The phosphorus and nitrogen nutrition of bambara groundnut (Vigna subterranea (L.) Verde.) in Botswana soils
An exploratory study Promotor: dri rM . Wessel emeritushoogleraa r in detropisch e plantenteelt
Co-promotor: dr ir W.G. Keltjens universitair hoofddocent bij het departement omgevingswetenschapppen yU/J09>? The phosphorus and nitrogen nutrition of bambara groundnut (Vigna subterranea (L.) Verde.) in Botswana soils An exploratory study Gaebewe Modirwa Ramolemana Proefschrift ter verkrijging van degraa d van doctor opgeza gva nd erecto r magnificus van deLandbouwuniversitei t Wageningen, dr CM. Karssen, inhe t openbaar te verdedigen opdinsda g 30maar t 1999 des namiddags tevie r uur ind eAul a <^i Ramolemana, G.M. Thephosphoru s and nitrogen nutrition ofbambar agroundnu t (Vigna subterranea (L.)Verde. ) inBotswan a soils.A n exploratory study. Thesis Agricultural University Wageningen. - With ref. -Wit h summary inDutch . ISBN 90-5808-020-X Also published in the series Tropical Research Management Papers, nr 26, Wageningen Agricultural University, ISSN0926-9495 . Subject headings:bambar a groundnut, phosphorus,nitrogen , soil moisture, Botswana. Cover design: Ernst van Cleef BIBLIOTHEEK LANDBOUWUNIVERSITHT WAGENINGEN PROPOSITIONS It is a paradox that an indigenous African crop, which produces an almost comple tely balanced food, is one of the most drought tolerant, easy to cultivate crops which makes little demand on the soil, should be relegated in its own countries without being accorded any research attention and, worse still, earns the "accolade" of poor man's crop. Modified after Doku, V.E. (1997). Problems and prospects for the improvement of bambara groundnut. Proceedings of the international bambara groundnut symposium, University of Nottingham U.K. 23-25 July 1996. University of Nottingham, U.K. pp. 19-27. 2. The suitability of bambara groundnut as a low-input smallholders' crop, depends on a combination of a low P requirement, an effective N, fixing capacity, flexibility in the length of the growing period and shade tolerance. The soil moisture content is in Botswana the most important factor determining soil P availability and P uptake by bambara groundnut. This thesis Application of P fertilizer to bambara groundnut improves P uptake and growth only if rates exceed a certain level. This thesis Reduced early plant growth due to P stress has severe negative consequences for pod filling. P application at a later stage cannot amend this situation. This thesis Although Botswana soils are deficient in P, farmers growing rainfed bambara groundnut cannot be advised to use P fertilizers for this crop. Application of P in a crop rotation including bambara groundnut, however, might be a feasible option to improve the P nutrition of this crop. 7. The success of the bambara groundnut crop in the preceding year can be observed the following year. Linnemann, A.R. (1990). Cultivation of bambara groundnut (Vigna subterranea (L.) Verde.) in Western Province Zambia. Tropical Crops Communication 16, Dept.Trop.Crop Sci., Wageningen Agricultural University, pp.33. In further research on phosphorus fertilization of bambara groundnut, the role and contribution of mycorrhizas in P nutrition should get attention, especially because bambara groundnut is often grown in P-poor soils and its rooting is often poor. Broek,HAM. vanden (1998), MSc thesis,WAU. 9. The future of Botswana agriculture lies below the topsoil. 10. Our minds are like a parachute, and we crush when they do not open up. Abstract Plant speciesdiffe r inthei rminera l nutrientrequiremen t andnutrien t useefficiency . Information onth ephosphoru s (P)an dnitroge n(N )requiremen t ofth eleguminou scro pbambar a groundnut (Vignasubterranea (L.)Verde. ) isscant y and sometimes contradictory. Thisha s ledt o a study onminera l nutrition ofbambar a groundnut with the following objectives: (i)t o investigate the effects ofapplie d Pan dN o ngrowt han ddevelopmen t ofbambar agroundnu t inBotswan asoils , and (ii) to investigate the P and N uptake, shoot concentrations at different growth and development stages of bambara groundnut and howuptake , internal concentration and growth are related. The experimental programme consisted of six pot experiments on P and N ferilization, a large field experiment with P and irrigation as treatments, and a farm survey to studyth eP statuso fsoil san dbambar a groundnut plantsgrow n infarmers ' fields. In low P soils,bambar a groundnut responded to fertilizer Ponl y when itbecome s available to the seedlingwithi ntw oweek safte r sowing.I twa sshow ntha t soilmoistur erathe rtha nP inth e soil determined P availability to bambara groundnut. Analysis of the critical shoot P concentration showed that in traditional farms of Botswana bambara groundnut was always grown under suboptimal Plevels .Therefore , in lowP soils,P fertilizatio n will be beneficial to bambara groundnut only whenth e soil moisture content during thefirst t o weeks after sowing isnea rfield capacit y andth e fertilizer isplace d closet oth e seedling. Both under P-limiting and P non-limiting conditions, bambara groundnut seemed to be able tomee tit sN requiremen t byN 2 fixation, probably supplemented withminera lN from the soil. Therefore, N fertilization is not necessary and can sometimes have negative effects on plant growth and leadt o reduced seedyields . Keywords: bambara groundnut {Vignasubterranea), phosphorus, nitrogen, soil moisture, Botswana. Dedicated to the memory of my parents Gabofetane and Modirwa, and grandparents Thapiso and Seretsenye Letsididi. Preface Thisthesi si sa spin-of ffrom a nEuropea nUnio n (EU) funded project onbambar agroundnu t in which my home institution, Botswana College of Agriculture (BCA), and Wageningen Agricultural University (WAU) collaborated with other universities in Europe and Africa to evaluate the potential for bambara groundnut as a food crop in semi-arid Africa. The work reported in this thesis addresses one of the project objectives "to recommend suitable managementpractice st ostabilis eth eyield so fbambar agroundnu tunde rrainfe d conditions"an d in particular the mineral nutrition of bambara groundnut. I hope the understanding of the phosphorusan dnitroge nnutritio no fbambar agroundnu