of soil that leaves theuppe r slope each year.Wit h a catchment of 0.545h a per 100m contourline,11.6 3k g x 0.545= 6.34 kg of soilpe rmetr e contourline perannu mwoul d beadde d to the lower slope. Nye (1955)estimate d theamoun t of soilerode d fromhi s gently sloping upper catena was one third of the total amount of soilbrough t to the surface by termites.Hi s estimate for theerode d amountwa s 5000 t.h a in 12000 years or 417 kg.ha - .a- .

Removal of soil material due to run-off from roek outcrops and boulders. Real depth of the soil layer removed by run-off ranged from 0.2-1 cmaroun d rocks on a slope of 10-20% to 0.5-3 cm on thesteepe r slopes.Erosio nwa s negligible if the soil bordering the rockwa swel l covered with grass.Th e measurements were representative fora small fringe of about 10c mwid e around the rocks O _1 with a surface area estimated at about 500m .ha .O na hectar e basis about 0.5-1 mm.a could have been displaced by run-off from rocks.Thi s soilmas s equals 5000-10000 kg.ha -1.-1 (s.m.= 1000kg . m-3). Ifal l this soil material left theuppe r slope itwoul d meana downslope addition of 27.3 to 54.5kg . a~ permetr e of contourline. Which fractionactuall y reaches the footslope,wa s not estimated. Imeson (1976)calculate d that on a wooded slope of 17" in the Luxembourg Ardennes70 - 80% of 20000k g of soilraise d to thesurfac e bymoles ,move d downslope.Th e rate atwhic h this material left the slope was estimated at 0.18 kg per metre of riverbank per annum or amas s fraction of 0.001. If such amas s fraction applied to the soil displaced by run-off,0.027-0.05 4 kg.a per metre of contourline would leave theuppe r slope.Becaus e of higher and more intensive rainfall in theKisi i area the figure cana t least be doubled.

7.2.3.4 Lateral transport of soilmaterial s in deepwel l drained soils

Downslope displacement of soil by mole rats. Estimateso fdisplacemen t bymol e ratactivit y amounted to 7k g air dry soil per 100m per 5day s or 50 t. ha .a ,assumin g that the 5da y period was representative for theyea r (see Section 2.7).O na slope of 12%, -r^L- x5 0 t.ha _1.a-:l = 3900 kg. ha-1.a-1 lou -83- wouldmov edownslop eove ra distanc eo f7 cm .I tmean stha ta soi llaye rwit h anaverag ethicknes so f0.0 4m mmove s7 c mdownslop eannually .

Creep, inFig .3 7thre ezone sca nb edistinguished : (1)Fro m0-70/8 0cm :zon ewit hgrea tvariatio ni nobserve ddisplacement , coincidingwit hth ezon eo fmajo rbiologica lactivity .Th eyearl ycree pwit h respectt oth ereferenc eleve la t30 0c mdept hwa s 1-4m mfo rth euppe r3 0c m and 1-1.5m mfro m30-70/8 0c mdepth . (2)70/80-20 0c mdepth .Th eyearl ycree pwit hrespec tt oth ereferenc eleve l at30 0c mdept hwa s0.5- 1mm . (3)200-30 0c mdepth .Th eyearl ycree pdecrease dfro m0.5- 1m ma t20 0c mdept h to0 a t30 0c mdepth . Thedisplacemen ti nzon e3 wa sbigge rtha ni nzon e2 .Thi scoul db edu e toa lowe rshea rstrengt hi nzon e3 tha ni nzon e2 an dals ot omor eshrinkin g andswellin ga sa resul to falternatin gdr yan dwe tcondition si nthi szon e (seemottlin gdat ai nth eNAR Sprofile ,App . 1).Micromorphologica lanalysi s

300 - Fig. 37. Downslope displacement of needles in the well drained profile upslope dounslope at NARS during a two year period. -84- of thin sections fromwel l drained soilhorizon s (Chapter 6)showe d that soil fauna seemed responsible formos t disturbances in thesehorizons . The displacement due only to the soil fauna could thenamoun t toa maximum of 3m per 1000year s in the upper 30cm , to0.5m/100 0year s from30 - 70/80 cm, to 0-0.5 m/1000year s from 70/80-200 cman d 0m/100 0year s below 200 cmdepth .

7.2.4 Discussion and conclusions

Results of studiesma y be represented by the following model (Fig.38) . 1°. Burrowing activity of termites inshallo w soils on steep slopes of the Wanjare catena is themos t important source of erodible soilmaterials .

2°. Soil materials eroded from the granitic upper part of the catena,ar e fed to theuppe r horizons of deeper soils at its food,wher e processes of creep dominate.

3°. Granitic soilmaterial s with agrain-siz e favorable for transport by termites remain concentrated in the upper meter of the deep profiles due to a high transport activity of soil fauna,compare d with a lowactivit y below1 metre. Incontrast ,materia l too coarse for transport by termites gradually sinks below the layer ofmai n activity, so that thatmateria l moves downslope very slowly.

4°. Decrease of the percentage of soilmateria l from granite during transport downslope is due to a gradual admixture with soil from diorite derived from deeper layers and due toa slow but gradual transport of soilmateria l from theuppe r metre to the part of the profile below 1m . This effect cumulates downslope in themixin g of soilmaterial s from granite throughout theprofile . Soilmaterial s from granite are then,however ,s o diluted that their minerals are no longer detectable. In fact, thedifferenc e invertica l distribution ofmaterial s from granite inprofile s upslope and downslope reflects the time of mixing. A similar difference due to time ofmixin gwa s observed between the distribution of glass fragments and magnetite-IIminerals .Glas s fragments representing relatively recent additions ofvolcani c ashwer e concentrated mainly in the upper metre of the profile,whil e themagnetite-I I minerals representing much -85-

.surface ^creep_ transport

Fig. 38.Arrow s indicate schematically transport activity in deep well drained soils on a slope (a)an d its cumulative effect on thedistri bution of stable (x)an d less stable (o)mineral s inver y fine sand from upslope in the soils downslope(b) .

older additions ofvolcani c ash,wer e rather evenly distributed throughout the profile.

5°. Lateral movement of soilmateria l is soslo w that even rather stable minerals selectively weathered. This excludes an important contribution of surface transport because itwoul d take only 3500year s for soilmaterial s moving with a speed of 7c mannuall y to reach Profile Wan 3;wit h that speed minerals would not have beenweathered . Due tomixin g soilparticle s will, however,no t remain in the surface layer. They will in fact move continuously from one layer to the next.Ho w fast aparticl e moves downslope will depend on the fraction of time it resides in the various layers. Because the fast moving surface layer is only 0.04 mm thick, soil particles willmainl y reside in the soil layer below it,wher e processes of creep dominate. If estimates of transport due tocree pappl y to theWanjar e catena, it would take 80000year s for soil material in theuppe r 30c m to reach Profile -86-

Wan 3an d 50000 0year s for soilmaterial s at a depth of 30-90c m toreac hit . Soilmateria l below that depthmove s at an even slower speed. Because rather stablemineral s disappeared during transport downslope, the estimated speeds seem rather high,whic h is probably because the slope of the Wanjare catena is more gentle than the slopewher e creepwa s measured. Theabsenc e of a detectable number of minerals from granite in soil materials deeper than 1metr e gives no indication that lateral transport was different in thepast .

6°. Continued reworking of soilmaterial s by termites shifts coarse median diameters of sand to lower values until they are "to their taste" (Section 7.1 Fig. 30).Latera lmovemen t in the deep soils is controlled by termites,s o that this effect of termites on grain-size distribution isals onoteworth y in a downslope direction (Fig.35) .

7.3 THE FORMATION OF TEXTURE PROFILES WITH STONE LINES

7.3.1 Prevailing theories

Even today a number of authors claim that all texture profiles with stone lines are the result of geogenetic processes. In their view the stone line is an erosion rest onwhic h a sediment without gravelan d stones is deposited. These authors seem unaware or canno t imagine that pedogenetic processes can form such profiles.S o they ignore thevie w ofman y soil scientists and biologists,wh o long ago showed that soilanimals ,especiall y termites,ca n form texture profiles with stone lines (Grasse, 1950;Nye , 1955;d e Heinzelin, 1955; Boyer, 1956a;Watson , 1960;d e Ploey, 1964;Levèque , 1969an d Moeyersons, 1978). The present study, ina nare awit h ahumi d to subhumid climate,confirm s the latter point of view. That also other processes can result in formation of similar profiles with a stone line,i s not ignored by the author even for areas with a similar climate. Some authors who believe that texture profiles with a stone line result from geogenetic processes only,stat e that they are formed ina semi-arid environment (see for instance Ab'Saber, 1982an d Bigarella et al. 1965). The bio-pedogenetic theory showed us that a dry climate isno t necessary for formation of such texture profiles.Thi s has important implications for -87- landscape development in the past. In the discussion below the geogenetic theory of formation of the texture profiles will be tested to see if the results from the area of study could also be explained by this theory.

The following conditions in theWanjar e catena,ar e not in linewit h this theory: - The absence of a lithologie discontinuity. This feature is one of the arguments Ruhe (1959)an d Fölster (1969)use d toprov e their theory of formation of the texture sequence,bu t soilmateria l directly above and below the stone line in the Wanjare catena,wa s not different. - The presence of a 3metr e deep soil cover on flat surface remnants of a dissected Tertiary erosion surface.Ho wcoul d soilmateria l removed by erosion from the flat surface have been replaced since the dissection of that surface during and after the Tertiary? Without replenishment of soil materials through upward transport from below theston e line,gradua l erosion of soilmaterial s would have left the stone line at the surface,whil e it isno w covered with threemete r ofwel l sorted fine earth.Furthermore , soilmaterial s would not have beenmixe d if noupwar d transport from below the stone lineha d taken place.

7.3.2 Other theories

When a desert pavement,a pediment ora coarse textured river alluvium gets covered by finer wind or water deposits,i tma y appear like a stone line sequence, particularly after sediments areweathered . The stone line on the Pleistocene Magena surface (Wielemaker and van Dijk, 1981)ma y have formed mainly as described above.Th e studies showed, however, that the fine earth above the stone line originated not only from a deposit on the stone line (volcanic ash)bu t also from theweathere d rock below the stone line.

Sharpe (1938)an d Ireland et al. (1939)explai n stone lines from the incorporation of coarse fragments in the subsurface by soil creep. The fragments would be detached from dikes or other resistant layers.Thei r theory does not explain adequately how the coarse material detached from dikes, becomes concentrated ina stone line,no r does it explain how stone lines are formed in soils on top of a ridge. -88-

Termites,whic h are responsible for vertical transport, fill the gap in this theory.

Laporte (1962)postulates ,tha t stonesma y sink in theprofil e due to physical processes.Thes e processes may work inclay s with ahig h plasticity when saturated withwater ,bu t not inwel l drained red soils,whic h are seldom saturated with water and have a low plasticity (Collinet,1969 ;Moeyersons , 1978). Inpoorl y drained soils in theare a of study stones did not sink,bu t were instead pushed upwards due toverti c properties (Section 7.1.3).

None of the theories discussed above gives anadequat e explanation for the formation of texture profiles with stone lines in the area of study. Neither the formation of thedee p layer ofwel lsorte d fine earth above the stone line,no r the stone line itself can be properly explained without the action of termites.I t implies that they played an essential role in processes of landscape and soil formation in theare a of study. Because termites occur also extensively inSout hAmerica , theywil lhav e also formed stone lines with fine material on top of them. So,thei r formation will nothav e been restricted to the dry climatic periods as postulated by Ab'Saber (1982)an d Bigarella et al. (1965). Deeply weathered soilswit h stone lines in Brazil are thus not necessarily the result of repeated erosion and sedimentation; their formation can be due to pedogenetic processes only. Such a view is also adhered to by Bennema (1965), who considers differences indept h and stage of weathering of Brazilian soils mainly the result of differences in time of soil development modified by conditions of topography, parentmateria l and drainage. 8 Termites and land-use

8.1 INTRODUCTION

In the preceding chapters itwa s shownho w soilfauna ,particularl y the termites,affecte d well drained soil profiles in theare a of study so thoroughly thatvirtuall y theirwhol e geometry was controlled by them, even in soilmaterial s at depths of 10m andmore . Itwa s shown thatminera l composition and grain-size distribution were controlled by termites and that they contributed to the formation of a soil mantle of at least 3m thick on these surfaces,whic hwer eTertiar y or older, and that they transformed an important part of saprolite materials below a depth of 3m . Deposition and transformation of soilmaterial s by termites occurred in all soilenvironment s studied, but onstee p slopes erosionwa s sostrong , that no effective accumulation of soilmaterial s could take place and, under poorly drained conditions,soi l materials outside the nest area rapidly lost their architecture made by termites.S o it seems that for aneffectiv e stabilisation of the termitehabita t resulting inaccumulatio n of soilmaterial s with a homogeneous and porous architecture,processe s of déstabilisation(Chapte r 6) such as erosionan d structure decay,shoul d not be too strong. Such interaction between termites and their environment willb e demonstrated in a flowchar t (Fig.39 )i n the next section. The chart is representative of thestud y area,whic h isa hig h altitude humid tropical region,enriche d with volcanic ash. Termite species there do not damage the crops and their activity seems positively related with soil conditions for plant growth. Inothe r areas of the tropics termites areno t always considered so benifical for agriculture. Some species pose problems for mechanized agriculture such asmoun d building termites and others as for instance the Hodotermitinae or harvester termites,ar e considered a pestbecaus e they eat the standing crop (Coaton, 1958). As regards soilconditions ,lo w humus content and lowwate rholdin g capacity ofOxisol s incertai n tropical regions -90- are ascribed by some people to the predominance of termites. When discussing significance of termites formaintenanc e of production levels inagricultur e such possible harmful effects will also be considered. The chapter concludes with some recommendations for future research.

8.2 TERMITES AND NATURAL ECOSYSTEMS

8.2.1 Inflow and outflow of energy and matter

What effect termites have on flows of energy and matter in natural ecosystems is demonstrated in Fig.39 .Th e outer circle represents the flow so fara s it is controlled by termites.Th e system is not closed, however; there isa n energy and material exchange with a larger ecosystem ofwhic h it forms a part: innatura lecosystems ,rain ,sunshine ,carbondioxide ,atmospheri c dust and colluvial oralluvia l deposits form an incoming flowo f energy and matter. Rainan d sunshine alsoaffec t the rate of decay of structure, the rateo f erosion of soil materials and the leaching rate of nutrients.Ho w fast this occurs depends on land characteristics, sucha s climate, type of soil materials, their drainage and geomorphic position,bu tno t in the least on the effect termite activity has onmodificatio n of such land characteristics (Fig. 39).

Chapters 6an d 7 showed that thehig h intake rates inwel l drained soils in theare a of study are due to termite activity and that also the rate of creep is controlled by theiractivity .Eve ni n theshallo w soils,rat eo f soil transport was regulated by them.Lepag e (1981)calculate d that foraging holes made by termites during one year ina dry savanna inKeny a occupy anare a of 2 —2 2-4m .km ,whic h illustrates their impact on intake rate and surface-runoff ina muc h drier area than theon e studied. So, termites largely control inflow and outflow of water and soil material,wit h impact ona much wider part of the environment thanwher e they are active.Withou t theiractivit y rates of erosionwil l strongly increase causing changes inhydrolog y which may result in downstream flodding and the silting up of reservoirs.

8.2.2 Suitability of the habitat for termites

Soils in theKisi i area have usually ahig h ironcontent .Th e iron,i f -91-

excretion rate

uptakerat e bytermite s animal biomass

esting \/l

macro-climate; leaching — - inputsan dmanage rate mentb yma n

Fig. 39.Schemati c presentationo fth e roleo ftermite si nagro - ecosystemsi nth eare ao fstud y accordingt oForester' s notations (Forester, 1961).

oxydized, can give soilsa hig h structural stability and thati sa conditio n fromwhic h termites profit becausei timprove s the qualityo fthei r building materials.S olan d characteristics, including climate determinet oa-larg e degree how suitablea sit ei sfo r termites,becaus ei tdecide st o what extent they can adapt conditionst othei r needs.Ho wwel l conditions are adaptedb y the termites,determine s the suitabilityo fthei rhabitat .I tplay sa centra l rolei nregulatin g food uptake for growth, for labour and formaintenanc eo f life functions (see Section5.2) .