ta sfoun di nthi sthesis ,wil lcomplemen t the work done by other partners in the EU-programme on bambara groundnut to increase the yield of the crop. Although my research work addressed part of the objective of the EU- programme on bambara groundnut research, it was however funded separately by Botswana College of Agriculture and Wageningen Agricultural University. Duringth e struggle to getth eresearc h work done inBotswana , Igo ttremendou s support from members of staff of BCAan d the Department ofAgricultura l Research (DAR) ofth e Ministry of Agriculture which is highly appreciated. I want to recognise the contribution of Dr. J.S. Kiazolu(forme r SeniorLecture ri nSoi lScience ,BCA )a tth ever yinceptio no fth eresearc hwor k which laid a solid foundation for the subsequent research programme. From DAR I want to recognise and express my gratitude to Dr. G.S. Maphanyane for her day-to-day guidance in conducting experiments and writing reports. Your commitment and the criticism and encouragement you gave helped this project to come to ahapp y ending. Ithan k my promotor Prof.Dr .Mariu sWesse l forhi sguidenc e inpreparatio n ofth e research proposal andth ethesis . You did not only enrich my stay at WAU academically, but also showed me part of the The Netherlands and itsheritage . Ithan k myco-promoto r Dr. Willem Keltjens for hispatienc e and alwaysfindin g timet orea dan ddiscus sm ymanuscripts .I twa s gratifying tohav e someone like you as a supervisor. The assistence and friendship of Conny Almekinders, Martin Brink, Jaap Nelemans, Mary Omosa,Ren e Rietra, Willeke van Tintelen, Chrisva nUffelen , Liping Weng, collegues at both Departments of Agronomy and Soil Science & Plant Nutrition at WAU is appreciated. Manypeopl ewer e involved inth eexecutio no fexperiment s andth e farm survey reported inthi s thesis.Firs t ofal lI than km yfield assistan t AgisanyangNgwak o andth e casual labourers from theConten t Farmi nSebel ewh ower einvolve d inth eday-to-da y activities ofbot hth efield an d pot experiments. Withoutthei rcooperatio n Iwoul dno thav e completed theprojec t intime . The irrigition design forth efield experimen t wasth e ideao fMr . Otsoseng Oagile and the late Petso Motswasele, and it is appreciated. Ia m gateful to Mrs.T . Mpuisang and Mrs.T.K.Y . Morake who assisted in determining the amount of irrigation water applied and matters regarding pest control,respectively . Iwan tt othan k KeikanelweNtone ,Kgatlen gDistric tAgricultura l staff and their farmers for assistance and cooperation inth e farm survey. Therewa sth elaborator y workt ob edon ean dI than kal lthos ewh oprovide d technical assistance andcleanin g ofth e glassware.I especiall y wantt othan k B.Ditshane ,A .Fisher , Mr. Gopane,L . Letsine, B. Makoba, O. Mogobe, C. Moloi, S. Moroka, L. Mosarwe, B. Mosomong, R. Mothokodise and L. Patrick. The data collected was statistically analysed with the assistance ofJ .Makori , Dr. R.M. Sakia, B. Sebolai,th e lateNe o Seitshiro and others,an d Ia m grateful. Ia mals oa part-tim e livestockfarmer , andalway sI reques trelative st otak eresponsibilit y during myabsence .I than km yuncle sE.P.S .an d S.G.G.Letsidid i andm ysiste rDori sRamoleman a for theirpatienc ean dacceptin gextr aresponsibilitie st omaintai nm yhobby .I als othan km yin-laws , Lizzy and George Rakgoasi, for the support they gavem y wife during my absence. Finally, I want to express my appreciation to my family, Sue, Bojotlhe, Thato, Tshepo and Maipelo,fo r their patience, and especially the encouragement and support of mywife . All this wasthroug h the love of God.Than kyou . GR. Table of contents Preface 1.Genera l Introduction 1 1.1 Bambaragroundnu t 1 1.2 Research 3 1.3 Bambara groundnut inBotswan a 6 1.4 Objectives ofth e study 6 1.5 Outline ofth e thesis 8 2. Response ofbambar a groundnut (Vigna subterranea (L.)Verde. )t o phosphorus fertilization. 9 2.1 Introduction 9 2.2 Apreliminar y pot experiment 9 2.2.1 Materials and methods 9 2.2.2 Results 11 2.3 Afiel d experiment 13 2.3.1 Materials and methods 13 2.3.2 Results 14 2.4 Discussion 16 3. Factors affecting therespons e ofbambar a groundnut {Vignasubterranea (L.) Verde.)t ophosphoru s fertilization. 21 3.1 Acompariso n ofth e Puptak epattern s ofbambar a groundnut, maize and pigeon pea. 21 3.1.1 Introduction 21 3.1.2 Materials and methods 21 3.1.3 Results 22 3.1.4 Discussion 24 3.2 Growth and development ofbambar a groundnut asaffecte d bytim e of phosphorus application. 28 3.2.1 Introduction 28 3.2.2 Materials and methods 29 3.2.3 Results 30 3.2.4 Discussion 32 4.Phosphoru s levels in shoots of bambara groundnut (Vignasubterranea (L.)Verde) . 35 4.1 Critical shootphosphoru s concentration ofbambar a groundnut. Asan d culture approach. 35 4.1.1 Introduction 35 4.1.2Material s andmethod s 35 4.1.3 Results 37 4.1.4Discussio n 39 4.2 Shoot phosphorus levels inbambar a groundnut plants from farmers' fields in Botswana. 40 4.2.1 Introduction 40 4.2.2 Materials andmethod s 40 4.2.3 Results 41 4.2.4 Discussion 43 5.Respons e of bambara groundnut {Vigna subterranea (L.)Verde. )t o mineral nitrogen fertilizer application if grown under non-limiting phosphorus conditions. 45 5.1 Introduction 45 5.2 Materials and methods 46 5.3 Results 46 5.4 Discussion 52 6.Respons e ofbambar a groundnut {Vigna subterranea (L.)Verde. )t ominera l nitrogen fertilizer. Effect oftim e ofapplication . 55 6.1 Introduction 55 6.2 Materials and methods 55 6.3 Results 57 6.4 Discussion 61 7.Genera l discussion 65 References 69 Summary 75 Samenvatting 79 Curriculum vitae 85 Appendix. 