Thewor k termitesd oi nsoil s and the excreta they produce,ma y further stabilize soilmaterial ss otha thabita t conditions are improved.I fi n addition, their activity raises the suitabilityo fsoil s for growtho fplants , biomass production will increase.Thi s againha sa positiv e influenceo n -92- habitat suitability,becaus e it improves the climate for termites and increases supply of dead organicmaterial , either directly as litter from plants or indirectly as faeces from grazing animals. More biomasswil l also decrease rates of loss tosoi lmaterials ,soi l structure and nutrients,becaus e itsupplie s a better soilcove r and usually exploits water and nutrients better.Thi s has via soil conditions a positive influence onhabita t suitability because it reduces the amount of work termites need todo , tomaintai n proper nestan d channelsystems . Sotermit e activity in soils has twoeffect s on the suitability of their habitat. - It improves theirhabita t conditions,s o that lesswor k isneede d for maintenance and repair and more energy remains for reproduction and growth. - It increases theamoun t of dead organicmaterial , fromwhic h both termites and other litter feeders profit. The effect of termiteactivit y on soil conditions and growth of plant biomass seemsmor e complicated than suggested in the flowchart . Termites probably create appropriate physical conditions for other soil dwelling organisms and so stimulate their functions in soils.Whe n litter is in ample supply aswa smos t likely the case in theKisi i area,a ver y diversified soil flora and faunawil l exist not only of termites,bu t also of other soil organisms. On the other handwhe n litter is inshor t supply,fo r instance in regions with a dry climate,i n poor soils or inovergraze d land, certain termite speciesma y still get enough food, whilemos t other litter feeders starve. Buxton (1981)wh o studied the relation between species diversity among termites and habitat productivity found that,als o among termites,diversit y declines if rainfall and habitat productivity decrease. Soa decreasing food availabilitymean s a decline infunction s of the soil flora and fauna,particularl y when termites are the dominant litter feeders. It is possible that people refer particularly to such situations when they consider termites harmful for soil conditions. Itsuggest s that for maintenance of proper soilcondition s for plant growth adiversifie d soil flora and fauna is necessary. The interaction of termites with other litter feeders is still poorly understood so that for the sake of simplicity only a direct relation between amount of dead organicmaterial , termite activity and soil suitability has been considered in the flowchart . -93- Table 15. Effects of termite activity on soil formation, on soil properties and on suitability of soils for plant production, in the area of study.

Effects on soil formation Effects on soil properties Effects on plant production. A. Beneficial effects: Formation of pores (0 > Increase in permeability Increase in rootabillty 1 mm ), lining and back for water and air and in and potential for moisture filling of channels and structural stability. uptake by plants. chambers with salivated Decrease in bulk density. - Increase in oxygen pellets of earth. availability. Decrease in risk of water Mixing of soil materials Formation of homogeneous stagnation, acid formation soil horizons without and shortage of oxygen. gravel and with diffuse transitions

Transformation of saprolite Increase in soil depth Decrease in risk of erosion. into soil materials Rapid decomposition of If supply of litter is More available nutrients dead organic material small,termites may cause through 1. decomposition of including wood. depletion of soil organic of material rich in lignin Concentration of bases matter so that structural and cellulose. 2. upward in organic matter. stability declines and transport of less weathered Formation of humus, rich poresystems become less soil materials. 3. creation in bases. heterogeneous (see text). of appropiate physical conditions for soil flora and fauna. Adverse effects: If litter is in short supply, termites may cause a rapid decline in functions of soil flora and fauna (see text).

What effect termite activity has on soil characteristics was demonstrated in the preceding chapters and is summarized in Table 15, mentioning also what it means for plant production. Not all soils in the area of study were so effectively modified. Table 16 shows which land conditions limited suitability of the habitat for termites and what this means for their effect on soil characteristics: The geometry of the Paleudolls with little or no limitation for termites was strongly controlled by termite activity. The geometry of Haplohumox did not reflect such a high actual termite activity as the Paleudolls did (Chapter 6). It could be that moisture is sometimes limiting, due to the rather hard and impermeable rocks, underlying these soils. Also their position on flat topped ridges may limit moisture availability. Food and nutrient availability in these infertile soils could be limiting causing insufficient functioning of the soil fauna, as explained in Section 8.2.2. The lithlc Humitropepts have limitations for subterranean termites, because soils are too shallow for nesting, while moisture may be sometimes limiting. No accumulation of soil materials can take place due to erosion on those steep slopes (Chapter 7.2). The poorly drained abruptic Tropaqualfs pose limitations for termites -94- Table 16. Factors limiting habitat suitability for termites in the area of study and a ranking of soils according to their suitability as a habitat for termites.

Soil type Limiting factors Ranking of habitat suitability Well drained, deep, No limitations. Availability of 1 stable Paleudolls moisture and of suitable temperature (Wan 3, NARS) fluctuate with climate.

Moderately well Slight limitation due to rather low 2 drained, deep, rather stability of building materials, stable Paleudolls. Climate as above (Wan 4, Wan 5)

Well drained, deep, Slight limitation due to low 2 stable Haplohumox. nutrient availability and moisture (Wan 9, 210) (during the dry season). Climate as above.

Somewhat excessively Severe limitation of soil depth for drained, shallow over termites with subterranean nests due hard rock on steep to high risk of erosion. Moisture may slopes. Lithic be limiting during dry season. Humitropept (Wan 1) Climate as above.

Poorly drained, deep Severe limitation due to high risk unstable abruptic of flooding and very low stability Tropaqualfs. of building materials. (Opapo, Wan 7).

because soil materials are unstable so that structures made by termites rapidly decay. Waterlogging and flooding is another limiting factor.

8.3 TERMITES AND AGRO-ECOSYSTEMS

How do agricultural activities affect the flow chart presented in Fig. 39 and what does it mean for the maintenance of adequate soil conditions for plant growth? To explain this, their separate effects on supply of dead organic material, on plant biomass and on soil characteristics will be discussed first. Then a discussion follows of their impact on the termites' habitat and the role of termites in soil. What this implies for land-use is dicussed in the last part of this section.

8.3.1 Effects on supply of dead organic material

The biomass growth rate. Rate of biomass production depends largely on the type of produce and how the farmer manages it. So crops like sugar-cane have a -95-

muchhighe r biomass growth rate thana cro p like pyrethrum.A farme r also raises thegrowt h ratewhe nh efertilize shi slan dan dwhe nh econtrol s diseasesan dpests .Dependin g onth eamoun th eharvests ,h ema ys oincreas e amount ofdea d organicmateria lfo rsoi l organisms comparedwit h natural systems.

Proportion of the plant biomass, which, is harvested. Howlarg ethi sproportio n is, dependso nth etyp eo fcro pan dth evariet y ofcro p grown. Through breeding,varietie s witha hig h harvestable fractionar eselected , sotha t crops optimally convert available energyan dresource s inharvestabl e biomass. Growtho fsuc h crops leaves little food forsoi lorganisms .Th eharvestabl e fraction depends very mucho nth etyp eo fproduce .S ofrui t trees never achieve ashig ha fractio na sfo rinstanc e graincrops ,no rd ograzin g systems.I nfact ,grazin g stimulates increasei nbiomas san di fgras si s fertilized, more dead organicmateria l becomes available thano nungraze dan d unfertilized grassland. Insummary ,crop s producing asmal lamoun to fbiomas san dcrop swit ha highharvestabl e fraction,decreas e amount oflitter ,compare d with natural systems.S othe y adversely affect thesuitabilit y ofth etermites 'habitat , either directly orindirectl y (Section 8.2.2).

8.3.2 Effects on plant biomass cover

As shown inFig .3 9plan t biomass alsoha sdirec t effectso nth e suitability ofth etermites 'habita t ifi taffect s soilclimate . Natural systems tend togiv e soilsa permanen t vegetative cover.Unde r agricultureth e soili ssometime s bare,particularl y whenannua l cropsar egrow ni nregion s witha pronounce d dryseason .

8.3.3 Effects on soil characteristics

Bates'of deterioration of geometry and of erosion. Apartfro mlan d characteristics (Section 8.2), rates depend among otherso nth etyp eo fcro p ando nho wi tprotect s thesoi l against rain impact.Us eo fheav y machinery and frequent tillagear epractice s whichma yfurthe r increase theris ko f erosionan dth eris ko fdeterioratio no fgeometr y (Chapter 6). Ifhorizo n -96- differentiationan d compactionar e not sufficiently counteracted by formation of root channels,b yactivit y of macrofauna,b y swellingan d shrinking or by deep ploughing, soilsma y degrade physically and become susceptible to erosion. Rate of teaching. Leaching dependsmainl y on theamoun t ofwate r passing below the root zonean d on the amount of nutrients it carries. Rainfall distribution, uptake of nutrients andwate r by plants in relation to nutrient and water supply,wate ruptak e and storage by the soilar e important rate determining factors in this respect. Beside indirect effects,agricultura l activities also have direct effects on the suitability of thehabita t for termites:tillage ,particularl y with a plough,ma y disturb nest and channel structures and it is likey that certain specieswil l disappear (Section 5.2) if they experience too muchstress .Th e use of pesticidesma y also bea stress for them. As explained in Section 8.2,non e of the effects stands alone:a more rapid deterioration of soilphysica l conditions decreases thesuitabilit y of soil for growth of plantsan d that,i n turn,decrease s the amount of litter available for soilorganisms ,s o that species diversity becomes lesswit h a consequent decline in their benificial effect on soilcondition s for plant growth (Section 8.2.2).

8.3.4 Farming and termite activity

Within the context of thesocio-economi c environment, farmers have the following options to maintain or stimulate beneficial effects of termite activity (or soil fauna in general): - choose crops or cropping systems,whic hminimiz e rates of loss tostructure , soil materialan d nutrients. - choose crops or cropping systems,whic h are least susceptible to termite attack (Mielke, 1978mention s inhi s paper,tha t indigenous African crops were hardly susceptible to termiteattack ,whil e introduced crops or varieties were quite susceptible). - suppress harmful termite species and promote beneficial ones. - choose crops with a low proportion ofharvestabl e biomass (forinstanc e a grazing system). - increase food available for thesoi l fauna (a)throug h application of -97-

- increase food available for the soil fauna (a)throug h application of fertilizer,mulch ,manur e or a combination of the three, (b)pes t and disease control in crops,withou t damaging termites. - apply cropping and tillage systems,whic hminimiz e the stress for termites. - alleviate stress for termites by changing land conditions as e.g. soil conservation and drainage measures. (Thenumerou s transitions which exist between the poorly drained unstable soils and thewel l drained stable soils in theare a of study, show that a slight improvement in drainage improves habitat conditions for termites,becaus e the stability of the soilfabri c increases due to theoxidatio n of iron.Pore s dono t collapse so easily so that the permeability increases favoring processes of weathering and leaching. Consequent changes incla ymineralog y further stabilize soilmaterials . In fact, it isno t only the work of termites thathelp s tostabiliz e their environment. They unlock the potential for stability that these iron rich clays have under well drained conditions).

8.4 SIGNIFICANCE OFTERMITE S INHIG HAN D LOW INPUT AGRICULTURE

Isi t possible tomaintai n production levels in agriculture without the beneficial effects of faunal activity asmentione d inTabl e 15?Whe n answering that question low and high input agriculture can best be separated, because low input agriculture has fewer means to improve land conditions thanhig h input agriculture. So the direct effects of termites on nutrient availability inhig h input agriculture are of no importance, in contrast to low input agriculture,wher e this may be important. More difficult and costly is the compensation of loss topor e space and structure.Ploughin g may improve structure and porosity in the topsoll but in some situations it can hardly prevent horizon differentiation, compaction and eventualwate r stagnation in the subsoil.Ensuin g erosion and shortage of water and oxygen for plants may toa certain extent be controlled by conservationmeasure s and soil management practices. Ifa t all possible,i twil l still raise costs of production, particularly in situations where soilmaterial s are susceptible toerosion . Further to this, intrinsic characteristics of soil materials cause great differences instabilit y of soilmaterials . So thewel l drained iron-rich kaolinitic soils in theare a of study are physically stable,bu t in certain areas of the tropics kaolinitic soils have a low iron content,whic h are very -98- susceptible tophysica l degradation,whe nheav y machinery isuse d (Bennema, 1982). Also the imperfectly to poorly drained soils in theare a of study are very susceptible tohorizo n differentiation, loss of structure and compaction. Insummary , sustained high input agriculture may see its profits greatly reduced if itha s tod owithou t effect of activities of fauna on physical properties asmentione d in Table 15,particularl y if soils have unstable structures and are susceptible tostructur e decay and erosion.

How is the situation with respect to low input agriculture, probably the most common type of agriculture in the tropics? Under this type of agriculture the farmer usually lacks the means to invest in costly management and conservation practices.When ,formerly ,h ewa s confronted with a decline in fertility of his land,h e could shift toa different piece of land. Inthi s practice of shifting cultivation he used theabilit y of natural ecosystems to restore,a t least partly, the fertility of his land. Due to the increase inpopulatio n pressure and the reduction infar m size,shiftin g cultivation ishardl y practical to restore production capacity of land.H e should now try tomanag e his soilresource s and the imcoming energy in such awa y thath e keeps the production capacity of his land as high as possible and stillget s enough fromhi s land. Toachiev e this,h e should derive the maximum benefit from faunal activity or from soil biological processes ingeneral .Whic h strategy a farmer can follow to stimulate faunal activity in soils was explained in Section 8.3. As explained in the introduction, theutilit y of termite or other fauna species for agriculture differs widely. Toderiv e thehighes t benefit from the activity of soil fauna, their damaging effects should be curtailed and their beneficial effects promoted. The introduction of earthworms in polders of the Netherlands (Hoogerkamp et al. 1983)show s that active manipulation can be quite successful. The introduction of social insects such as termites and ants as soil ameliorators seems not to be described so far. Theefficienc y of, particularly, the fungus growing termites inutilizin g low quality food resources, pleads for amor e efficient use of those insects for maintenance of production capacity inagro-ecosystems . Because they also convert thesam e low quality food resources in food with ahig h quality for human consumption, in the form of flying termites or alates (Mielke, 1978), theyma y be important as -99-

secondary producers and so raise thevalu e of the plant biomass production, which canno t be harvested.

8.5 RESEARCH NEEDS

This chapter did not give a clearcut answer to thequestion ,whic h part of theannua l biomass production should be directed tomaintenanc e and repair functions of the ecosystem. Neither did it give an answer towhic h farming practices should be adopted. There are many variables involved which have been rarely investigated inagricultur e and are still poorly understood by ecologists (Swift, 1984). Itemphasize s thenee d for systematic knowledge regarding the ecological requirements and the role of various soil fauna,particularl y in thetropics . Inaddition , information is required, regarding the effects of farming practices on the interaction between soils and soil fauna and regarding beneficial orharmfu l effects oncro p production.Comparativ e studies between agricultural and natural ecosystems could fill this gap in our present knowledge. The need for such studies is also emphasized ina proposal fora collaborative programme of research on soil,biologica l processes and tropical soil fertility (Swift, 1984). Samenvatting

Titel: BODEMVORMING DOORTERMIETEN ,ee n studie inhe t Kisii gebied,Kenia .

De studie laat zienho e belangrijk de aktiviteit van termieten isvoo r bodem- en landschapsvorming in de tropen;ee n onderwerp,da t in bodemkundige literatuur teweini g aandacht heeft gekregen.Wa t de resultaten van deze stu diekunne n betekenen voor het landbouwkundig gebruik,word t inee n apart hoofdstuk behandeld. Twee omstandigheden inhe t studiegebied richtten de aandacht op het be lang van de termieten:enerzijd s het feit,da t in alle bodems vulkanische as aangetroffenwerd , die behalve inhe t oosten intensief gemengd isme t bodem materiaal afkomstig van de lokale rots enanderzijd s het feit,da t in alle bodems grote aantallen termieten aangetroffenwerden .D e vraagwa s nu of ter mieten verantwoordelijk waren voor deze menging en zo ja,ho e dit in zijn werk gegaanwa s enwa t hun aktiviteit verder voor consequenties kon hebben gehad voor bodem-e n landschapsontwikkeling. Het studiegebied, gelegen op eenhoogt e van 1450-2000m vlak ten zuiden van de evenaar,heef t eenhumie d tropisch hoogland klimaat met een jaarlijkse neerslag van+ 1500-2000mm , verdeeld over twee drogere en twee nattere seizoenen.No g slechts 10.000 jaar geleden ishe t klimaat belangrijk droger geweest dannu . Het landschap isheuvelachti g en bedektme t diepe doorlatende rode tot roodbruine bodems,di e nu alle intensief voor landbouw in gebruik zijn. Alleen invlakk e dalen en op slecht gedraineerde schiervlaktenkome n grondenvoo r met een zeer compacte klei -B (Planosols)e nme t eennatuurlijk e vegetatie van gras en plaatselijk struikgewas. Het hele gebied heeftmogelij k al sedert hetTertiai r de invloed van vul kanische as ondergaan.Dez e as isafkomsti g van de Rift Vallei,20 0 km ten oostenva nKisii .D e bijzondere chemische enmineralogisch e samenstelling van deze peralkalische as,blijk t een belangrijk hulpmiddel bijd eanalys e van de menging van bodemmaterialen. Het verdelingspatroon van vulkanisch glas is vooral representatief voor menging met meer recente vulkanische as.Glasfrag mentenworde n indiep e goed gedraineerde gemengde profielen tot dieptes van meer dan 7m aangetroffen,maa r het meeste glas bevindt zich in de bovenste -101- meterva n deze profielen.Magnetiet-II ,da t ôôk representatief isvoo r ouder bodemmateriaal gevormd uit vulkanische as,kom t regelmatig verspreid door het hele profielvoor ,echte r niet in aantoonbare hoeveelheden beneden de "stone- line". Het percentage bodemmateriaal afkomstig vanvulkanisch e as inmengbodem s isgescha t door vergelijking van de gehaltes aanmagnetiet-I Ii nd e bodems ontwikkeld op asme t die inmengbodems .O p dezelfde manier is gebruik gemaakt van deverhoudin g van ilmeniet tot zirkoon en van de verhouding van Al tot Fe, er vanuitgaand e dat de samenstellende bodemmaterialen elk een karakteristieke verhouding hebben. Naast glas blijkt ook de distributie van zink en kalium in profielen,ee nmaa t voor recente bijmengingme t as. Ind ehoofdstukke n 5, 6e n 7word t verder ingegaan opd e rol van termie ten bijmengin g en transport van bodemmateriaal. Inhe t gebied komen vooral schimmelverbouwende termieten voor. Sommige soorten maken bovengrondse nest heuvels,ander e nestelenuitsluiten d ondergronds.Oo k humus etende termieten met kleine nestheuvels werden aangetroffen. Termieten passenhe t milieu zo mogelijk aan hun behoeften aan.Wa t deze behoeften zijne nho e hun aktiviteit in de bodem samenhangt met deze behoeften wordt inhoofdstu k 5uitgelegd . De chemische gegevens laten zien,da t termieten ookmoeilij k afbreekbaar orga nisch materiaal zoals hout,grondi g omzetten, zodat een rest overblijft die rijk is aan basen.

Het effekt van termietenaktiviteit op de fysische eigenschappenword t in hoofdstuk 6gedemonstreerd . Met behulp vanmikromorfologisch e analyses van slijpplatenword t een schatting van de interaktie van twee processen gemaakt. Enerzijds is dit devormin g van struktuur en poriën,waaro p termieten grote invloed hebben en anderzijds het verval van de door hen gevormde "architek- tuur",waaro p eigenschappen vanhe t bodemmateriaal,maa r ook landgebruik en ontwateringstoestand, grote invloed hebben.D e interaktie van deze beide processen isva n groot belang voor menging en transport van bodemmateriaal. Tevens blijkt,da t termieten bodemmateriaal uit de bovengrond door het hele profiel verplaatsen tot dieptes van meer dan 7m omholte s op te vullen en kanalen tebepleisteren . Omgekeerd vindt ook transport naar boven plaats,da t sterker is danhe t benedenwaarts gericht transport. Hoofdstuk 7laa t zien,da t degraafaktivitei t van termieten ingoe d gedraineerde gronden verantwoordelijk isvoo r vorming van diepe homogene en fijn getextureerde grondenme t op zekere dieptemeesta l boven de saproliet, een laag bestaande uit grof materiaal,d e zg."stone-line" . In Planosols op -102- slecht gedraineerde vlakken treedt naast bioturbatie ook pedoturbatie op,al s gevolg van zwel en krimp;di t veroorzaakt menging van zowel grof als fijn materiaal. Het bodemmateriaal indez e gronden is echter zo instabiel,da t de homogenisatie teniet wordt gedaan als gevolg van illuvatie van klei, silt en fijn zand, zodat Planosols blijven bestaan. Mineralogische en ook textuur gegevens van de Wanjare katena tonenaan , dat"creep "d e belangrijkste vormva n lateraal transportwa s ind e diepe goed gedraineerde gronden. Termieten aktiviteit blijkt verantwoordelijkvoo r de mate van "creep".O p ondiepe bodems vondwe lmateriaa l transport langs het oppervlak plaats,voora l geleverd door termieten. Inhoofdstu k 8word t het effektva n termietenaktiviteit op energie en stofkringlopen gedemonstreerd inee nmodel .Ee n adequaat funktionerende makrofauna (termieten)blijk t voor eenhandhavin g van fysische bodemeigen schappen van groot belang. Instandhouding van deze aktiviteit isnie t alleen belangrijk voor de landbouw,maa r ookvoo r de bodembescherming en dewater huishouding van grotere gebieden.Landbouw , inhe t bijzonder "low input farming", zou zoveelmogelij k de (gunstige)bodembiologisch e processen moeten stimuleren enmanipuleren . Tenslotte konkludeert de auteur,da t veel meer vergelijkende studies nodig zijn tussen landbouw ennatuurlijk e ecostystemen omhe t effekt van landgebruik op de interaktie tussen bodem en bodemfauna te lerenkennen . Appendix 1.Profil edescription s andanalytica l data.

Soils were classified according to Soil Taxonomy (Soil Survey Staff, 1975)an d according toFAO/Unesc o (1974).Ne w catagories proposed by Kenya Soil Survey (seeWielemake r and Boxem, 1982)t ob einclude d in the FAO/Unesco (1975)legen d of thesoi lMa p of theWorl d and awaiting international approval,ar e indicated with an asterisk in Appendix 1.

Profile Exc. 8

Classification: SoilTaxonomy : AquicArgiudoll , FAO /Unesco:Luvi c Phaeozem Location: South-East Kisii,semi-detaile d survey,Kisi i district,shee t 130/4; 34"57,23"E. , 0°48'42" S.;Elevatio n 2145m . Described by I.Guiking ;decembe r 1975 Physiography: gently undulating top of ridge near scarp Surrounding landform: steeply dissected Meso relief: undulating plateau Slope: 2 % Parentmaterial : rhyolite (Bukoban) Land-use: maize and pyrethrum Soilfauna : termites,ant s and few worms Drainage: well drained Root development: finean d very fine roots,mainl y inuppe r 60c m (maize)

Al 0- 22c m dark reddish brown (5Y R 3/2,moist )silt y clay;moderat e to strong, fine subangular blocky;man y very fine and fine biopores;soft , very friable,slightl y sticky and slightly plastic;gradua l and smooth transition to:

A3 22- 66c m dark reddish brown to dusky red (2,5- 5Y R 3/2,moist) ; few,distinct , fine reddish yellow (5Y R 6/8) mottles;silt y clay;weak ,ver y fine subangular blocky;man y very fine and fine biopores;ver y friable, slightly sticky and slightly plastic;gradua lan d smooth transition to:

B21t 66- 125c m dark reddish brown (2,5 YR 3/4 - 5Y R 3/3, moist)clay ;weak ,ver y fine subangular blocky; moderately thick,continuou s clay cutans;man y very fine and fine biopores;friable ,stick y and plastic; clear andwav y transitionto :

B22t 125- > 175c m dark reddish brown (2,5 YR 3/4 - 5 YR 3/3, moist)clay ;moderate , fine angular blocky; clay-manganese cutans,thick ,continuous , black; fewver y fine and fine biopores;firm , sticky and plastic. -104-

Profile 53

Soil classification: dystro -mollic Nitosol;ST :orthoxi c Palehumult Ecological zone: IIa Observation: 130/2-3s, Kisii District, E 714.9, N 9926.3, 1880ra, 22-3-1975 Geological formation: bukoban system Local petrography: andesite Physiography: saddle between hill-tops Surrounding landform: hilly Relief -meso : uniform slope of 500m Vegetation: 80%grasses ,fe w bushes Land use:grazin g Erosion: nil Surface stoniness: nil Soil fauna: termites and ants Slope gradient: 5% Drainage class:wel l drained

All 0-50 cm dark reddish brown (5YR 3/2, moist) clay; moderate fine subangular blocky;ver y friable, slightly sticky and slightly plastic; many fine and very fine, common medium and few coarse pores; clear and smooth transition to:

A12 50-79 cm dark reddish brown (2.5YR 3/3,moist ) clay; moderate fine subangular blocky; friable, slightly sticky, slightlyplastic ;man y fine and very fine,commo n medium pores; gradual and smooth transition to:

B21 79-270 cm dark reddish brown (5YR 3.5/4 moist) clay; moderate fine angular- subangular blocky; friable, slightly sticky and slightly plastic; moderate thick broken clay-cutans;man y fine and very fine pores,fe wmediu m pores;fro m 205-210 cm very many, small soft manganese concretions; clear and smooth transition to:

B22t 270-700 cm yellowish red (5YR 4/6, moist) clay;moderat e medium angular blocky; firm, slightly sticky and slightly plastic; common fine and very fine pores thick broken clay cutans;blac k continuous coatings from iron and manganese; gradual and smooth transition to:

B23t 700-750+ cm yellowish red (5YR 4/6,moist )ver y fine clay; moderate medium-coarse prismatic; firm,slight ly sticky and slightly plastic;patch y thick clay cutans,continuou s thick cutans of iron and manganese; commqn fine and very finepores . -105-

Profile 210

Soil classification: humicFerralsol ;ST :typi c Haplohumox Ecological zone: IIa Observation: 130/3-210, Kisii District, E 6886.3,N 9912.6, 1833m , 26-2-1974 Geological formation:Middl e Kisii series,bukoba n system Local petrography: quartzites and siltstone Physiography: concave slope near summit of plateau-remnant Surrounding landform: undulating to rolling Relief -meso : concave single slope Vegetation: 80%grasses ,20 %shrub s Land use:extensiv e grazing Erosion: nil Surface stoniness:n o stones Soil fauna: termites Slope gradient: 3% Drainage class:wel l drained

All 0-12 cm dark reddishbrow n (5YR 3/4,dry ) clay;mod erate very fine and fine subangular blocky; common, very fine continuous random, inped, tubular biopores;very hard, non-sticky and slightly plastic; common fine roots;fe w coar se,blac k sand grains;clea r and smooth tran sitionto :

A12 12-31 cm dark reddish brown (2.5YR 3/4,moist )clay ; moderate very fine and fine subangular blocky; common very fine, fine and medium, random biopores; friable, non-sticky and slightly plastic; few coarse,blac k sand grains;abun dant very fine and fine roots; clear and smooth transition to:

Bl 31-48 cm dark reddish brown (2.5YR 3/5,moist ) clay; moderate, very fine to fine subangular to angular blocky; few very fine and medium, random, biopores; friable,non - to slightly sticky, slightly plastic; common very fine roots; gradual and smooth transition to:

B21 48-95 cm dark reddish brown (2.5YR 3/6,moist ) clay; moderate fine angular blocky; fewver y fine and finebiopores ; friable to firm, slightly sticky and slightly plastic; thinpatch y cutans; fewver y fine roots gradual and smooth transition to:

B22 89-155 cm dark reddish brown (2.5YR 3/6,moist )clay ; moderate fine subangular to angular blocky; common very fine and fine, random, tubular biopores; friable to firm, slightly sticky and slightly plastic; thinpatch y cutans;a t 130-140 cm depth few large,angula r and slight ly weathered quartzite gravels; clear and wavy transitionto : -106-

Profile Wan 1

Soil classification:Ranker ;ST : lithic Humitropept Ecological zone:II b Observation: 130/1-WAN 1, Kisii District, E.689.9, N 9930.1, 1647m , 19-12-1975 Geological formation: Precambrium/post kavirondian system Local petrography: Wanjare granite Physiography: hill top Surrounding landform: rolling to hilly Relief -meso : eroded topwit h termite mounds (30 cm) Vegetation: dense bushland; 30%shrubs ,20 %herbs ,30 %grasse s Land use:extensiv e grazing Erosion: slight sheet and gully erosion Surface stoniness/rockiness: extremely rocky, few stones Soil fauna: termites and ants Slope gradient: 2% Drainage class: somewhat excessively drained

All 0-5/14c m dark brown (7.5YR 3/2, dry),dar k reddish brown (5YR 3/2 moist) slightly gravelly sandy clay loam; strong very fine granular; due to termite activity abundant pores inal l size classes up to 10mm ; soft,friable , slightly sticky and slightly plastic;clea r and broken transition to:

A12 5/14-20/35 cm dark brown (7.5YR 3/2 dry),dar k reddish brown (5YR 3/2,moist ) slightly gravelly sandy clay loam; strong very fine granular; many fine and fine biopores; soft, friable slightly sticky and slightly plastic; abrupt and broken transition to:

20/35+ cm rock (granite)

ProfileWa n 2

Classification: SoilTaxonomy : orthoxic Tropohumult FAO /Unesco:humi c Acrisol Location: sheet 130/1 (Gem),nea r Irigonga school on the slope of theKebuy e range 34°41'53"E., 0°37'58"S.,elevatio n 5360 ft./1634m. , described by S. Slager and ii.va nReule r(19 - 12-1975). Physiography: upper part of a footslope, just beneath the steeper hillslope to the top Surrounding landform: rolling tohill y Meso relief: flat, scattered granite boulders Micro relief: termitemound s up to 30c m Slope: 12% Parentmaterial : Wanjare granite (Pre-Cambrian/Post Kavirondian) Vegetation/land-use: dense bush land with trees; 15%trees , 60% shrubs,15 %herbs ,40 %grasse s (5%bar e ground), extensively grazed. Surface stoniness: very stony (up to 1m . diameter) Drainageclass : somewhat excessively drained Soilfauna : termites,ants , some beetles andworm s -107-

Al0 - 2 2c m darkbrow nt obrow n(7. 5Y R4/2 ,dry) ,dark . brown(7. 5Y R3/2 ,moist) ,gravell y (angular) loamysand ;strongl ydevelope dgranula r structure(diamete r3 mm. )abundan tver yfin e biopores;soft ,ver yfriable ,slightl ystick y andslightl yplastic ,gradua lan dsmoot h transitionto :

B2t2 2- 4 0c m yellowishre d( 5Y R5/6 ,dry) ,yellowis hre dt o red (5Y R4/ 6- 2. 5Y R4/6 ,moist) ,gravell y (angular)loam ysand ;n omacr ostructure , abundantver yfin ebiopore s (manyaggrotubule s filledwit hA lmaterial) ;soft ,ver yfriable , nonstick yan dno nplastic ;abrup tan dbroke n transitionto :

R4 0- 45/7 1c m rottenroc kwit hhar dnucleu so funweathere d granite,abrup tan dbroke ntransitio nto :

IIB21 t45/7 1- 8 5c m yellowishre d( 5Y R4/6 ,dry) ,dar kre d(2. 5Y R 3/6,moist) ,ver ygravell y (angular)sand y loam;fe wcompletel yrotte nroc kpiece su pt o 15c mdiameter ;n omacr ostructures ,man yver y finebiopores ;hard ,friable ,no nstick yan d nonplastic ;gradua lan dsmoot htransitio nto :

IIB22 t8 5- 13 5c m yellowishre d( 5Y R4/6 ,dry) ,yellowis hre dt o darkre d (5Y R3/ 6- 2. 5Y R3/6 ,moist) ,ver y gravelly (angular)sand yloa mwit hfe w completelyrotte nroc kpiece su pt o1 5c m diameter,n omacr ostructure ,man yver yfin e biopores;hard ,friable ;no nstick yan dno n plastic;gradua lan dsmoot htransitio nto :

IIC I13 5- 15 5c m yellowishre d (5Y R5/6 ,dry) ,re dt oyellowis h red(2. 5Y R4/ 6- 5 Y R4/6 ,moist) ,ver y gravelly (angular)sand yloa mwit hfe w completelyrotte nroc kpiece su pt o1 5c m diameter;n omacr ostructure ,commo nver yfin e biopores;abrup tan dsmoot htransitio nto :

IIC 215 5- 165c m hardenedmateria lwit hiron/manganes e concretions,20 %distinc tbrow nmottle s (10m m diameter),60 %distinc tre dmottle s(1 0m m diameter),20 %prominen tblac kmottle s (15m m diameter),plat ystructure ,fe waggrotubule s filledwit hA lmaterial ,hardl ybreakable , abruptan dsmoot htransitio nto :

IIR + 165c m weatheredgranite ,rapidl ychangin gint ofres h graniterock .