87 Appendix 1.Map s of Botswana showing the location where the study was conducted (Gaborone) andth e annual rainfall pattern 87 Appendix 2.Long-ter m weather data Gaborone 88 Appendix 3.Th e distribution ofrainfal l andth etim e and quantity of irrigation from sowing onth e 18 th December 1995unti l final harvest onth e 22n d April 1996 ina field experiment at Sebele,Botswan a 89 Generalintroduction Chapter1 General introduction 1.1 Bambara groundnut Importance Inth epas tdecade sa combinatio n ofhig hyieldin g cultivars,fertilization , irrigation or adequate rainfall gavea spectacula r increase infoo d production inth ehumi dtropic so fAsia .Thi sdi d not happen inAfric a andespeciall y inth e semi-arid zonesfoo dproductio n didhardl y increase.Lac k ofwate r and income precluded the use ofth e socalle d Green Revolution Technology, and this means that the farmers have to rely on rainfed food crops which are adapted to the prevailing environment. One ofth e leguminous cropsadapte dt o awid erang eo fclimati c conditions,bu t especially semi-arid conditions isbambar a groundnut (Vignasubterranea (L.) Verde). Bambara groundnut is adapted to awid erang e of soilsan d itspopularit y with small farmers in Africa is attributed to the ability to produce yields onpoo r soils.Lik e other legumes it can fix atmospheric nitrogen through symbiosis with rhizobia. As a leguminous crop, bambara groundnut is useful in crop rotations because it may improve the nitrogen status of the soil. Mukurumbira (1985) found that bambara groundnut has ahighe r residual nitrogen effect than groundnut, maize or fallow. The crop hasals o been reported to bedrough t tolerant and ablet o produce some yield where other crops such asgroundnu t (Arachis hypogaea) fail (Linnemann andAzam-Ali, 1993). Bambara groundnut contributes to the livelihood of small farmers as a source of protein and income (Linnemann andAzam-Ali , 1993; Brink etal., 1996 ;Mulila-Mitt i and Kanenga, 1997; Sesay etal., 1997) .Th ecro pi smainl y grownfo rth e seeds,bu tth evegetativ epart sma yb euse d as fodder. An important advantage of bambara groundnut is that not only mature, but also immature seeds canb e consumed byhumans . Themajo r producers ofbambar a groundnut areNigeria ,Niger , Ghana and Ivory Coast, but the crop is also widely grown in eastern and southern Africa (Mazhani and Appa Rao, 1985; Linnemann andAzam-Ali , 1993).Th e annual worldproductio n isestimate d to be around 330 000 MT, with about 50% produced in West Africa (Coudert, 1982). The importance of bambara groundnut inBotswan a isdiscusse d insectio n 1.3. Chapter 1 Growth anddevelopment Reproductive development inbambar agroundnu t isknow nt ob einfluence d bytemperatur e and photoperiod. Inglasshous e studiesBrin k (1998) found that bambara groundnut selections from Botswana and Zimbabwe were mostly not sensitive to day length with regard to onset of flowering butal lwer estrongl y sensitivewit hregar dt oth eonse to fpodding .Th erat eo fprogres s from sowing to flowering was influenced bytemperatur e only, while the rate ofprogres s from sowingt opoddin g wasinfluence d bybot hphotoperio d and temperature. Itwa s also found that leafformatio n continuesfo ra longe rperio dunde rlon gda y length,resultin g ina greate r leafare a perplant .However ,tota ldr yweigh twa sno tinfluence d byda y lengthbu tth epercentag e of dry matteraccumulate d inpo d structures was influenced indirectly by the effect ofphotoperio d on podding. For bambara groundnut selections from Botswana ('DipC94' and 'GabC92'), it was found that under day lengths of 12.0t o 14.0hour s day"1i ttake s about 45 days from sowing to flowering andabou t 31day sfrom flowerin g topoddin g whenth e averagemonthl y temperatures vary between 26.8-32.5 °C and 13.2-19.3 °C during day and night, respectively. The onset of podding isth e most important event in bambara groundnut development asi t coincides witha major shift in assimilate partitioning, mainly directed towards pod growth. The photoperiod difference between locations in Botswana was found not to be large enough to merit using selectionswit h different photoperiod sensitivities for different locations(Brink , 1998).Th e time from sowing to maturity of bambara groundnut plants inth efield i sreporte d to range from 90 to 170day s (Linnemann and Azam-Ali, 1993). Inglasshous e studiesa tNottingha m University, soilmoistur edefici t wasfoun d todecreas eplan t growth,bu tdi dno taffec t thetim et oflowering . Theonse to fpoddin gwa sdelaye d inth estresse d plants, and the dry matter/water use ratio varied from 1.8 to 3 g kg"1 for irrigated and low moisturetreatments ,respectively . Dry matter partitioning to roots was found to be greater than 40%durin g early growthan dwa sconsistentl y greater for the lowmoistur e treatments. Harvest indiceswer eextremel y small under lowmoistur e conditions with amea n of0.03 , compared to 0.47 under irrigation (Collinson etal, 1996;Anon. , 1997). AtWageninge n Agricultural University inglasshous e studies,bambar a groundnut plants grown inhydroponic s with high P-supply continued vegetative growth longertha n plants with lowP - supply.Lo wP-suppl y plants invested morei nthei rroo t system thanplant swit h non-limiting P- supply, leading to shoot-root ratios of 2.6-3.0 in low-Pplants , compared to ratios of 4.1t o 7.3 for high-Pplant s (Anon., 1997). Generalintroduction 1.2 Research 1.2.1 General Accordingt oRachi ean dSilvestr e(1977) ,th emos t importantlegum ecrop si nAfric a arecowpe a (Vigna unguiculata), groundnut (Arachishypogea) and bambara groundnut. Seed yields of bambaragroundnu t inAfric a areo naverag e 650-850 kgha" 1.Compare d toth e othertw olegum e crops (cowpea and groundnut), bambara groundnut has received very little research attention althoughi tha sa goo dpotentia l asa hard ycro p(Linnemann , 1994).However , recently bambara groundnutbenefite d from aresearc hprogramm e"Evaluatin gth epotentia lfo rbambar agroundnu t asa foo d cropi nsemi-ari d Africa" (Anon., 1997).Th eproject ,whic hwa sfunde d byth eEU ,wa s acollaboratio n between European (University ofNottingham-Unite d Kingdoman d Wageningen Agricultural University-The Netherlands) and African (University of Botswana-Botswana, University of Sierra Leone-Sierra Leone and Sokoine University of Agriculture-Tanzania) countries. Some ofth eprojec t objectives wereth e following: 1. To identify suitable agro-ecological regions, seasons and agronomic practices for the cultivation of bambara groundnut genotypes inBotswana , Sierra Leone and Tanzania. 2. To predict yields of bambara groundnut genotypes in these zones and the sensitivity ofyiel d to genetic and environmental factors. 3. Torecommen d suitable management practicest o stabilize the yields of bambara groundnut inrainfe d environments. 