Remark:I nth etopsoi l0 - 45/7 1c msom eartefact swer efoun despeciall y inth euppe rpar to fth eI IB21 thorizon . -108- Profile Wan 3

Soil classification: mollic Nitosol;ST : typic Paleudoll Ecological zone: IIb Observation: 130/1-WAN3, Kisii District, E 689.5, N 9930.1, 1567 m, 30-12-1975 Geological formation: Post kavirondian intrusive inNyanzia n system Local petrography: quartz diorite Physiography: upper part of a footslope Surrounding landform: undulating - rolling Relief -meso : terraces,develope d from thrash-lines Vegetation: no natural vegetation Land use:cultivate d area (maize) Erosion: few rills Surface stoniness:ni l Soil fauna: termite activity Slope gradient: 8% Drainage class:wel l drained

Al 0-30 cm reddish brown (5YR 4/2, dry) dark reddish gray (5YR 3/2,moist )clay ;moderat e medium compound subangular blocky,man y very fine, fine and few medium biopores; hard, very friable, slightly sticky and slightly plastic; clear and smooth transition to:

A3 30-55 cm reddish brown (2.5YR 4/3-5YR 4/3,whe ndry) , dark reddish brown (5YR 3/2,moist ) clay to sandy clay; moderate very fine and finean gular blocky many very fine,fin e and common medium biopores; slightly hard,ver y friable, slightly sticky and slightly plastic; gradual clear and smooth transitionto :

Bit 55-85 cm reddish brown (5YR 4/4, dry) reddish brown (5YR 4/3,moist ) clay;moderat e fine subangular blocky;wea k to common clay and organic matter cutans; many fine, common medium and a few coarse biopores;hard , friable,slightl y sticky and slightly plastic; gradual and wavy trans itionto :

B21t 85-110 cm reddish brown (2.5YR 4/4-5YR 4/4,dry) ,wea k red to reddish brown (2.5YR 4/3,moist ) clay; moderate fine and medium subangular blocky; moderate common clay/organic matter cutans; many very fine and fine,commo n medium, few coarse and very coarse biopores;hard , friable, slightly sticky and slightly plastic; gradual and wavy transition to:

B22t 110-235 cm reddishbrow n to red (2.5YR 4/5, dry),reddis h brown (2.5YR 4/4,moist )clay ;moderat e fine and medium subangular blocky;moderat e common to abundant clay/organic matter cutans;man y very fine, common medium and few coarse biopores;ver y hard, friable,slightl y sticky and slightly plastic,gradua l and wavy transi tion to: -109-

B23t 235-290 cm red (2.5YR 4/6, dry),reddis h brown to red (2.5YR 4/5,moist ) clay;moderat e medium subangular blocky;moderat e common clay/organic matter cutans; many very fine, common fine and few medium, coarse and very coarsebiopor - es; hard, very friable,slightl y sticky and slightly plastic; clear and wavy transition to:

B24t 290-245 cm red (2.5YR 3/6, dry),reddis h brown (2.5YR 4/4, moist) clay;wea k medium and fine subangular blocky;moderat e common clay/organic matter cutans; 15%3- 4 mm diameter iron-man ganese concretions; many very fine, common fine and few common coarse-ver y coarsebio - pores; hard, friable; slightly sticky and slightly plastic; abrupt and wavy transition to:

B31 345-485 cm yellowish red to red (5YR 5/6-2.5YR 5/6,dry) , dark red (2.5YR 3/6, moist), common prominent 3-6 mm diameter manganese mottles;ver y grav elly clay, no macro structure; common fine and medium biopores,ver y hard, firm non-sticky and non-plastic; clear and wavy transition to:

B32 485-550 cm yellowish red to red (5YR5/6-2.5Y R 5/6,dry) , dark red (2.5YR 3/6, moist); abundant 5-10 mm diameter manganese mottles;ver y gravelly clay; nomacr o structure; common fine and medium bio pores; very hard,ver y firm,non-stick y and non-plastic; gradual and wavy transition to:

C 550-750 cm multicoloured, varying from red (2.5YR 5/8, dry) to dark red (2.5YR 3/6 moist), red (10YR 5/8,dry )t o dark red (2.5YR 3/6 moist), red (10YR 5/8,dry )t o red (10YR 5/6 moist), withman y white and yellow rotten rockspots ; abundant 5-10 mm diameter manganesemottles ; very gravelly clay;n omacro-structure ; com mon fine and medium biopores (see remark), very hard,ver y firm non-sticky and non-plas tic.

Remark: many agrotubules diameter 5m m filled with clay reddish brown (2.5YR 4/4) whenmoist , they occur from about 5.50 m to the bottom of thepi t (7.50 m). -110-

Profile Wan9

Soil classification: humic Ferralsol ST:typi c Haplohumox Ecological zone: IIb Observation: 130/1-WAN9 ,Kisii District, E 687.4, N 9931.0, 1571m , 2-3-1976 Geological formation: Post Kavirondian intrusive inNyanzia n system Local petrography: quartz-diorite Physiography: flat-topped ridges Surrounding landform: undulating to rolling Relief -meso : nil Vegetation: 60%grass ,40 %herb s Land use: grazing Erosion: nil Surface stoniness:ni l Soil fauna: termite activity to 2.17 m Slope gradient: 0% Drainage class:wel l drained

All 0-16 cm dark reddish brown (5YR 3/3)whe n moist,clay ; weak fine angular to subangular blocky; fri able, slightly sticky and slightly plastic; fewver y fine pores;clea r and smooth transi tionto :

A12 16-38 cm dark reddish brown (2.5YR 3/4) whenmoist , clay; moderate fine subangular blocky; fri able, slightly sticky and slightly plastic; few medium and fine, many very finepore s; common termite chambers 5-20 cm indiameter ; gradual and smooth transition to:

Bl 38-80 cm dark red (2.5YR 3/5, moist clay; moderate fine subangular blocky; friable, slightly sticky and slightly plastic; fewmediu m and fine, many very fine pores; common termite chambers 5-20mm in diameter; gradual and smooth transition to:

B21t 80-130 cm red (2.5YR 4/5) when moist, clay; moderate very fine angular to subangular blocky; fri able, slightly sticky and slightly plastic; moderately thick, common clay cutans; many fine and very fine, fewmediu m pores;smal l termite holes of 2 cm diameter; gradual and smooth transition to:

B22t 130-210 cm red (2.3YR 4/6) when moist, clay with some quartz gravel; moderate fine and very fine angular blocky; slightly sticky and slightly plastic;moderatel y thick common clay cutans; many fine and very fine pores small termite holes of less than 2 cm diameter; one piece of rotten rock with diameter of 5 cm; gradual and smooth transition to: -in-

Bas 210-280 cm likeB22 ,bu t with 5% rounded soft manganese concretions of 5 cm diameter, clear and smooth transition to:

B24 280-310 cm like B22,bu twit h 10%sof t rounded manganese concretions of 5-10 cm diameter,abrup t and wavy transition to:

B3 310-330 cm dark red (2.5YR 3/6) when moist, clay with 5% small quartz gravel; moderate very fine angular blocky; firm, slightly sticky and slightly plastic; moderate thick, common, clay and manganese cutans; few finepores ; 75% rounded and angular manganese concretions or 3-8 cm diameter; clear broken ana wavy transition to:

330+ cm mixtures of rotten rock,quart z diorite ston es and gravelly soil.

Remark: at 310-330 cm allochthonic stone fragments were found.

Profile Wan 4

Classification: SoilTaxonomy : typic Paleudoll ace. to FAO (1974) / mollic* Nitosol Coordinates: 34°42'55" Ean d 0°37'43" S;Altitude : 1528m described by R.F. Breimero n 15-1-1976 Petrography: diorite (intrusion inWanjar e granite formation) Physiography: general: rounded hills with long straight slopes land unit: lower part of linear slope Macro relief: undulating to rolling Slope: 700m , convex-linear-slightly concave,10 % gradient Land-use: arable land-used for maize Rock outcrops: General ground water level: deeper than 2m Drainageclass : somewhat imperfectly towel l drained Soilfauna : very high activity of termites,ant s andmoles , in theuppe r metre,fro m 1-2m depth ahig h activity and below 2m a moderate activity.

Ap 0- 10c m dark reddish brown/grey (5Y R3 - 4/2,moist) , reddishbrow n (5Y R 5/3,dry) ,clay ; moderate very fine subangular blocky and crumby structure, common fine and very finepores ; hard,ver y friable,slightl y sticky and slightly plastic; clear smooth transition to: -112- Al 10- 43c m dark reddish brown (5-2.5 YR 3/3, moist), reddish brown (5-2.5 YR 4/3,dry) ,clay ; moderate fine subangular blocky,man y very fine and fine,fe wmediu m and very coarse pores; hard, friable,slightl y sticky and slightly plastic;gradua l and smooth transition to:

Bl4 3 - 80c m dark reddish brown (2.5 YR 3-4/4, moist), reddish brown (2.5Y R 5/4,dry) ,wit h dark reddish brown (2.5Y R 3/4) ped surfaces,havin g at places fine faint darkmottles ;clay ; moderate fine subangular blocky,man y moderate clay cutans,man y fine and very fine,fe w medium and very coarse pores;hard ,friable , slightly sticky and slightly plastic;diffus e and smooth transition to:

B21t 80- 210c m reddish brown (2.5Y R 4/4,moist )an d (2.5 YR 5/4,dry) ,wit h dark reddish brown (2.5Y R 3/4) ped surfaces having fine faint darkmottling ;clay ;stron g fine subangular blocky toangula r blocky,man y continuous thick clay (+manganese )cutans ;man y very fine and fine, fewmediu m and coarse pores;ver y hard, firm to friable, slightly sticky and plastic; clear and slightly wavy transitionto :

B22t 210- 235c m reddish brown (5-2.5 YR 4/4,moist) , and (5- 2.5 YR 5/4,dry) ,wit hman y finean d medium prominent blackmottles ;clay ;stron g fine subangular toangula r blocky,man y moderate clay (+manganese )cutans ,man y very finean d fine,fe wmediu m and coarsepores ; very hard,firm , slightly sticky and plastic; common 2-20m m big black manganese (+ iron ?) concretions;gradua lwav y transition toB3 :

B3 235- 320c m reddish brown (5Y R 4/4,moist )an d (5Y R 5/4, dry),wit h many medium prominent black mottles; slightlu gravelly clay;moderat e fine tomediu m subangular toangula r blocky, common moderate clay (+manganese )cutans ;commo nver y fine, few finean d medium pores;ver y hard, firm, slightly sticky and plastic; common 1-5 mm big black manganese/iron concretions, gradualan dwav y transitionto :

BC 320- 360c m reddish brown (5Y R 7/4) mixed with black and reddishyello w rotten rock colours;ver y gravelly clay;wea k fine angular blocky structure; fewver y fine pores;ver y hard, firm, sticky and plastic,man y 1-5m m big rounded blackmanganese/iro n concretions,clea r and wavy transition to thehar d rock.

horizon: diorite rock at 350- 370c m depthwit h a thin decomposed rock cover. -113- Proflle Waft 5

Classification: SoilTaxonomy : typicArgiudoll . FAO /Unesco:Luvi c Phaeozem (Hl). coordinates: 34°42*56" Ean d 0*37'43" S; altitude: 1523m described byR.F . Breimer on 5-1-1976. Petrography: diorlte (intrusion inWanjar e granite formation) Physiography: general: rounded hills with long straight slopes land unit: lower part of linear slope. Macro relief: undulating (5-8% )t orollin g (8-16 %) Slope: 700m , convex-linear-slightly concave, 10% slope Land-use: maize Rockoutcrops : General ground water level: from 60-120c m tover y deep Drainageclass : moderately well drained Soilfauna : 0-100 cm:hig h activity of ants and termites 100-150cm :moderat e activity over 150cm :n o visible activity

Al 0- 30c m dark brown (7.5 YR 3/2,moist) , brown (7.5Y R 5/2,dry) ,sand y clay loam;moderat e fine tomediu m subangular blocky and moderate very fine crumby;man y very fine and fine, few medium and coarse pores;slightl y hard,ver y friable, slightly sticky and slightly plastic; gradual and smooth transition to:

Bl 30- 55c m dark reddish grey (5Y R 4/2,moist) , clay loam; moderate medium tofin e subangular blocky; many very fine,commo n fine and fewmediu m and coarse pores;hard , friable,slightl y sticky and slightly plastic; clear andwav y transition to:

B21t 55- 95c m reddish brown (5Y R4/ 3 - 4,moist) , clay loam; with many coarse prominent blackmottles ; moderate medium to fine subangular blocky, common cutans; man y very fine,commo n fine, few medium and coarse pores;ver y hard, friable to firm, slightly sticky and slightly plastic; many manganese/iron concretions of 1-5 mm diameter;gradua l and wavy transition to:

B22t 95- 150c m mixed brown colours (7.5 YR 5/2+ 5/4, moist), clay loamwit h many coarse prominent black mottles;moderat e medium to fine subangular blocky,commo n clay and manganese cutans, common very fine, few fine,mediu m and coarse pores;ver y hard, friable tofirm , sticky and slightly plastic;man y 1-5m m big manganese concretions; clear and wavy transition to: -114-

At the transition a 5c m thick layer occurs,consistin g of small gravels and black concretions.

B23tg 150- 215c m mixed grey and brown (7.5Y R 5/1 + 5/3,moist) , slightly gravelly clay with common medium to coarse prominent blackmottle s and many fine prominent strong brown (7.5 YR 5/6)mottles ; moderate to strong fine tomediu m angular blocky,man y strong clay and manganese pressure cutans;fe wver y fine and fine pores;ver y hard, firm, sticky and plastic;man y 0.5-2 cm big manganese/iron concretions;clea r and wavy transition to:

B3g 215- 310c m dominant colour is pinkish grey (7.5Y R6/2) , mixed with brown (7.5 YR 5/2,moist) ; slightly gravelly clay withman ymediu m to coarse prominent strong brown (7.5 YR 5/6)an d black mottles;wea k fine tomediu m angular blocky; many strong manganese and pressure cutans,fe w very fine and fine pores;ver y hard, firm, sticky and plastic;abrup t and wavy transition to the rock boulders. The depth of the rockvarie d between 2.75 andmor e than 3.20m , due to boulder weathering of the diorite rock.