4. To establish general principles underlying the growth and development of bambara groundnut inrainfe d environments. Experimental evidencefro m theprojec t onbambar agroundnu t indicatedtha tbambar agroundnu t iscapabl e ofattainin g significantly greateryield stha nthos eachieve d by farmers. In controlled- environment glasshouse experiments, yields of moretha n 3.5 t ha"1wer e achieved under non- limiting soil moisture. Under rainfed conditions,field crop s achieved yields in excess of 3.0 t ha1. Because of the effect of day length on podding, sowing date was identified as having a crucial influence onfinal po d yield (Anon., 1997). 1.2.2 Fertilizer studies Attempts havebee n madet o increasebambar a groundnut yield through application of nutrients andimprovemen t ofth enitroge n fixing capacity. Chapter 1 Phosphorus Sandy soils in Africa areusuall y deficient in nutrients such asphosphorus , and these soils are predominantly usedfo rbambar agroundnut . Researchconducte d inMalawi ,Nigeria , Swaziland, Zambiaha sfaile d to showan yrespons eo fbambar agroundnu t toth e application of phosphorus fertilizers at rates up to 147 kg P ha"', and sometimes yields were even depressed by P fertilization (Anon., 1975;Cumberland , 1978;Anon. , 1979;Anon. , 1980;Nnad i etal., 1981 ; Musonda, 1988).However , Tanimuan dYayoc k(1990 )reporte d inNigeri ai na fiel d experiment withdifferen t combinations ofphosphorus ,potassiu m andnitroge n asignificantl y higher yield from 22 kg P ha"1tha n at other phosphorus levels. Wassermann etal., (1984) in South Africa reported an increase intota l DMa ta P rat e of 30k g Pha' 1,bu t seed yield was not increased. In a hydroponic experiment at Wageningen Agricultural University, bambara groundnut has shown P-toxicity symptoms in anutrien t solution where other crops were grown without any problem (Kamstra, 1995).However , itshoul d benote dtha tnutrien t uptake ina nutrien t solution may be different from that under soil conditions whereth e uptake isaffecte d by more factors. Nitrogen The response to nitrogen application is generally poor. For example, in Nigeria a starter applicationrat eo f 20k gN ha" ' did not increase yield (Nnadi etal., 1976)o r even decreased it, although the effect was not significant (Tanimu and Yayock, 1990). Increasing the level of nitrogenu pt o 140k gN ha" 'di dno t significantly increase yield inMalaw i andZambi a (Anon., 1971, 1973;Lung u and Mbewe, 1986; Musonda, 1988), and in some cases yields decreased (Anon., 1970).However ,Dadso nan dBrook s(1989) ,i nTogo-Wes tAfrica , reported a significant increasei ntota ldr yweigh tan dsee dyiel d from atreatmen t of5 0k gN ha" 1.Stanto n etal., (1966) (accordingt oLinneman n andAzam-Ali , 1993)reporte da yiel d increasefro m fertilizing with 8.5 kgN ha" 1(4 0k gha" ' sulphate of ammonia)thre e weeks after sowing. An increase inyiel d was alsoreporte d inMalaw iwhe n 38k gN ha" 1(18 0k gha" 1 sulphateo fammonia )wa sbande d onth e ridge one month after planting (Anon., 1962).Lat e application of nitrogen has been found to improve plant growth and yield, and reduced senescence in other leguminous crops such as cowpea (Tayo, 1981,1986 :Elowa d and Hall, 1987) and soybean (Isfa, 1991). A combination of nitrogen and phosphorus applied at sowing in Malawi, Nigeria and Zambia generally gavea margina l increase orcoul d evendepres syield s (Anon. , 1970, 1973;Musonda , 1988;Tanim u and Yayock, 1989, 1990). Generalintroduction Dinitrogen fixation Bambaragroundnu t hasbee nreporte d tonodulat ewit hbacteri ao fth e cowpeainoculatio n group (Somasegaran et ah, 1990), while Gueye and Bordeleau (1988) in Senegal found that both indigenous and introduced Rhizobiumstrain s nodulated all 24 bambara groundnut genotypes tested. Experiments haveals obee nconducte d toevaluat eth eeffec t ofRhizobium inoculation on thequantitie so fnitroge nfixed b ybambar agroundnut .I nbot hgreenhous ean dfield experiments , inoculation with Rhizobiumincrease d nodule number, nodule weight,N content ofth e leaves, stems and roots, and N accumulations surpassed those gained by the mineral N fertilizer treatment (Thompson andDennis , 1977;Brook set ah, 1988;Mafongoya , 1988;Gueye , 1992). Goodnodulatio n throughinoculatio n hasbee nobserve dt oincreas eth enitroge n fixing capacity of bambara groundnut and could provide up to 70%o fth eplan t N (Gueye, 1992). In Ghana, nitrogen fixation by bambara groundnut was also reported to increase soilN by 106k g ha"' in anEutri cNitoso l (Dennis, 1977). Applicationo fnitroge n ata rat eo f 50k gN ha" 1i nTogo ,Wes tAfric a was reportedt o suppress nodulation in bambara groundnut (Dadson and Brooks, 1989). Suppression of nodulation has also been reported in other legumes sucha scowpe awhe nN was applied atrate s ofmor etha n 20 kg ha"1(Graha m and Scott, 1984). Thompson and Dennis (1977) in Ghana also observed a decrease inth e number ofnodule s of bambara groundnut when Caan d Mgwer e added. Potassiumand other nutrients An application of 25k g Kha" 1increase d yields in an experiment in Swaziland (Anon., 1979). Tanimu and Yayock (1990) in Nigeria also reported an improvement in seed yield from an application of 25k g Kha" 1.However , Musonda (1988) didno t get arespons e withrate s upt o 183k g K ha1. For an increase of yields, Messiaen (1975) recommended the use of a fertilizer containing phosphorus andpotassium .Indee d somestudie ssee mt oindicat etha tpotassiu m either combined with phosphorus or nitrogen would increase yield. Tanimu and Yayock (1989) obtained the highest yield from a combination of 25 kg K ha"1 plus 11 kg P ha . Bambara groundnut responded favourably also to the application of 50k gN ha" 1(22 9k gha' 1 ammonium sulphate) plus 67 kg K ha"1 (113 kg ha"1 potassium chloride) (Musonda,1988) . A combination of three nutrients (N,P an dK )i nNigeria , Senegal andZambi aseeme d to have noeffec t and sometimes depressedyield s (Tardieu, 1958; Golian dNg , 1987; Anon., 1989). InCongo ,Ghan aan d South Africa, liming was reported to increase bambara groundnut shoot dry weights and yields (NILCO, 1960;Thompso n and Dennis, 1977;Wasserman n etai, 1984). Chapter 1 1.3 Bambara groundnut in Botswana Botswana liesi nth e centre of Southern Africa, between 18°an d27 ° south of equator,an d 20° and29 °eas to fth eGreenwich .Th eclimat ei ssemi-ari d toarid ,averag erainfal l variesfro m 650- 700m mi nth eextrem enorth ,t o 150- 20 0m mi nth eextrem esouthwest , with 400 -50 0m mi n theeaster nsid eo fth ecountry .Mos trai nfall s inth e summer,an dgenerall y starts in late October andcontinue stil l March/April.Ther e isgrea tvariatio n inth eamoun t ofrainfal l and distribution each year. Summer temperatures are around 23 to 28 °C with arang e of about 10° Cabov e or below this average (Sims, 1981). Between October and April, day length changes from 12.7 hours in Octobert o 13.7hour s inDecembe r andbac k to lesstha n 12hour s inApri l (Harris and Azam-Ali, 1993).Betwee n 60 and 70%o fth e country iscovere d by the Kalahari sands.Mos t of the crop farming activities are in the eastern side of the country where the soils are sandy loams, and loamy sands with small areas of heavier soils in depressions and valley bottoms. Natural fertility of these soils is low, being deficient in P,havin g low levels of mineral N and very little organic matter (Sims, 19981;Beynon , 1991). InBotswana , bambara groundnut isusuall y grown for bothhom e consumption and for salean d this dual purpose is the main reason for growing the crop. The main problem for bambara groundnutcultivatio n inBotswan ai slo wrainfall . Thecro pi s mainlygrow nb ywome nwh oals o carryou tmos to fth ework .Th eare aunde rcultivatio npe rfar m issmall ,usuall y lesstha n 0.5ha . Intercropping with cereal cropsoccur sbu t most ofth efarmer s growi ta s a solecrop .Th eplan t densities vary between 2 500 and 16700 0plant s ha"1.Abou t half ofth e farmers plant the crop in rows, while the other half broadcast and plough to cover the seed with soil. With regard to crop management, most farmers ploughthei r fields and weed once a season, while all farmers practice earthing-up (cover the stems and pods with soil). The use of external inputs such as fertilizers andpesticide s islow .Yield s obtained are low,usuall y lesstha n 500k gha" 1(Brin ket al.,1996) .Informatio n islackin go nfertilize r responseo fbambar agroundnu t inBotswan asoils . Thema p of Botswana isgive n inAppendi x 1 and climatic data inAppendi x 2. 1.4 Objectives ofth e study Reasons to study the effects of Pan dN fertilizer application on growth ofbambar a groundnut were: - that little systematic research has been carried out in this field and the results so far obtained are not in agreement with each other; -tha ta n important objective ofth eE Uresearc hprogramm e was"t o recommend suitable management practices to stabilize theyield s ofbambar a groundnut in rainfed environments", which includes fertilizer use; -tha t phosphorus isth e most limiting nutrient inBotswan a soils;an d - that furthermore soils are low inorgani c matter, thus also lowi nmineralizabl eN . Generalintroduction In view of these facts, a study was undertaken to increase the understanding of the P and N nutrition ofbambar agroundnu t andi fpossibl e to make fertilizer recommendations for growers in Botswana. Thehypothese s ofth e study were that: (1) phosphorus islimitin g growthan ddevelopmen t ofbambar a groundnut inBotswan a soils; (2) soil moisture content may limit the Puptak e andth eN 2 fixation process orboth ; (3) bambara groundnut has ahig h Pefficienc y (uptake and internal use); (4) limiting P supply affects N supply tobambar a groundnut through suboptimal N2 fixation; (5) therema y benee d for additional Ni ncertai n development stageswhen : (i)th eN 2 fixation process isno t yet active (early growth); (ii)whe nthi s process slows down (podfilling stage) ;an d (iii) when Psuppl y isn o longer agrowt h limiting factor. The main factors underlying these hypotheses and their interrelationships studied are schematically shown in Figure 1.1. Soilmoistur e ' 1 Soil and Soilan d N2 fixation M i— fertilizerN fertilizerP ^*. ' * Figure 1.1. Diagram showing the possible direct and indirect effects of phosphorus (P) and nitrogen (N) fertilization, N2 fixation and soil moisture on crop production. Chapter 1 1.5 Outline ofth ethesi s Followingth eIntroduction ,Chapte r2 describe sth erespons eo fbambar agroundnu tt oP fertilize r in a pot and a field experiment under different soil moisture regimes. In both experiments the samesoi lwit ha lo wP conten t wasused . Thefindin g that theresponse s of bambara groundnut toP fertilizatio n inth epo tan dfiel d experiment differed ledt othre e hypotheses; (i)tha t lack of response inth efiel d mayb ea nindicatio n oflo wP requirement , (ii)tha tbambar a groundnut has a high P efficiency (uptake and internal use), and (iii) that the positive response in the pot experiment inwhic h Pwa sthoroughl y mixedwit hth e soil isa nindicatio n that extra Pmigh t be only effective if applied and available atth e very early stage ofplan t development. In Chapter 3.1, the Prequiremen t and Pefficienc y ofbambar a groundnut are compared to those of maize andpigeo npea .Th etimin g ofP applicatio n onbambar a groundnut isdiscusse d in Chapter 3.2. The findings that in the previous experiments addition of fertilizer P did not affect the P concentrations inth eshoot swa sa reaso nt oinvestigat eth ecritica l Pconcentration s of bambara groundnut. These were established in a sand culture experiment reported in Chapter 4.1. The critical Pconcentration s were subsequently usedt oasses sth e Pnutritio n ofbambar a groundnut plants in farmers' fields (Chapter 4.2). The critical P concentrations indicated that in the preceding experiments bambara groundnut plants,als othos e receiving additional P,ha d been growing under P limiting conditions. Thisprompte d the research reported in Chapters 5an d6 to find out if nitrogen would become the next limiting factor once the P supply and uptake is adequate. Finally in Chapter 7 the main findings of the study are discussed in relation to the hypotheses, the results from research on other tropical legumes and the bambara groundnut growing practice in Botswana. Responseto phosphorus fertilization Chapter2 Response of bambara groundnut (Vigna subterranea (L.) Verde.) to phosphorus fertilization 2.1 Introduction Asreporte d inChapte r 1 bambara groundnut generally shows apoo r responset o P fertilization even when grown in P-poor soils.Furthermor e no information is available onth erespons e of bambara groundnut to P fertilization in Botswana, where low soil P availability is one of the limiting factors incro pproductio n (Beynon, 1991).Therefore , aserie so fstudie swa sundertake n to investigate the growth- and development response of bambara groundnut to applied P in a representative low available-P soil.A sa firs t orientation, ashort-duratio n pot experiment was conducted to investigate the effect of Pfertilizatio n onth e early growth ofbambar a groundnut (from sowing to flowering). Subsequently the response to Pwa s studied in a field experiment with and without irrigation. 