Termite mound sequence atOpap o

Profiles,MIA ,D and E are associated soils.MI A is located in the active part of the termite mound,MI Di s located on the slope of the mound andM1 E is located in the Plain,wher e no influence of themoun d isnoticeable .Profil e M1E is similar toM 3 ofwhic h description and analytical data are given asdescriptio n no. 36 inApp . 1o f Wielemaker and Boxern,1982 . Date of observation: 19-7-1978;describe d by M.Nacheniu s and W.G. Wielemaker Location: Opapo, SouthNyanz a district,Nyanz a Province inKeny a (Sheet 130/1,gri d lines 9923/671); 1340m . Physiography: Plain;slope :0 - 1% Parentmaterial : Granite overlain by alluvium and volcanic ash. Macro relief: Gently undulating;meso-relief-flat ; micro relief: termitemound s of 1- 2metr e high andwit h a diamètre of 3- 10metr e at distances of 30- 50metre . Vegetation and land use:O n themoun d shrubs;betwee n themound s grass, which is extensively grazed. Drainage: On themound , well drained, but between the mounds, poorly drained. Surface stoniness: nil;erosion :ni l Classification: M1E isa n abruptic Tropaqualf (ST)an d a eutric Planosol (FAO/Unesco) Processes: Inprofil eMI A termites have thoroughly mixed the different horizons as occurring in profile M1E.MI D is a transition toM1E . -115-

ProfileMI A

+ 190- + 30c m Darkbrow n(7. 5Y R3/2 )dry ,man ymediu mfain t brown(7. 5Y R4/4 )mottles ,clay ;stron gver y fineangula rblocky ;slightl ysticky ,slightl y plastic,ver yhar ddry ;gradual ,smoot h transitionto :

+ 30- 110c m Verydar kgre y(1 0Y R3/1 )dry ,commo nmediu m distinctdar kgreyis hbrow n(1 0Y R4/2 ) mottles;clay ;stron gmediu m tocoars eangula r blocky;har ddry ,slightl ysticky ,no nplastic ; clearwav ytransitio nto :

110- 165c m Mixedcolours ,dar kgreyis hbrow n(1 0Y R4/2 ) clayan dwhit e(1 0Y R8/1 )dry ;gravell yheav y clay;massiv ewit hman yrounde dstone s( 5t o 20cm) ,man ygravel ,clea rwav ytransitio nto :

C16 5+ c m Mixedcolour sgreyis hbrow n(1 0Y R5/2 )an d paleoliv e( 5Y R6/4 )moist ,gravell yclay ; massive;man yangula rquart zgravel .

ProfileMI D

Al0 - 2 8c m Lightgre yt oligh tbrownis hgre y(1 0Y R6/1.5 ) dry,ver ydar kgreyis hbrow n(1 0Y R3/2 )moist ; clayloam ;wea kver yfin eangula rblocky ; friable;slightl ysticky ,slightl yplastic ;fe w gravel;clea rwav ytransitio nto :

A22 2- 4 5c m Lightgre y (7.5Y R7/0 )dry ,commo nmediu m distinctyellowis hbrow n(1 0Y R5/4 )mottles ; clayloam ;wea kcoars eprismati ct oporou s massive;friable ;no nsticky ,no nplastic ;fe w gravel;gradua lsmoot htransitio nto : Remark:commo ncontinuou sthic kcla ycutan s linepore san dhole sespeciall yi nth elowe r parto fth eA „horizon .

B2t45- 8 5c m Verydar kgreyis hbrow n(1 0Y R3/2 )moist ,dar k greyishbrow n(1 0Y R4/2 )clay ,commo nfin e faintyellowis hre d( 5Y R4/6 )mottles ;clay ; moderatemediu mangula rblocky ;slightl y sticky,slightl yplasti cdry ;fe wgravel ; abruptan dwav ytransitio nto : Remark:A „materia li spresen talon gcracks ;

85- 10 5c m 90% grave lan dstones ,embedde di ncla y

105+c m Asi nBo »abov e -116-

Profile MIE

Al 0- 12c m Very dark greyish brown todar kgreyis h brown (10Y R3.5/2 )moist ,dar kgre y (10Y R 4/1)dry , common faint dark yellowish brownmottles ; loam;wea k tomoderat e very fine subangular blocky;ver y friable;slightl y sticky and slightly plastic;fe wver y finean d fine pores; few fine and very many very fine roots;clea r and smooth transition to:

A21 12- 42c m Dark greyish brown (10Y R4/2 ) moist,gre y (10Y R 5/1) dry;commo n fine faint dark yellowish brownmottles ; sandy clay loam; weak very fine subangular blocky;ver y friable, slightly sticky,slightl y plastic; fewver y fine and fine pores;commo n very fine and very few fine roots;clea r andwav y transitionto : A224 2- 50c m Like A-,above ,bu t gravelly coarse sandy loam; abrupt and broken transitionto :

B2tl 50- 60c m Very dark brown (10Y R 2/2)moist ,man y medium prominent red (2.5Y R 5/8) mottles; clay; strong fineangula r blocky; firm;stick y and slightly plastic;thic kcontinuou s clay cutans; fewver y fine pores;commo n fine roots;clea r and wavy transition to:

B2t2 60- 105c m Very dark brown (10 YR 2/2)moist ,man y distinct fine reddish yellowmottles ;gravell y and stony clay; strong fine angular blocky; few weakno t intersecting slickensides;stick y and slightly plastic; thick continuous clay cutans; no pores;commo n fine roots;stone sar e mainly quartzitic with a few jaspis andweathere d granites;clea r andwav y transition to:

B3 105+c m Dark grey (10Y R 4/1) moist;commo n fine distinct yellowish brownan d whitemottles ; clay withweathere d rock particles;wea k coarse angular blocky;fir mmoist ,stick y and slightly plastic,wet ;n opores .

Profile NARS.

Soilclassification : mollic Nitosol;ST :typi c Paleudoll Ecological zone: lib Observation: 130 NARS,Kisi iDistrict , E698.2,N9925.50 , 1885m , 18-10-1976. Geological formation: bukoban system (Precambrium) Local petrography: basalt Physiography: middle part of straight lateral slope Surrounding landform: rolling upland -117-

Relief-meso: nil;micro :an t and termite heaps Land-use: grazing Erosion: nil Surface stoniness: nil Soil fauna: termites; dung beetles;ant s andmol e rats Slopegradient : 12% Drainage class: well drained

Al 0- 40c m dark reddishbrow n (5Y R 3/4 dryan d moist) slightly gravelly clay;moderat e fine subangular blocky;hard ,firm ,no n sticky and nonplastic ;clea ran d smooth transition to:

Bl 40- 70c m dark reddish brown (5Y R 3/3,moist )gravell y clay; strong fine tomediu m subangular blocky; hard, firm, slightly sticky and slightly plastic;clea r and smooth transitionto :

B21t 70- 80 cm dark reddish brown (5Y R 3/4,moist )clay ;fin e broken dark clay cutans;wea kmediu m subangular blocky;hard , firm, slightly sticky and slightly plastic;gradua lan d smooth transition to:

B22t 80- 150c m dark red (5Y R 3/6,moist )clay ; strong medium angular blocky;broke n fine darkcla y cutans; hard, firm,no n sticky and slightly plastic; diffuse and smooth transition to:

B23t 150- 200c m dark red (2.5 YR 3/6,moist )cla y with black spots; strong subangular blocky; fewfin e clay cutans; friable ,slightl y sticky and slightly plastic;diffus e and smooth transition to:

B24t 200- 300c m dark red (2.5 YR 3/6,moist )cla y with common, medium and distinct blackmottles ;fe wmoderat e clay andFe/M n cutans;stron g very fine,fin e and medium subangular blocky; friable, slightly sticky and slightly plastic;abrup t and smooth transition to:

B3 300- 330c m Gravelly,boulder y and stony clay with common black and yellow spots;fe wmoderat e clay and Fe/Mn cutans;moderate ,ver y fine angular blocky; friable,slightl y sticky, slightly plastic. -118 Profile Depth Particle sizedistribution % Density separations(% ) (em) sand fractions(m m ) sand silt clay insan d (0.05-C .1mm) 2-0.25 0.25-0.1 0.1-0.05 d<2.5 2.5-2.9 >2.9

V26 0- 23 n.d. n.d. 17 70 15 15 92.6 6 1 4 133/1 10- 20 8.7 7.0 3.1 18 8 46.8 34.4 22.5 77.2 0 3 40- 50 8.6 7.5 3.1 19 2 44.4 •36.4 37.3 60.5 2 1 80- 90 9.2 7.0 2.6 18 8 43.3 37.9 46.1 51.7 2 3 118/4 10- 20 8.6 2.6 2.2 13 2 24.3 62.2 52.5 45.4 2 1 40- 50 2.4 3.6 2.4 8 4 63.9 27.6 58.0 39.8 2 1 80- 90 1.5 2.8 1.5 5 8 39.3 54.9 40.9 56.4 2 6 132/2 10- 20 7.9 5.2 2.8 15 9 43.9 40.2 20.7 75.9 3 4 40- 50 2.6 3.5 2.7 8 8 49.5 41.7 25.0 72.5 2 5 80- 90 2.7 4.7 3.5 10 9 45.2 43.9 26.1 70.8 3 0 131/4 10- 20 3.0 2.7 0.8 6 5 77.4 15.9 21.0 77.3 1 7 40- 50 2.5 2.9 0.8 6 2 57.2 36.6 28.9 69.5 1 6 150-200 10.4 5.3 4.0 19 7 37.3 42.9 28.6 70.1 1 3

Particlesiz e distrlbuti on (mass %) Density sépara tions (Z) Prof lieDept h sand fractions inm m insan d (0.05 - 0.1 mm) (cm) 1-2 0.5-1 .00.25- 0 50.1-02 S 0.05-3.1 sand silt clay d<2.5 2. 5-2 .9 >2.9

Exe. 8 0- 22 3.1 6.8 7.4 8.0 7.8 33.1 45.7 21.2 22- 66 4.0 5.0 5.5 6.2 6.2 26.9 41.6 31.4 13.8 84 .7 1.5 90- 105 4.6 5.2 6.0 6.5 6.0 28.3 38.9 32.7 150- 165 6.0 7.4 6.6 6.3 5.6 32.0 46.1 21.9 . 2.1 200-25 0 S3 0- 50 0.9 1.6 2.3 1.9 6.7 31.2 62.1 6.9 89 4 1.1 50- 79 0.6 1.7 2.3 2.3 6.9 28.4 64.7 7.7 91 0 1.3 79- 270 0.4 1.8 2.0 1.6 5.8 26.1 68.1 21.8 77 1 1.1 340-34 5 0.5 1.8 2.6 2.1 7.0 21.0 72.0 15.6 83 2 1.2 600-61 0 1.0 2.1 2.7 1.8 7.6 19.3 73.1 15.4 83 9 1.6 740-75 0 0.7 1.8 3.1 4.2 2.5 13.0 17.3 69.7 12.8 85 7 1.4 S1 950-1000 . . 20.7 50 29.3 41.8 57 6 0.6 210 10- 20 2.2 1.9 2.8 2.3 9.2 20.9 69.9 9.6 90 0.4 30- 40 1.4 1.8 2.6 2.2 8 19.8 72.2 10.6 89 0 0.4 50- 70 1.2 1.5 2.3 2.2 7.2 17.2 75.6 13.9 85 6 0.5

Prof LieDept h Org.C pH Exchangeable bases(mmol/kg ) CEC(NH4Ac) Base Bulk

(cm) g/kg) H20 KCl iCa }Mg K Na at pH7 0 saturati on density (mmol/kg) ( 1 ) 'g/kg)

Exe. 8 0-2 2 6.2 5.5 200 30 24 250 100 . 22-6 6 6.1 5.2 126 21 6 168 91+ 90-105 6.0 5.0 53 16 8 86 90+ 150-165 6.0 5.3 37 26 15 89 88+ 200-250 5.9 4.9 37 25 23 95 89+ . S3 0-5 0 24 5.2 4.8 86 30 4 203 59+ 1090 50- 79 17 5.5 4.8 67 35 5 187 57+ 79-270 4 5.4 4.2 11 20 7 136 30+ 1100 340-345 4.5 4.0 16 6 7 140 25+ 1190 600-610 4.7 4.1 39 11 8 150 39+ 1220 740-750 5.0 4.1 47 17 8 144 50+ 1220 210 10- 20 n 5.0 4.0 2 1 1 118 3.4 940 30- 40 7 5.0 4.1 tr tr tr 100 1.0 n.d. 50- 70 9 5.1 4.2 tr tr tr 87 1.1 1130 90-110 4 5.3 4.3 1 tr tr 88 2.3 1180 130-150 3 5.0 4.3 tr tr tr 81 1.2 1150 - 119- Magnetic separates(% ) zirconspe r Profile Dept tl inssan d (0.05-0.lmm,d>2.9 )2.9 ) 100grain si n Analyticaldat a ofprofile s onvolcani cas h (cm) s.m. n.m. n.m. fraction outside themai nare ao fstudv .Chemical , somemineraloglca lan dmorphologica ldat a V26 0- 23 18.8 37.6 43.7 - are giveni nWielemake ran dWakatsukl ,1984 ; 133/1 10- 20 25.0 69.0 6.0 . profile V26- 1;133/ 1- 2; 118/4- 3; 132/2- 4 40- 50 25.9 72.4 1.7 13 and 131:4-5; .- no t determined. 80- 90 21.4 76.9 1.7 118/4 10- 20 21.9 73.4 4.7 40- 50 20.5 72.7 6.8 80- 90 28.3 62.2 9.6 132/2 10- 20 34.5 63.3 2.1 40- 50 30.5 67.8 1.7 34 80- 90 47.1 51.1 1.8 131/4 10- 20 11.0 84.1 4.9 40- 50 11.0 81.6 7.4 57 150-200 11.0 84.0 5.0 AAnalyticar l data of profiles within magnetic separates Xi n minerals per 100 themai n area of study. Profilesan d (O.05-0.1mm,d>2.9) transparent grains in thenon-magneti c fraction zirconAegirin e -Amphib o LePyroxen e Exc.8 augite 13.5 66 7 19.8 23 _ - Comp ositionaccordin g toX-ra y fluorescence strongly magnetic weaklymagneti c 5.8 39 9 54.3 • Ml M2 IH R Z I R A M2M l H S 3 4.0 81 0 15.0 +++ +++ tr- Mill tr+ - tr tr 13.1 74 3 12.6 20 8 26 1 -H-+ +++ tr Mill tr+ trt r tr 10.6 73 6 15.6 13 40 20 1 +++ +++ tr- Mill tr+ + tr - 1.1 86 0 12.9 26 11 26 0 + - Mill tr+ - tr tr 17.8 68 0 14.1 38 7 22 0 +++ +++ tr- Mill tr+ - - + 5.7 84 6 9.8 70 tr 5 0 +++ +++ tr- Mill tr+ - - + S1 3.0 45 9 51.1 0.0 epidote 100 INN tr- Mill tr+ - + - 210 14.8 63 0 22.2 12.5 62. 5 25.0 4.5 50 45.4

Profile Depth Moisture% Mass %o f majorelement s in (cm) volumic massa tp F fine earth(calculated asoxid e s)

2.0 3.0 4.2 Si02 A1203 Fe203 K20

Exc. 8 0-2 2 58.4 12.1 8.0 2.6 22- 66 60.7 13.4 8.9 2.8 90-105 61.7 14.8 9.5 3.1 150-165 61.7 14.8 10.4 2.7 200-250 57.7 15.0 12.7 2.7 S3 0-5 0 48.5 28.9 22.5 47.5 20.2 13.8 1.7 50- 79 79-270 45.6 32.4 25.5 46.5 24.0 14.7 1.4 340-345 46.2 38.8 31.1 44.3 23.9 16.1 1.1 600-610 52.9 34.1 29.9 740-750 58.8 46.8 28.8 45.4 22.2 17.2 0.8 210 10- 20 47.8 20.8 11.8 0.7 30-4 0 43.7 23.7 21.2 46.9 21.8 14.6 0.7 50- 70 45.9 27.3 25.0 47.7 23.0 14.0 0.7 90-110 47.0 37.3 26.7. 47.5 23.7 14.3 0.8 130-150 41.9 37.7 27.5 47.5 23.4 14.6 0.8 Profileso f theuppe rWanjar e catena -120- Profile Depth Particle size distribution (mass %) Density separations(%) Magnetic separates(%) in sand(005-0.106mm, (cm) gravel sandfractions (mm) in sand (0.05-0.106mm) d>2.89) >2.0 1-2 0.5-1.0 0.25-0.5 0.1-0.25 0.05-0.1 sand silt clay d<2.5 2.5-2.89 >2.89 s.m w.m n.m

Wan 1 0- 35 20.2 19.6 10.5 7 3 7 7 4.6 49.5 19.2 31.3 4.1 92.2 3.7 19.9 50.4 29.7 Wan 1A 0- 14 47.2 28.9 6.1 6 3 9 .1 3.1 54.4 12.8 32.8 . 30.3 58.8 10.9 Wan 2 0-16 32.2 54.8 10.9 5 7 5 5 1.9 78.9 6.7 14.4 6.6 90.9 2.5 . . 15- 30 41.5 68.4 10.3 3 0 2 6 1.2 85.5 4.1 10.4 8.6 88.1 3.2 37.8 34.4 30.9 55- 70 54.5 53.6 11.4 2 0 0 4 0.9 24.2 4.9 26.8 36.3 60.8 2.9 . . . 83- 98 42.7 67.6 6.9 0 5 0 .6 0.4 76.1 3.8 20.1 23.9 73.8 2.3 15.4 42.0 42.5 120-135 61.8 68.2 5.5 0 6 1 0 1.2 32.3 4.9 18.7 24.5 73.0 2.5 24.8 41.5 33.7 160-165 40.1 21.2 10.3 3 8 3 .3 7.7 42.6 35.2 22.2 31.3 65.0 3.7 0.6 12.3 87.1 Wan 3 10- 25 1.7 11.3 15.8 10 9 6 7 2.9 46.6 1.0 52.4 14.0 76.9 9.1 2.4 73.0 24.6 40- 55 1.5 6.7 10.2 7 5 5 2 1.5 31.0 7.3 61.7 6.3 80.4 13.3 2.9 73.6 23.4 70- 85 3.2 5.3 8.9 7 4 5 8 2.5 27.4 10.7 61.9 15.7 72.2 12.1 2.2 74.4 23.4 85-100 5.3 5.2 8.2 6 3 4 9 2.6 27.1 9.8 63.1 28.3 60.3 11.4 2.1 89.3 8.6 145-160 2.2 7.5 10.2 6 5 5 2 3.1 32.6 6.2 61.2 26.6 60.6 12.7 2.6 73.8 23.5 190-205 0.8 6.9 7.6 4 8 3 5 1.8 24.6 8.6 66.8 17.6 69.3 13.0 1.8 74.0 24.2 240-255 7.1 8.9 6.1 4 6 4 8 3.9 28.4 9.0 62.6 24.4 67.2 8.3 1.7 76.0 22.4 255-270 2.1 7.7 6.0 4 6 4 7 5.9 29.0 7.9 63.1 8.9 81.8 9.3 1.1 73.4 25.5 285-300 3.5 10.0 7.2 5 6 5 4 3.6 31.9 3.9 64.2 3.3 86.4 10.3 0.0 66.3 33.7 315-330 5.0 12.6 6.8 4 9 4 7 6.4 35.4 7.0 57.6 9.8 80.8 9.3 1.2 72.0 26.9 385-400 29.9 34.4 5.7 2 0 1 5 0.8 44.4 5.4 50.2 12.2 78.8 9.0 0.9 70.4 28.7 435-450 19.0 20.1 7.1 3 7 2 9 2.0 35.7 10.3 54.0 13.9 74.1 12.0 0.7 69.8 29.5 485-500 14.7 16.5 7.5 4 2 34 2.4 34 28.6 37.4 n.d. n.d. 8.8 0.8 74.4 24.8 535-550 9.6 12.8 12.7 6 4 5 0 1.3 38.3 21.3 40.4 21.6 66.7 9.4 0.7 66.0 33.3 635-650a 11.8 17.8 9.7 3 3 2 4 1.5 34.6 <-65.4-> 25.6 65.2 9.2 0.4 62.1 37.5 termite soil saprolite b 6.3 27.0 16.2 6 3 4 2 1.6 55.4 <-44.6-> n.d. n.d. 6.1 n.d. n.d. n.d. 735-750 5.1 7.8 11.1 8 3 6 5 2.8 36.6 27.1 36.3 18.3 71.8 9.0 0.9 65.1 34.0

Profile Depth Org.C pH Exchangeab] e bases (mmol/kg) CEC(NH4Ac) Base Bulk

(cm) (g/kg) H20 KCl iCa iMg K Na at pH 7.0 saturation density mmol/kg ( % ) (g/kg)

Wan 1 0- 35 24.1 4.0 3 6 9 4 5 n.d. 189 9+ 970 Wan 1A 0- 14 n.d. 5.5 4 2 2 1 2 n.d. 123 4.1 . Wan 2 0- 16 27 4.6 3 9 22 6 4 n.d. 104 30.8 1490 15- 30 24 4.5 3 7 9 2 3 n.d. 92 15.2 1440 55- 70 n.d. 4.6 3 9 5 1 2 n.d. 57 14.0 n.d. 83- 98 4 5.0 3 8 6 3 2 n.d. 61 18.0 1510 120- 135 3 4.9 3 8 5 3 4 n.d. 74 16.2 160-165 6 5.0 4 1 10 10 2 n.d. 52 42.3 . Wan 3 0- 25 22.0 6.0 5 2 136 28 10 n.d. 175 99.4 1170 25- 55 18.0 5.2 4 3 63 15 4 n.d. 133 61.6 1210 55- 85 14.0 5.2 4 5 60 14 4 n.d. 144 62.9 1240 85- 100 18.0 5.3 4 5 62 14 3 n.d. 120 65.8 1170 145- 160 8.0 5.2 4 7 36 17 5 n.d. 87 66.7 190-205 5.0 5.5 5 1 40 19 8 n.d. 91 73.6 1240 240- 255 3.0 5.1 4 4 20 6 13 n.d. 85 45.9 255- 270 2.0 5.2 4 5 50 14 18 n.d. 80 100+ 285- 330 2.0 4.8 4 0 20 7 13 n.d. 82 49.1 485- 550 tr. 4.8 3 8 19 7 4 n.d. 47 70.0 730-740 tr. 4.8 3 8 22 6 3 n.d. 59 52.5 -121-

Depth Compositl onacc o rding toX - ray fluorescence Minerals per 100 transparent grains

in ve cy fine sand

(cm) Stronglymagnet ! c weaklymagn . d<2.5 d>2.89, non-magnetic Ml M2 Ilm Hm Ru Z Hm Dm Z A glass plant opal zircon Amphibole+py roxene

0-3 5 +++ - ++ +-H- tr - ++ ++ ++ -H-+ 22.1 4.1 78 15 0- 14 11 1 1 - +++ + - tr - - +++++ . . 50 n.d. 0- 16 10.4 2.6 n.d. n.d. 15- 30 ++ - ++ +++ tr + ++ +++ +++ + 8.6 2.0 68 17 55- 70 0 1.1 n.d. n.d 83- 98 ++ - ++ +++ + tr ++ -H-+ ++ tr 0 0 96 1 120-135 ++ - ++ ++<- + + ++ +++ ++ tr 0 0 96 1 160-165 ++ - ++ +++ + - ++ +++ ++ - tr 0 96 1 10-2 5 -H- + 11 1 1 ++ + tr - •m-H- + - 19.6 11.0 100 - 40- 55 + + 11 1 1 + + tr - +++++ + - 2.7 0.5 99 - 70- 85 + + 11 1 1 + + tr - +++++ + - 1.5 0.7 100 tr 85-100 + + 11 1 1 + + tr - +++++ + - 3.1 - 100 tr 145-160 + + 11 1 1 tr + tr - +++++ tr - 0 2.0 100 - V 190-205 + + 11 1 1 tr + tr - -H+++ tr - tr 0.5 100 - 240-255 + + 11 1 1 tr + - - +++++ tr - 0 tr 100 tr 255-270 ++ + 11 1 1 ++ + - - +++++ tr - 1.4 1.0 100 - 285-300 tr 1.8 100 - 315-330 ++ + +++ -H- + - - •H+H- tr - •0 0 100 tr 385-400 ++ - +++ ++ tr - - +++++ tr - tr 0 98 - 435-450 + - + - - - - -H-t-H- tr - 0.8 tr 100 - 485-500 ++ - ++ +++ - - - +•+++ tr - . 99 0.6 535-550 ++ - •H- +++ - - - -H++H- tr - tr tr 96 tr 635-650a 1.0 tr 96 2 termite soil saprollte b ++ - -H- +++ - - •m-H- tr - 0 0 87 - 735-750 -H- - ++ +++ - - - +++++ tr - tr tr 96 -

Profile Depth Moisture% , Massao ichemica lelement s in fine earth (ca] cu Lated as oxides] (cm) volumicmas sa t

2.0 3.0 4 .2 SIC Fe MnO MgO CaO Na 0 K Ti02 P 2 A12°3 !°3 2 1° 2 °5 Wan 1 0- 35 35.2 17.7 11 .0 67 12.5 4 .5 0 0 0.2 0.2 1.3 2 .2 0 5 0 1 Wan 1A 0- 14 • 79 10 3 .9 0 0 0.1 0.1 1.0 1 .6 0 2 0 1 Wan 2 0- 16 23.7 12.9 8 .8 79 1 9.3 3 .1 0 1 0.1 0.1 0.6 1. 6 0 4 0 1 15- 30 21.4 10.4 8 0 80 4 9.4 3 .2 tr 0.1 0.1 0.7 1 .4 0, 3 0 1 55- 70 n.d. n.d. n. i. 81 0 10.4 3 .6 0 0 0.1 tr 0.6 0 .5 0 3 0 1 83- 98 22.1 15.0 9 4 73 6 14.3 5 .1 tr 0.1 tr 0.2 0 .6 0 4 0 1 120-135 • • 78 4 11.2 4 .6 ti 0.1 tr 0.1 0 .7 0 4 0 1 160-165 • 52 2 14.6 21 .2 1 1 0.1 1.1 tr 0 9 0 5 0 1 Wan3 0- 25 34.3 25.1 18 0 60 8 14.9 8 .6 0 3 0.2 0.3 tr 0 9 1 1 0 2 25- 55 37.2 29.1 22 0 56 3 19.8 10 .5 0 3 0.1 0.3 tr 0 7 1 2 0 2 55- 85 35.5 31.5 27 6 52 9 20.7 11 .2 0 3 0.1 0.3 tr 0 7 1 3 0 2 85-100 42.6 35.7 25 2 56 9 20.0 10 .7 0 3 0.1 0.3 tr 0 7 1 2 0 2 145- 60 • 53 5 22.7 11 .5 0 1 0.1 0.1 tr 0 6 1 3 0 1 190-205 40.4 33.0 26 4 53 1 22.6 11 .8 0 2 0.2 0.2 tr 0 7 1 3 0 1 240-255 • • 54 7 22.4 11 .3 0 2 0.1 0.2 tr 0 7 1 3 0 1 255-270 • 52 6 22.8 11 .8 0 2 0.1 0.2 tr 0 7 0 9 0 1 285-330 55 1 21.7 11 .5 0 2 0.1 0.2 tr 0 6 0 9 0 1 485-550 • 59 9 17.5 11 .2 0 1 0.1 0.1 tr 0 9 0 7 0 1 730-740 . 63 5 17.4 10 .3 0 1 0.1 0.1 tr 1 0 0 7 0 1 -122-

Deepaugering s of theuppe rWanjar ecaten aan dcompositio n ofgranit ean ddiorite .

Deep Depth Particle sizedistributio n (mass %) magnetic separates mass%o fsan dwit h augering (cm) sand fractions (mm) (fraction0.05-0.42mm )d>2. 9i n fraction d> 2. 9 0.05-0.42mm 1-2.0 0.5-1.00.25-0. 50.1-0.2 50.05-0. 1san d siltcla y s.m w.m n.m

1 0-1 5 45.9 11.6 7.2 7.3 2.6 74.6 9.7 15.6 30.5 37.6 31.9 . 2 0-1 5 38.2 10.9 5.3 4.5 1.7 60.6 14.2 25.1 14.3 54.5 31.2 . 60- 65 25.5 5.3 2.8 3.0 1.5 38.1 17.2 44.7 2.5 60.8 36.7 3 0-1 5 20.3 11.7 5.2 4.5 3.2 44.9 15.3 39.9 26.5 43.0 30.5 . 60- 65 22.4 8.1 5.7 5.1 2.5 43.8 6.9 49.3 1.2 51.8 47.0 . 6 0-1 5 26.1 10.1 5.5 5.7 2.3 49.7 13.0 37.4 4.0 73.0 23.0 60- 65 18.7 9.2 4.6 3.7 1.4 37.6 36.0 26.4 0.9 75.4 23.7 3.2 165-170 14.2 7.1 4.4 4.2 1.4 31.3 7.5 61.1 0.4 74.2 25.4 . 245-250 23.6 6.0 3.0 3.2 1.9 37.7 9.1 53.2 0.6 70.2 29.2 . 350-355 33.3 8.2 3.4 3.0 2.0 49.9 8.3 41.9 3.0 74.6 22.4 5 0-1 5 12.3 8.2 7.1 7.0 3.0 37.6 14.0 48.5 0.1 79.8 19.2 . 60-6 5 10.8 7.6 5.1 4.4 2.0 29.9 9.6 60.6 0.4 82.8 16.8 3.6 120-125 7.5 7.6 5.0 4.3 1.6 26.0 19.4 54.6 0.3 77.5 22.2 . 160-170 7.5 7.7 5.4 4.6 1.3 26.5 7.7 65.8 0.2 73.1 26.7 3.9 245-270 11.7 5.6 3.8 4.8 1.5 27.4 9.0 63.6 0.1 75.6 24.3 . 345-350 12.1 6.1 3.9 3.9 2.3 28.3 34.5 37.2 0.2 78.1 21.7 4.7 390-395 15.4 5.8 3.2 2.8 1.4 28.6 10.4 60.9 0.6 63.7 35.8 3.3 440-445 16.4 8.5 4.2 3.4 2.2 34.7 10.0 55.4 0.5 89.6 9.9 . 4 0-1 5 16.4 11.7 7.1 5.5 2.5 43.2 12.4 44.4 0.8 74.7 24.5 1.7 60-6 5 9.8 10.1 5.9 3.3 0.6 29.7 6.5 63.9 0.7 64.7 34.6 3.4 120-125 13.7 7.1 4.8 3.7 0.3 29.6 37.1 33.1 0.4 87.2 12.4 3.9 165-170 10.3 6.2 3.8 4.0 2.1 26.4 8.5 65.2 0.4 71.5 28.2 3.9 245-250 11.2 5.8 3.6 3.4 2.1 26.1 9.1 64.8 0.4 69.6 30.0 3.9 350-355 14.1 7.1 4.0 3.0 1.7 29.9 11.6 58.4 0.4 80.2 19.4 4.1 390-395 23.6 9.2 3.6 2.8 1.9 41.1 7.6 51.4 0.7 66.7 32.7 0.7 7 0-1 5 5.3 8.7 7.4 5.9 2.4 29.7 11.0 59.4 0.5 84.0 15.5 1.7 60- 65 5.0 7.3 4.8 3.8 1.6 22.5 14.7 62.7 0.3 78.9 20.8 3.8 120-125 5.7 7.2 5.2 4.5 1.6 24.2 8.3 67.4 0.4 75.3 24.4 3.8 160-170 15.1 8.3 4.2 3.2 1.4 32.2 9.1 58.7 0.2 79.3 20.5 1.6 8 0-1 5 10.3 8.8 7.3 7.1 3.1 36.6 16.5 47.0 3.0 75.7 21.3 2.6 granite rock 30.3 58.8 10.9 3.8 diorite rock 2.3 11.0 86.7 43.7 -123-