2.2 Apreliminar y pot experiment This was a short term preliminary pot experiment conducted in a soil collected from a site intended for a P fertilization field experiment to determine whether on low P soils bambara groundnut responds to P fertilization. Furthermore, some attention was paid to possible differences inrespons eamon g cultivarsan dth ecapacit y ofth eN 2fixatio n system under natural soilconditions . Two bambara groundnut selections were grown ina P-poo r soil fertilized with P at sevenrates ,wit h orwithou t inoculation ofth e seedwit h awel lknow nRhizobium strain. 2.2.1 Materials and methods Theexperimen t wasconducte d onth e grounds ofth e forestry nursery ofth e Botswana College ofAgriculture ,Sebel e (24°33' S;25°54 'E ;99 4m )fo r aperio d of 51day s and it ranfro m 13t h December 1994t o 2n dFebruar y 1995.Maximu m andminimu mtemperature s varied between 33.1-34.7 °C and 19.2-19.5 °C, respectively. A loamy sand (Ferric Luvisol; Joshua, 1991) collected from a4 0 year old bush fallow field was used. The soil was sieved toremov e stones andanalyse d to determine the main chemical characteristics whichar e given inTabl e 2.1. The determination of pH was in CaCl2 (0.02 M) using a 1:5 soil/solution ratio; P by Bray no.2 extractant; exchangeablebasi ccation san d CECb y ammonium acetate, and organiccarbo n by the Walkley and Black (1934)method . 9 Chapter 2 The treatments were seven P fertilizer levels (0, 80, 160,240 , 320,400 ,48 0 mg Ppe r pot; on weightbasi so f2 000 000k g soil ha"1 corresponding to 0, 10,20, 30,40, 50,an d 60k g Pha 1), twoinoculu mlevel s(wit han dwithout) ,an dtw obambar agroundnu t selections('Diphir i Cream' and 'Zimbabwe Red'). 'Diphiri Cream' isa loca l selection (cultivar) collected from afarme r in the Southern part of Botswana near Kanye, while 'Zimbabwe Red' is a selection from Zimbabwe. The experiment was arranged in a completely randomized design with treatments replicatedthre etimes . Single superphosphate was crushed and thoroughly mixed withth e soil. No mineral nutrients other than P were added. The bean inoculant bacteria -Rhizobium leguminosarumbiovarphaseolin- (5x 10scell s g"1) was applied atth e rate of 500 gpe r 50k g seed, with sugar solution used as sticker. Whiteplasti c bags with ato p diameter of 25 cm and adept h of2 4 cm holding 16 kg of soilwer e used.Thre e seeds were sownpe rba g andthinne d toon eseedlin g 17day safte r sowing(DAS) .Th esoi l moisture content wasmaintaine d at about field capacity,an dth eplant swer eexpose dt onatura l light intensity,temperatur e andda y length. The number of leaves of individual plants was taken weekly starting from 28 DAS. The experiment was terminated at flowering stage, 51DAS .A t harvest, theplant s were separated into shoots(leaves ,petiole s and stems),root s andnodules .Fres h weight of noduleswa staken , andothe rplan tpart s were oven-dried at 70° Cfo r 24hour s and weighed. The shoots were then separated into leaf blades, petioles and stems. The dried leaf blades were ground in a plant grinder with a two mm sieve and subsequently analysed for N and P. Samples of 1.25g dry material were digested ina 2 0 ml sulphuric acid-selenium mixture at 330° Cfo r 2hour s using aTecto r202 0Digester .Afte r cooling,fou r mlo fhydroge nperoxid ewa sadde dt oth ediges tan d the digestion was continued for two hours at 330°C .Th e digest was made up to 200 ml with distilled water, and the above mentioned elements were then determined. Nitrogen was determined from asampl e of2 5m lusin g aBUCH I 323distillatio n unitan dtitrate d with0.005 5 MH2S04. Phosphoruswa sdetermine d with aU V Spectrophotometer. Thedat awer e subjected toANOV A using SAS(1985 )statistica lprogramm e anddifference s betweenmean swer eteste d for significance withth e least significant difference (LSD005). Table 2.1. Themai n chemical characteristics ofth e soil used inth epreliminar y pot experiment. pH- P-Bray Organic Exchangeable cations CEC 1 1 1 CaCl2 (mgPkg- ) carbon (cmol(+ )kg ) (cmol (+) kg ) (wt.%) Ca Mg K Na 4.8 8.4 0.4 1.30 0.93 0.60 0.04 4.69 10 Responseto phosphorus fertilization 2.2.2 Results Numberof leaves The number of leavespe rplan t significantly increased with anincreas e inphosphoru s rate (Fig. 2.1a). The highest number of leavespe rplan t wasobserve d atth ehighes t Prate .Th e two selections differed significantly at 35, 42 and 49DA Swit h higher number of leaves for 'Diphiri Cream' than 'Zimbabwe Red'. There wasn o significant effect of adding bean inoculant. Drymatter production Phosphorus fertilization significantly increased shootdr y matter (DM) ofth etw obambar a groundnut selections at 51DA S withn o significant difference between them (Fig. 2.1b). The shoot DMo f both selections increased linearly (r2= 0.90 ) with increasing Prate .A t aP rat e of48 0m g Ppot" 1shoo t DMwa s increased with about 50%relativ e toth e unfertilized treatment. Phosphorus fertilization hadn o significant effect onroo t DMan d nodule fresh weight. There was also no significant effect of addingbea n inoculant on shoot androo t DM, neither onnodul e fresh weight. (a) (b) Q. E 0 80 160 240 320 400 480 0 80 160 240 320 400 480 1 Phosphorusrat e(m gP pot ) Phosphorus rate(m gP pot 1) Figure 2.1. Effect ofphosphoru s fertilization onnumbe r of leaves (a) for 'Diphiri Cream' (•) and 'Zimbabwe Red' (•) at4 9day s after sowing and on shoot dry weight (b) at 51day s after sowing. 