in in in m

o ~* -i oo oo o\ —< CM in

-*-*{s|(NC"l

—i i-. CM

m m CM co o -H PI CM oo £ i in m in in m in

+ + + + + + + + + + + + + + + eMO^o^Oe^r^r— -TooP^-ï^fnOm co co CM <- -* —<*o

CM CM (M CM m CM CM CM CM CO CM

in m CM CM \0 CT* ~H CM

00 CM 00

H est CM *o co CM v£> CM CM co vO co CM CM ••£> -3- r-*

» -* •* -* •* -* "* ** •* ** -tf -* .* -* -i O C* - o - m m -«• in

in oo oo r— m as**

m m CM CO CO P> ^ co CM r* co CM in co -H CM CM CO ^H CM CM

CM m cn —I —I *-t CS -124-

ON CM 0\ -* m m m

en m ci

en CM ut in CM CM ~-> o 4J w

CO \Û Kl H Ifl _, rsl —i -^ —1 O

r*. vo r-- oo oo

CM O -H CM

o> •40 CO M? O m M3 m co en en en m tO ^ \D o CM f^. •43 co o m a\ m m a* m O o> -» co -* co CS ey* o O CO o» •J3 co en co r- -Ï m - - "* •^ -* •* \D m co - en «£> \o CM in m CJ\ m O m f- - « - O in - o (M «M (M CM CM CM (O ~H

co co co m co

m *£> r^ m m ao m

o o co co inuococOininrM-* —i m r-cocMmcocMco*-^

•—icooôco-HQOCMm -H CM CM -125-

Appendix2 Method s Morphological and chemical methods

Morphologyo fth eprofile swa sdescribe daccordin gt oth estandard san d terminologyo fFA O (1967).Laborator yanalysi sar eliste dbelow : (1)Particl esize :organi cmatte rwa sdestroye dwit hhydroge nperoxid e (volumefractio n0.3) ;carbonate san diro ncoating swer eremove d withHC l(cone .2 mol/1) .Th esampl ewa sdilute dan dsiphone dthre e times,we tsieve dwit ha 0.05m msiev et oseparat esand .Th eres twa s collectedi na sedimentatio ncylinde ran ddisperse dwit hsodiu m pyrophosphate (cone.0.12 0mol/1) .Afte rshaking ,sil t(<0.05mm )an d clay(<0.002mm )wer epipette doff .Afte rdryin gth esan dwa ssieve d intofraction s2.0-1.0 ,1.0-0.50 ,0.05-0.25 ,0.25-0.1 0an d0.10 - 0.05mm. (2)pH-H 20an dKCl :soi lwa sintermittentl y shakenfo r2 d wit h distilledwate ran dpotassiu mchlorid e (cone.1 mol/1 )i na volum e ratioo f1:2.5 .Th ep Ho fth esupernatan twa smeasure dwit ha combinedglas scalome lelectrode . (3)Mas sfractio no fcarbon :Walkle yan dBlac kmetho d (Black,1965 ,p . 1372-1376)wit h1.4 3a scorrectio nfactor . (4)Conten to fexchangeabl ecations :leac hsoi lwit hammoniu macetat e (cone.1 mol/1 )o fp H7.0 .Estimat eNa ,K an dC ab yemissio n spectrometryan dwit hadditio no flanthanu mchlorid efo rcalcium . EstimateM gb yatomi cabsorptio nspectrometry . (5)Cation-exchang ecapacity :Afte rleachin gou texchangeabl ecation s (seemetho d 4),was hth esoi lwit haqueou sethano l (vol.fractio n 0.95)an dpercolat ewit hacidifie dNaCl .Steam-disti lof fth e ammoniaan dtitrat eagains tHC l (cone.1 0mmol/1 )(Houb ae tal . 1979). (6)Tota lchemica lcomposition :X-ra y fluorescencespectrometr y (Si,Al , Fe,Mg ,Ca ,Mn ,K ,Na ,P ,Ba) ;afte rdestructio no fth esampl ewit h HFan dH 2S0,,N awa sdetermine dspectrometricall y(Begheijn ,1980) . (7)Zinc :air-drie dfin eeart hwa sgrine di na mortar .Th epowdere d samplemateria lwa sshake nwit hHC l(cone .6 mol/1) ;zin cwa s determinedwit ha spectrophotomete r(Departmen to fmine san d geology,Nairobi ,Kenya) .

Minevalogical methods

Aweighte dsampl eo fsan d(afte rtreatmen twit h Rn®?ant *^CB :se e method 1above )preferrabl yi nth e50-10 0à 105p msiz efractio nwa sseparate d inheav yan dligh tminera lfraction swit hbromofor m (d>2.89). For furtherseparatio no fth eligh tfraction ,a mixtur eo fbromofor man d decaline (d= 2.5 )wa sused .

Heavymineral swer efurthe rseparate dwit ha han dmagne ti na stronglymagneti c (s.m.),a weakl ymagneti c(w.m. )an dnon-magneti c fraction(n.m.) .Th estrongl ymagneti cmineral swer eseparate dmovin ga handmagnetrepeatedl yunde ra thic kpiec eo fpaper ,covere dwit hheav y mineralgrains .Remainin gmineral swer ethe nrepeatedl y touchedwit hth e samehandmagnet .Grain sadherin g toi twer eremove dfro mi twit ha brus h andcollected .Thi swa sth eweakl ymagneti cfraction .Th eremainin g mineralswer ecalle dnon-magnetic .Al lfraction swer eweighed .

Compositiono fstrongl yan dweakl ymagneti cfraction swer e -126-

estimated by X-ray fluorescence analysis of powdered samples.Grai n mounts were made in Canada resin of the fractions with d<2.5, with 2.5

Methods to estimate the fraction of Al from volcanic ash in mixed profiles. Estimates according to the magnetite-II content .Principl eo fth e method: themas s ratio ofmagnetite-I I toA l ina sample of soil from volcanic ash is used toestimat e the fraction ofA l ina soil from volcanic ash and local rock (mixed soil). For that purpose also the mass ratio ofmagnetite-I I in the fine earth of asampl e ofmixe d soil to its total Alconten t should be known.

Calculation procedure:mas s fraction ofA120 3 derived from volcanic ash in a sample of fine earth from amixe d soil is the result of the following calculation: ^- x-2 -i nwhic hA =mas s fraction of magnetite-II in the 50-100(imsan d fraction asmas s fraction of fine earth. Iti s calculated as follows: (mass fraction of 50-100(jm fraction)x (mass fraction of stronglymagneti c fraction)x (mass fraction of strongly magnetic fraction)x (mass fraction ofmagnetite-I I in strongly magnetic fraction): Al ismas s fraction ofmagnetite-I I in sample of fine earth from volcanic ash A2, same asA l but in a sample of fine earth from amixe d soil B ismas s fraction of Al^Oo ina sample of fine earth from volcanic ash C is as B,bu t ina sample of fine earth from amixe d soil. Al This calculation procedure was used toestimat e-^ -value s of samples from profile S3.Th e average values of 5 samples ranged from (31-62)x 10 .Wit h those values and with the Can d A2value s of samples of profileWa n 3,averag e fractions of Al from volcanic ash in samples of profile Wan 3coul d be approximated.

Estimates according to the molar ratio of Al to Fe

Calculation procedure:A + x (B-A)= C A =mola r ratio of Al to Fei nvolcani c ash B =mola r ratio ofA l toF e in thecountr y rock C =mola r ratio of Al to Fe in sample of mixed soil x = fraction ofA l derived from country rock 1-x= fraction of Alderive d from volcanic ash -127-

Appendix 3 Plasmlc fabric in theprofile s studied

The dominant plasmic fabric in thewel l drained profiles Wan 2,3 and 9wa s insepic,apar t from the topsoil,whic h wasmor e undulic. In profiles Wan 4an d 5th e insepic fabric changed at depth ina masepic fabric, especially in thegreyis h part of thematrix . The masepic fabric,whic h is indicative of swelling and shrinking,wa s stronger expressed in the subsoil ofWa n 5tha n in thesubsoi lo f Wan 4,whil e the subsoil of M1E was very strongly masepic.

Appendix 4 Fission trackdat a (Source "Stichting voor Isotopen Geologisch Onderzoek"a t Amsterdam).

Profile Sample Ml neral T depth (era) 6 6 * 66 6f, 6 xlO , xl0 2 xl0 2 xlO xlO * t/cn. t/cm t/cm yt yr grains r,s U ppm

S3 50- 79 zl rcon 0.19 16.1 0.99 0.7 0.2 10 0.83 470 (58) (2462) (2061)

S3 600-610 " 0.16 12.3 " 0.8 0.2 12 0.38 359 (47) (1751)

S3 740-750 " 0.17 13.1 " 0.8 0.2 9 0.38 382 (62) (2313) 11 131/4 10- 20 " 0.09 12.7 0.5 0.2 11 0.23 344 (30) (1928)

131/4 80- 90 " 0.14 12.0 " 0.7 0.2 10 0.7 351 (52) (2176)

Standard » 0.43 0.97 " 26.7 1.2 6 0.99 283

1 l 235 238 2 Note: \d = 1.55 E-10 yr"" ; \f = 7.03 E-17 yr ; U/ U = 0.00725; a= 580.2 E-24 cm the numbers in the brackets refer to the tracks counted the uranium content is a biased estimate of the actual sample value for the standard zircon an age of 28.4 + 2.3 Ma is recommended -128-

Abaturov,B.D. , 1972.Th erol eo f burrowinganimal s inth etranspor t ofminera l substances insoil .Pedobiologi a 12:261-266 .

Ab'Saber,A.N. , 1982.Th epaleoclimat ean d paleoecology of BrazilianAmazonia .41-5 9 in:Biologica l diversification in theTropics ,Prance ,G.T . (ed.). ColombiaUniv .Press ,Ne wYork .

Alvarado,A. ,C.W .Beris han dF .Peralta , 1981.Lea f cutterAn t (Atta cephalotes).Influenc eo n themorpholog y of Andepts inCost a Rica,Soi l Sei. Soc.Am .J . 45:790-793 .

Arshad,M.A. , 1981.Physica lan d chemicalpropertie s of termitemound s of twospecie so fMacroterme s (Isoptera, Termitidae)an d thesurroundin g soils of thesemi-ari d savanna ofKenya .Soi lScienc e B2:161-174 .

Arshad,M.A. , 1982. Influence of the termiteMacroterme smichaelsen i (Sjöst)o nsoi lfertilit y andvegetatio n ina semi-arid savannah ecosystem.Agro-ecosystem s 8:47-58 .

Bachelier,G. , 1978.L a faunede s sols;so nécologi ee tso naction .Documentatio n techniqueno .38 ,ORSTOM ,Paris . 391pp .

Bal, L., 1973 .Micromorphologica lanalysi s of soils.Soi l Survey papersno .6 . SoilSurve y Inst.Wageningen ,th e Netherlands.17 4pp .

Bennema,J. , 1965.Geologi e enBodemkunde .Geologi e enMijnbou w44 :197-207 .

Bennema,J. , 1982.Aci d soils of thehumi d tropics ofSout hAmerica ,wit h special reference to thewel l drained soils onol dalluvia lsediments .Proceeding so f theWorksho p on theManagemen t oflo w fertilityaci dsoil so f theAmerica nhumi d tropics,hel d inParamaribo ,Suriname ,23-2 6nov . 1981eds .J.F . Wienkan d H.A.d eWit , IICA,Sa nJosé ,Cost aRica .p .105-125 .

Begheijn,L .Th. ,1980 .Method s of chemicalanalysi s forsoil san dwaters .Dept .Soil . Sei.Geol. ,Agric .Univ . Wageningen. 100pp .

Bigarella,J.J. , M.R.Mousinh oan d J.X,D aSilva ,1965 .Processe s andenvironment s of theBrazilia nQuaternary . Imprensad oUniversidad e deParana ,Curitiba ,Brasil .7 1pp .

Black,CA. , 1965.Method s of soilanalysis .Agronom y No.9 .America n Society ofAgronom y Inc.,Madison , Wisconsin,U.S.A. , 2vol .157 2pp .

Bouillon,A. , 1970.Termite s of theEthiopia nRegio ni n 'Biology ofTermites * (K.Krishn aan d F.M. Weesner, eds.). Vol. 2:153-280 .

Boyer,P. , 1955.Première sétude spédologique s et bactériologiques des termitières.C.r .hebd .Séanc eAcad .Sei . Paris 240:569-571 .

Boyer,P. , 1956a.Actio nde sTermite s constructeurs surcertain s solsd'Afriqu e tropicale.6 e Congr,Intern .Sc . Sol.,Paris ,vol .C III:95-103 .

Boyer,P. , 1956b.Etud epedologiqu e de la repartitias etd udosag ede s bases totalesdan s lesmatériau xd e Bellicositermesnatalensi s (de Hav.). Ibid.242 :801-803 .

un, H.M.H.,1980 .Agroclimati czone s ofKenya .Pape rpres e Soc. ofEas t Africa,Arusha .Octobe r 1980.Keny aSoi l Survey,Nairobi .1 2pp .

Brewer,R. , 1976.Fabri can dminera lanalysi s of soils.R.E .Kriege r Publishing Company.Huntington ,Ne wYork .48 2 pp.

Brussaard,L. , 1983.Mestkever smake nd ekringloo p rond.Landbouwkundi g Tijdschrift 95:27-32 .

Butzer,K.W. ,G.L . Isaac,J.L . Richardsonan d C.Washbourn-Kamau , 1972.Radiocarbo n dating of EastAfrica nLak e levels. Science 175,no .4026 :1069-1072 .

Buxton,R.D. , 1981.Change s inth ecompositio n andactivitie s of termitecommunitie s inrelatio n tochangin g rainfall.Oecologi a (Berl)51 :371-378 .

Carson,M.A .an dM.J .Kirkby , 1972.Hillslope ,for man d process.Cambridg e UniversityPress .47 5pp .

Cloud,P. ,L.B .Gustafso nan d J.A.L. Watson, 1980.Th ework s of living socialinsect s aspseudofossil s and theag e of theoldes tknow nMetazoa , Science 210:1013-1015 -

Coaton,W.G.H. ,1958 .Th ehodotermiti dharveste r termiteso fSout hAfrica .Scienc eBulletin ,Entomolog y Series, 375n o 43).Dep .o fAgr .Unio no f SouthAfrica .11 2pp .

Collinet,J. , 1969.Contributio nà l'étud ede s 'stonelines 'dan sl a régiond uMoye nOgoou ê (Gabon).Seri e PédologieORSTO M 7:0-42 .

Comaby, B.W.an d J.E.Krebs , 1975.Propiedade s quimicasd elo ssuelo s ferruginosos denido sd e termites ("comejen)y d e lasarea s adyacentes enVenezuela .Communicacione sTurrialb a 25:202-204 .

Darwin,C , 1881.Th eformatio n ofvegetabl emoul d through theactio no fworms ,wit hobservation s of theirhabits . -129-

Murray,London .32 6pp .

Dijk,H ,van ,1961 .Samenstellin g eneigenschappe nva norganisch e stof ind egrond .In :Bodemkund eMin .va n Landbouwe nVisserij .Th eHague .59-66 .

Dorst,J . and P.Dandelot ,1970 .A fiel d guide toth elarge rmammal so fAfrica ,Collin sSt .Jame' sPlace ,London . 287pp .

Drummond,H. , 1887.O n the termitea s the tropicalanalogu e of theeart hworm .Proc .Roy .Soc .Edinburg h 13:137 - 146,

FAO, 1967.Guideline s forsoi l profile descriptions.Foo d andAgr .Org .Rome .5 3pp .

FAO/Unesco, 1974.Soi lMa po f theWorld .Paris .

Ferrar,P. , 1982.Termite s ofa SouthAfrica nSavann a IV.Subterranea n Populations,Mas sdetermination s and biomass estimations.Oecologi a(Berl )52 :147-151 .

Fölster,H. , 1969.Slop edevelopmen t inS.W . Nigeria during latePleistocen ean dHolocene .GÖttinge r Bodenkundliche Berichte 10:3-56 .

Forester,J.W. ,1961 .Industria ldynamics .M.I.T .press ,Cambridg eMass. ,U.S.A .

Larov 144.

Grasse,P.P. , 1950.Termite se tsol s tropicaux.Revu e Internationale duBotaniqu eAppliquée :549-554 .