11 Chapter2 Mineralcomposition ofleaf blades Application of Pfertilize r or inoculum had no significant effect onN an d Pconcentration s of leafblade s at 51DA S (Table 2.2).Als oth e selections wereno t significantly different inN and Pconcentrations . However, total shoot Pan dN content s were significantly increased with increase inP rat e (Fig.2.2a-b) .Th e shoot Pan dN content s ofth etw o selections were not significantly different. Atth ehighes t Prate ,th e shoot Pconten t of 'Diphiri Cream' was doubled, relativet oth eunfertilize d treatment (Fig.2.2a) .Th eincreas e inshoo tN content at thehighes t Prat ewa s 30%relativ et oth eunfertilize d treatment (Fig.2.2b) . Table 2.2. Effect ofphosphoru s fertilization onminera l nutrient concentrations of 'Diphiri Cream' and 'Zimbabwe Red' leafblade s at 51day s after sowing averaged overplu s and minus inoculation. Prate Nitrogen (%) Phosphoruss (%) 1 (mg por ) 'Dip. Cream ' 'Zim. Red' 'Dip. Cream' ' Zim. Red' 0 2.87 3.00 0.14 0.21 80 3.11 2.57 0.17 0.14 160 2.70 3.11 0.19 0.17 240 2.68 2.72 0.16 0.16 320 2.87 2.85 0.17 0.17 400 2.91 2.80 0.21 0.19 480 2.66 2.80 0.19 0.20 Mean 2.83 2.85 0.18 0.18 S.E. (56 D.F.) 0.102 0.017 C.V. (%) 12.5 32.8 12 Responseto phosphorus fertilization (a) (b) 550 T-p L,iJL/U.0LSDo.o5s H a. 500 D_ 450 400 350 300 + 250 -+- •+- -+—i—i—i 0 80 160 240 320 400 480 0 80 160 240 320 400 480 Phosphorus rate (mgP pot 1) Phosphorus rate (mg Ppot 1) Figure 2.2.Effec t ofphosphoru s fertilization on shoot Pconten t (a)an d shootN content (b)o f 'Diphiri Cream' (•) and 'Zimbabwe Red' (•) at 51day safte r sowing. 2.3 Afield experiment . The increase in shoot DM of bambara groundnut after P fertilization as observed in the preliminary pot experiment was linearwit h increasing Prate s upt oth e highest rate of48 0m g P pot"', corresponding to 60 kg P ha1 .T o avoid limited P supply in the field experiment the maximum Prat e was increased to 80k gP ha"' .A sth e selections showedth e samerespons e to P fertilization, only the selection 'Diphiri Cream' was used. As the results of the preliminary study showed that the indigenous Rhizobiawer e sufficiently effective in nodulating bambara groundnut noadditiona l Rhizobium wasused . With the lowrainfal l and itspoo r distribution in Botswana, low soil moisture availability might be a growth limiting factor, directly due to an absolute water shortage, or indirectly by lowering P availability in the soil. Therefore, in this field experiment different P rates were combined with two soil moisture regimes, rainfed and irrigated. 2.3.1 Materials and methods The experiment was conducted from 18t h December 1995 to 22 nd April 1996 at Botswana College ofAgricultur e usinga soi l similart otha tuse d inth epreliminar y pot experiment (Table 2.1) with a P level of 6.2 mg P kg"1 ( P-Bray) in the top 0-20 cm. Maximum and minimum temperatures varied between 26.2-30.2 °C and 11.5-18.6 °C, respectively. Day length varied 13 Chapter 2 from about 13.7hour s in December to 11.3 hours in April. The treatments, two soil moisture levels (rainfed and irrigated) and five P fertilizer levels (0, 10,20 , 40 and 80 kg P ha1) were arranged ina split-plo t designan dreplicate d four times.Th emoistur e levelswer eth emai nplot s andth eP rate s subplots.Ther ewa sa 10 mwid espac ebetwee nth emai nplot s ineac hbloc k and betweenth eblock st oavoi d mist-spray between thetreatment s during overhead irrigation. The subplots were 6 m x 6 m with a path of one metre in between. Single superphosphate was broadcast andworke d intoth e soil with adiggin g fork. No mineral nutrients other than Pwer e applied. 'DiphiriCream 'seed ,dresse dwit hcapta na ta rat eo f 12 g/10k gseed ,wa s sowni nrow s at adistanc e often cm and adept h offive centimetres . Thedistanc e between the rows was 40 cm.Plant swer ethinne d at2 3DA St oa distanc eo f5 0c mgivin ga plan tdensit y ofabou t 5400 0 plantsha' 1.Earthin g up(coverin gth estem san dpod swit hsoil )wa sdon ewit ha ho e atth e onset ofpoddin g (64DAS) . Sprinklers with a radius of 20 m were used for irrigation. The quantity of water applied was estimated byplacin g severalcan si nth eplo ta tth ebeginnin g ofth e irrigation, andmeasurin g the collected waterwit ha measurin g cylinder atth eend .Tensiometer s wereuse d for monitoring the soil moisture potential. Irrigation was applied whenth e soil moisture potential was pF 2.6,a s recommended for loamy sands (Stegman etal., 1983).Th etota l amount of water received by eachplo tthroughou tth egrowin gperio dwa s 414m mfo r irrigatedan d36 0m mfo rrainfe d plots. The distribution ofrainfal l andth e time and quantity of irrigation isgive n inAppendi x3 . At each harvest, asampl e of sixplant s wastake n from one spot in eachplo t during vegetative growth (28 DAS), 50%flowerin g (49 DAS,whe n 50%o fth e plants in aplo t had at least one open flower), 50%poddin g (78 DAS,whe n half of the 12plant s marked for monitoring pod development had at least one pod 0.5 cm long), and mid-podding (99 DAS,thre e weeks after 50% podding). At sampling, the number of leaves per plant was counted and the plants were separated intoshoots ,roots ,nodule s andpod s when applicable. Fresh nodule weight wastake n and all other plant parts were oven-dried at 70°C for at least 24 hours and weighed. Thefinal harvest of 12 plantspe rplo twa stake n from thecentr eo fth eplo twhe nth eleave sturne d yellow (126 DAS).Fro m this sample,th e weight of the aerial parts,po d number, pod yield, and total seed yield were recorded. Ground leaf blades from all the sampling periods were analysed for N and Pconcentration s asdescribe d for the preliminary pot experiment. Statistical analysis of datawa s done asdescribe d before. 2.3.2 Results Effectof phosphorus fertilization Therewa sn osignifican t effect ofphosphoru s fertilization nora phosphorus-moistur e interaction ongrowt ho fbambar a groundnut except for ashor t lived increase inroo t DM and nodule fresh 14 Responseto phosphorus fertilization Table2.3 .Effec t ofphosphoru s fertilization andsoi lmoistur elevel so nphosphoru san dnitroge n concentrations (%) of 'Diphiri Cream' leafblade sa tdifferen t growthstage si na fiel d experiment at Sebele, Botswana. Prate 28 DAS1 49DA S 78DA S 99DA S 1 (kg ha ) Rain Irrig.2 Rain Irrig. Rain Irrig. Rain Irrig. Phosphorus concentration 0 0.18 0.24 0.22 0.23 0.13 0.12 0.12 0.11 10 0.18 0.23 0.21 0.20 0.17 0.15 0.13 0.12 20 0.21 0.20 0.20 0.20 0.14 0.14 0.11 0.12 40 0.20 0.22 0.20 0.22 0.16 0.13 0.11 0.13 80 0.21 0.22 0.21 0.19 0.15 0.16 0.11 0.11 Mean 0.19b 0.22a 0.21a 0.21a 0.15a 0.14a 0.12a 0.12a S.E. (24 0.019 0.014 0.072 0.010 D.F.) C.V.