Grove,A.T. , 1969.Lan d formsan d climatic change inth eKalahar i andNgamiland .Geogr .J . 135:191-212 .

Grzimek, 1970.He t levende rdieren .Vol .X e nXI .He tSpectrum ,Utrecht/Antwerpen .

Hayes,M.B.H. ,1983 .Darwin' svegetabl emoul d and somemoder nconcept s ofhumu s structurean d soilaggregation . In: Satchell I.E. (ed.), Earthworm ecology.Chapma nan dHall ,Londo nan d NewYork ,pp .19-35 .

Hammen,T .va nder , 1974.Th epleistocen e changes ofvegetatio nan d climate intropica l SouthAmerica .J . Biogeogr.Vol .1 :3-26 .

Heinzelin,J .de , 1955.Observation s surl agenès e desnappe sd egravat sdan s lessol s tropicaux.I.N.E.A.C. ,Ser . Se.64 .

Hesse,P.R. , 1955.A chemica lan d physical study of thesoil s of termitemound s inEas tAfrica .J . Ecol. 43;449 - 461.

Hesse,P.R. , 1953.A chemica lan d physical studyo f thesoil s of termitemound s inEas tAfrica .Ph .D .Thesis, LondonUniv .

Hoeksema,K.J .an dEdelman ,C.H. , 1960.Th e role ofbiologica lhomogenizatio nI nth eformatio n ofGray-Brow n Podzolic Soils.7 th Intern.Congr .Soi l Sei.Trans .(Madison )4 :402-405 .

Hole, F.D., 1981.Effect so f animals onsoil .Geoderm a 25:75-112 .

Hoogerkamp,M. ,H . Rogaaran d H.J.P.Eijsackers ,1983 .Effec to fearthworm s ongrasslan d onrecentl y reclaimed poldersoil s inth eNetherlands .In :Satchel l I.E. (ed.), Earthworm Ecology.Chapma nan d Hall,Londo nan dNe w York,pp .85-107 .

Huddieston,M.J. , 1951.Geolog y of theKIsi iDistrict ,Rep .Geol . Surv.Keny a18 .

Imeson,A.C. ,1976 .Som eeffect s of burrowing animals on slopeprocesse s inth eLuxembour g Ardennes.Par t 1.Th e excavationo f animalmound s inexperimenta lplots .Geografisk aAnnale r 58A:115-125 .

Ireland,H.A. , C.F.S. Sharpean d D.H. Eargle,1939 .Principle s of gully erosion in thePiedmon t of SouthCarolina , U.S. Dep.Agr .Tech .Bull .633 .

Kalshoven,L.G.E. ,1936 .Onz ekenni sva nd eJavaansch e termieten.Handelin g 7,Nederland s IndischNatuur-Wetensch . Congr.:427-435 .

Kalshoven,L.G.E. ,1941 .Invloe d vand e localemacroscopisch e fauna,e nme t nameva nd e termieteno pd e vruchtbaarheid vand ebodem .Tecton a 34:568-581 .

Kendall,R.L. , 1969.A necologica lhistor y of theLak eVictori aBasin .Ecol .Monogr .39 :121-176 .

Krumbein,W.C .an d F.J. Pettijohn, 1938.Manua lo f sedimentary petrography. Appleton- centur y - Crafts,Inc .Ne w York.54 9pp .

Kubiena,W.L. , 1948.Entwicklungslehr e desBodens .Springer ,Wien .21 5pp .

Kubiena,W.L. ,1955 .Anima lactivit y insoil s asa decisiv e factor inestablishmen t ofhumu s forms.In :Kevan ,D . KeithM c E. (ed.), SoilZoology .Butterworths ,London ,pp .73-82 . -130-

Laatsch,W .von , 1957.Dynami kde rMitte leuropäische nMineralböden .4 Aufl .Theodo rStenkopff .Dresde nu Leipzig , Germany.28 0pp .

LaFage ,J.P .an dW.L .Nutting , 1978.Nutrien t dynamicso f termites.In :Brian ,M.V . (ed.),Productio necolog yo f antsan d termites,Cambridg eUniversit y Press.,pp .165-232 .

Laporte,G. , 1962.Reconnaissanc e pedologique lelon gd e lavoi e ferrée.COMILOG .Burea ude s rechuchesgéologique s etminiers ,Libreville ,6 5A 9 .

Lavelle,P. , 1983.Th esoi lfaun ao f tropicalsavannas .I ITh eearthworms .In :F .Bourlièr e (ed.): Tropical Savannas,Elsevier ,pp.485-504 .

Lee,K.E. ,1983 .Earthworm s of tropical regions,som easpect s of theirecolog y and relationshipswit hsoils .In : Satchell,J.E .ed .Earthwor m Ecology. Chapmanan d Hall,Londo nan d NewYork ,pp .179-195 .

Lee,K.E .an d T.G.Wood , 1971a.Termite s and Soils.Academi c Press,London .25 0pp .

Lee,K.E .an d T.G. Wood, 1971b.Physica l andchemica l effects onsoil s of someAustralia n termitesan d their pedologicalsignificance .Pedobiologi a 11:376-409 .

Lepage,M.G. , 1981.L'impac t despopulation s récoltantes de üa<\>&oï£Q{za{nJinerEüwev i(1 )(Sjostedt )(Isoptera , Macrotermitinae)dan s unécosystèm e semi-aride (Kajiado- KENYA), I- L'activit é de récoltee tso n déterminisme.Insecte sSociaux ,28 :29 7- 308 .

Levèque,A. , 1969.L eproblèm ede s solsa nappe sd egravat sa uTogo .Cah .ORSTO M Sêr.Pédol .VII :43-69 .

Livingstone,D.A. , 1962.Ag eo fdéglaciatio ni nth eRuwenzor i range.Uganda .Natur e 194:859-860 .

Livingstone,D.A. , 1975.Lat equaternar y climatic change inAfrica .Annua lRevie wo fecolog y and systematicsV6 : 249-280.

Lunt,H.A .an d G.M.Jacobson , 1944.Th echemica lcompositio n ofearthwor m casts.Soi l Sei.58 :367-375 .

Madge,D.S. , 1969.Fiel dan d laboratory studies on theactivitie s of twospecie s of tropicalearthworms . Pedobiologia 9:188-214 .

Meester,T .de , 1971.Highl y calcareous lacustrine soils inth eGrea t KonyaBasin ,Turkey .Agr .Res .Repor t752 , Pudoc,Wageningen , theNetherlands .16 9pp .

Miedema,R .an dW .va nVuure , 1977.Termit emound s inSierr aLeone .J . of SoilScienc e 28:112-124 .

Mielke,H.W. , 1978.Termitari a and shifting cultivation:th e dynamic role of thetermit e insoil s of tropicalwet - dryAfrica .Tropica lEcolog y vol.1 9no .1 :117-122 .

Milne,G. , 1935.Som esuggeste d unitso f classificationan d mapping,particularl y forEas t African Soils.Soi l Research 4:183-198 .

Ministry ofAgriculture , 1975.Agricultur e inKisi iDistrict .Min .o fAgr. ,Kisii ,Kenya .

Moeyersons,J. , 1978.Th ebehaviou r ofstone san d stone implements,burie d inconsolidatin g and creeping Kalahari sands.Eart hsurfac e Processes,3 :115-128 .

Mohr,E.C.J. ,F.A .va nBare nan d J.va nSchuylenborgh , 1972.Tropica lsoils ,a comprehensive study of their genesis,3 r edition.Mouton-Ichtia rBaru-va nHoeve ,Th eHague .48 0pp .

Morrison,M.E .an dC.A . Hamilton, 1974.Vegetatio nan dclimat ei nupland so f southwesternUgand adurin g later pleistocene period.2 Fores t clearance and othervegetationa l changes inRukig ahighlands .Journa l ofecolog y V62. Nl, pi.

Nye,P.H. , 1955.Soi lformin g processes inth ehumi d tropics IV.Th eactio no f thesoi lfauna .J . Soil Sei.6 :51 - 83.

Ochieng,W.A. ,1974 .A pre-colonialhistor y of theGusl io fWester nKeny aC.A.D . 1500-1914.Eas tAfr .Literatur e Bureau,Nairobi .25 7pp .

Ogot,B.A. , 1967.Histor y ofSouther nLuo .Vol .I .Eas tAfr .Literatur eBureau ,Nairobi .

Pendleton,R.L . 1942.Importanc e of termites inmodifyin g certain Siamese Soils.J .Amer .Soc .Agron .34 .

Ploey,J . de, 1964.Nappe sd egravat s etcouverture s argilosableusesa uBas-Congo ;leu rgenès ee t l'action des termites.In :Bouillon ,A (ed),Etude s sur les termitesAfricains .299-314 .Leopoldvill eUniv. ,Leopoldville .

Pomeroy,D.E. , 1976a.Som eeffect s ofmoun dbuildin g termiteso nsoil s inUganda .J .o fSoi lSei .27 :377-394 .

Pomeroy,D.E. , 1976b.Studie s ona population of large termitemound s inUganda .Ecologica l Entomology 1:49-61 .

Robinson,J.B.D. ,1958 .Som e chemical characteristics of termitesoil s inKeny acoffe e fields.J . SoilSei .9 : 58- 65. -131-

Rode,A.A. , 1962.Soi lScience ,translate d fromRussian .Israë lProgra m forScientifi cTranslations ,Jerusalem . 517pp .

Ruhe,R.V. , 1959.Ston e lines insoils .Soi l Sei.87 :223-231 .

Russell, 1910.Th eeffec t ofearthworm s insoi lproductiveness .J .Agric . Sei.Camb .2 :245-257 .

Sands,W.A. , 1965.Termit e distribution inman-modifie dhabita t inWes tAfric awit h special reference tospecie s segregation inth egenu sTrinerviterroe s(Isopter aTermitidae ,Nasutitermitinae )J .Anim .Ecol .34 :557-571 .

Sarntheim,M. , 1978.San d deserts during glacialmaximu m and climatic optimum.Natur evol .272 :43-46 .

Scheltema,W. , 1964.He tKawfoeto e verschijnsel inSuriname .Diens tBodemkarterin gSurinam ean dM.Sc .thesi s lodged inth elibrar y of theDept .o f SoilScienc e andGeology .Agric .Univ .Wageningen ,th eNetherland s as Research Rep.No .37 .

Schuylenborgh,J .van ,S .Slage ran dA.G .Jongmans , 1970.O nsoi lgenesi si ntemperat ehumi d climate VIII.Th e formationo f a"Udalfic " Eutrochept.Neth .J . ofAgr .Sei .18 :207-214 .

ScientificAmerican , 1983.Soi l Se. Soc.o fAmerica nJourna l 46,1982 .5 :917-921 .

Scrickande,J.C .an d A.N.Pathak ,1951 .A comparativ e study of thephysico-chemica l characteristics of the castings ofdifferen t insects.India nJ .Agric .Sei .21 :401-407 .

Shackleton,R.M. , 1951.A contributio n toth egeolog yo f theKavirond oRif tValley .Quart .J .Geol .Soc .106 :345 - 392.

Sharpe,C.F.S. ,1938 .Landslide s and related phenomena.Colombi a Univ.Press .Ne wYork .

Slager,S. ,1966 .Morphologica l Studies of some Cultivated Soils.Agr .Res .Rep .670. ,Pudoc » Wageningen.Il lpp .

Soil Survey Staff, 1975.A basi c systemo f soilclassificatio nfo rmakin g andinterpretin g soilsurveys .Unite d StatesDep .o fAgric .Agr .Handboo k no.436 .Governmen t Printing Office,Washingto n D.C., U.S.A.

Stakman,W.P. ,G.A .Val k andG .va nde rHarst ,1969 .Determinatio no f soilmoistur e retentioncurves-Vol.1 . Sandboxapparatu s range pF0-2. 7 (19pp. )Vol .2 .Pressur eMembran eapparatu s range pF3.0-4. 2(1 0pp. )3 edition Inst,fo rLan d andWate rManagemen t Research,Wageningen , theNetherlands .

Street,F.A .an dA.T .Grove ,1976 .Environmenta lan d climatic implications of lateQuaternar y lake-level fluctuations inAfrica .Natur evol .261 :385-389 .

Swift,M.J . (ed),1984 .Soi lbiologica lprocesse s and tropical soilfertility :A proposalfo ra collaborative programme of research. Specialissu e 5Biolog y International.Th enew smagazin e of theIntern .Unio no f Biological Sciences (IUBS), 38pp .

UNESCC—UNEP-FAO, 1978.Natura lresource s research.Tropica l forest ecosystems,a stat e ofknowledg ereport .

Volubuev,V.R. , 1964.Ecolog y of Soils,translate d fromRussian . Israel program forscientifi c translations, Jerusalem. 260pp .

Wallwork,J.A. ,1970 .Th edistributio nan d diversity ofsoi lfauna .Acad .Press ,London ,Ne wYork .S .Fransisco , 355pp .

Wasawo,D.P.S .an d S.A.Visser , 1959.Swam pworm s and tussockmound s inth eswamp so fTeso ,Uganda .Eas tAfr . Agric.J.25 :86-90 .

Watson,J.P. ,1960 .Som eobservation s onsoi lhorizon s and insectactivit y ingranit esoils .Proc . 1 Fed. Sei. Congr.Rhodesi aan d Nyasaland,271-276 .

Watson,J.P. ,1974 .Termite s inrelatio n tosoi lformation ,groundwate ran dgeochemica lprospecting . Soilsan d Fert.5 :111-114 .

Watson,J.P. ,1976 .Th ecompositio no fmound so f thetermit eMacroterme s Falciger (Gerstäcker)o nsoi lderive d fromgranit e inthre e rainfall zones of Rhodesia.Journa l of SoilScienc e 27:495-503 .

Werff,P .va nder , 1978.Activit y of themes oan dmacrofaun a ina poorl y drainedvalley-botto m soili nKisi i district,Kenya .M . Sc.Thesi s lodged inth e library of theDep .o fSoi lScienc ean d Geology,Agr .Univ. , Wageningen, theNetherlands , 56pp .

Wielemaker,W.G. , 1984.Th e importance ofvariabl e charge,du et oamorphou s siliceous soilconstituent s insom e soils fromKenya .Geoderm a32 :9-20 .

Wielemaker,W.G .an d H.W. Boxera(eds.) , 1982.Soil s of theKisi iarea ,Kenya .Agr .Researc hRep .922 ,Pudoc , Wageningen,th eNetherlands ,20 8pp .an d 5maps .

Wielemaker,W.G .an d G.R. vanDijk ,1981 .Mappin g andcorrelatio n oferosio nsurfaces ,variousl y affectedb y -132-

volcanic ash,tectonisr aan d lavaflows .A cas estud y inSouthwester n Kenya.Zeitschrif t fürGeomorph .Neu e Folge 25:54-72 .

Wielemaker,W.G .an dT .Wakatsuki , 1984.Properties ,weatherin gan dclassificatio no f some soils formed in peralkalinevolcani cas hi nKenya .Geoderm a 32:21-44 .

Wit, H.A.de , 1978.Soil s and Grassland typeso f Serengeti Plains (Tanzania).Ph .D .Thesis -Dept .o f SoilScienc e and Geology,Agric .Univ .Wageningen .

Wood,T.G. ,1976 .Th e roleo f termites (Isoptera)i ndecompositio nprocesses .In :Anderson ,J.M . andA .Macfadye n (eds.). Therol eo f terrestrialan d aquatic organisms indecompositio n processes:145-168 .

Wood,T.G. , 1978.Foo dan d feedinghabit s of termites.In :Brian ,M.V . (ed.), Productionecolog y ofant san d termites. CambridgeUniv .Press ,pp .55-80 .

Wood,T.G .an dW.A .Sands ,1978 .Th erol eo f termites inecosystems .In :Brian ,M.V . (ed.), Productionecolog yo f antsan d termites.Cambridg eUniv .Press ,pp .245-292 .

Zimmerman,P.R, ,J.P .Greenberg ,S.O .Wandig aan d P.J. Crutzen, 1982.Termites :a potentially largesourc eo f atmospheric methane,carbo ndioxid ean dmolecula r hydrogen.Scienc e 218:563-565 .