(%) 17.9 13.1 14.6 16.9 Nitrogen concentration 0 3.30 3.58 3.79 3.76 2.55 2.80 2.45 2.36 10 3.35 3.53 3.76 3.71 2.69 2.76 2.63 2.43 20 3.52 3.42 3.83 3.71 2.64 2.68 2.32 2.51 40 3.51 3.61 3.84 3.55 2.63 2.67 2.16 2.68 80 3.52 3.25 3.89 3.79 2.97 2.64 2.22 2.31 Mean 3.44a 3.48a 3.82a 3.70a 2.70a 2.71a 2.35a 2.46a S.E. (24 0.14 0.10 0.19 0.14 D.F.) C.V.(%) 8.3 5.1 13.7 11.3 1DA S =day s after sowing. 2Rai n and Irrig.mean s rainfed and irrigated treatments, respectively. Within agrowt h stage,differen t letters indicate significant differences. 15 Chapter 2 weight atth e 50%flowerin g stage.Th eyiel d (number ofpod spe rplant , pod dry weight and seed weight) and leafblad e Pan dN concentration s (Table 2.3) were also not significantly affected by Pfertilization . The leafblad eP an dN concentration s inth efield a t4 9 DAS were onaverag e 0.20%an d 3.75%,respectivel y and similart othos eo fth epreliminar y pot experiment at 51DA S (Table 2.2).Phosphoru s fertilization had no effect on shoot P content (Table 2.4). Effectsof irrigation Irrigation significantly increased all plant parameters measured except root DM.Th e high shoot DM for irrigated- compared to rainfed plantsresulte d ina significantly higher number ofpod s and total seed weight perplan t after irrigation (Table 2.5).O na hectar e basisth e mean final seed yields for rainfed and irrigated plots were 2.81 and 4.21 ha"1, respectively. Irrigation had no significant effect on leaf bladeP an dN concentrations (Table 2.3),bu ta positive effect onth e Puptak e expressed byth e shoot Pcontent s (Table 2.4). Irrigation significantly increased shoot Pconten t at2 8an d 99DAS .Th e shoot Pcontent s increased with plant age up to 78DA S and decreased at 99DA S (mid-podding).Thi s reflects are allocation of shoot Pt o the pods.Base d onliteratur e (Linnemann, 1986; Linnemann and Azam-Ali, 1993)th e Pquantitie s inth e pods ofrainfe d and irrigated treatments are ofth e order of 130m gP an d 146m g Pplant" 1,respectively . 2.4 Discussion In apo t experiment with asmal l soil volume alinea rrespons e to added Pwa s found when this was thoroughly mixed withth e soiljus t before sowing,whil e ina field experimen t on the same soil no response to Pwa s found. Anobviou s explanation for this phenomenon istha t in thepot sth e root densities weretha t hightha tth e lowquantit y of available Pbecam e depleted andtha tthi sresulte d in arespons e to anexterna l Psource .I nth e field, plants can exploit a larger volume of soil andthu s have accesst o more soil P.Thi s effect of soil volume per plant onP availabilit y isreflecte d ina lowe rtota l above ground dry matter production ofplant s grown inpot s with unfertilized soil (13.0 gshoo t DMplant" 1a t 51DAS ) thantha to f comparable plants grown inth efield (16. 0g shoot DMplant" 1 at4 9DAS ) atth e same growth stage. Another explanation for the difference inrespons e toP betwee n the pot and field experiment might beth e mode of Papplication . Inth e pot experiment Pwa s mixed with soil before sowing, giving plants accesst o added Palread y ata nearl y age.I nth efield P wa s applied to the soil surface and superficially worked into the soil,givin g asituatio n inwhic h added P might stay out ofth ereac h ofplan t roots during the early orth e entire growing period. This idea leads to the hypothesis (studied inChapte r 3.2) thatth e response ofbambar a groundnut to added Pdepend s onth e growing stage inwhic h itbecome s available to the plant. 16 Responseto phosphorus fertilization Table 2.4. Effect ofphosphoru s fertilization and soil moisture levelso n shoot phosphorus content (mgplant" 1)o f 'Diphiri Cream' plants atdifferen t growth stages (DAS =day s after sowing). Prate 28 DAS' 49DA S 78DA S 99DA S (kgha"' ) Rain rrig.2 Rain Irrig. Rain Irrig. Rain Irrig. 0 2.3 4.7 35.4 37.8 89.6 87.4 76.1 88.1 10 2.1 3.8 38.7 43.3 109.9 116.9 85.3 98.0 20 3.4 3.2 35.8 34.0 111.0 110.3 78.1 91.7 40 2.8 3.2 42.5 34.5 108.8 104.3 76.5 92.8 80 3.9 4.1 41.2 44.0 116.9 137.8 82.2 91.3 Mean 2.9b 3.8a 38.7a 38.7a 107.0a 111.3a 79.6b 92.4a S.E. (24 0.23 2.35 5.17 3.69 D.F.) C.V.(%) 31.2 27.1 13.4 19.2 1DA S =day s after sowing. 2Rai n and Irrig. means rainfed and irrigated treatments, respectively. Within a growth stage,differen t letters indicate significant differences. Table 2.5.Effec t of soilmoistur e levels on shoot dry weight, number ofpod san d total seed weight of 'Diphiri Cream' atdifferen t growth stages ina fiel d experiment at Sebele, Botswana. Period Shoot dry weight Number ofpod spe r Total seed weight (DAS) (g plant1) plant (gplant 1) Rainfed Irrigated Rainfed Irrigated Rainfed Irrigated 78 72.6b 80.7a 98.2b 133.0a - 99 68.9b 80.0a 152.4b 178.0a - 126 69.6b 80.2a 127.2b 173.1a 51.4b 77.2a Within aparameter , different letters indicate significant differences. 17 Chapter 2 Phosphorus fertilization had no effect on leaf blade P concentrations in both the preliminary pot experiment and the field experiment, but there was an increase in P uptake with increase in shoot DM. Shoot P concentrations in the field experiment for rainfed and irrigated treatments were maintained around the critical P level for bambara groundnut, being around 0.15% at 78 DAS as determined in Chapter 4. This points to a maximum dilution of the P taken up, and a maximum P use efficiency. The fact that the level of the leaf blade P concentration was maintained at the same level at all P rates in the pot experiment but also under irrigated field conditions with relatively high yields, indicates that P availability was always marginal or even deficient. Under all conditions, P availability was limiting plant growth. In the pot experiment this is shown by the linear response of the shoot DM to P fertilization (Fig. 2.1b). In the field, applied P was not utilized and bambara groundnut seemed to cover its requirement with the P already present in the soil, also when growth and P uptake were relatively high with irrigation. At a soil level of 6.2 mg P kg"1 soil (P-Bray) seed yields of even 4.2 tons ha"1 were achieved, but still at leaf P concentrations at around the critical level and without any response to P fertilization. Obviously, under rainfed conditions only a small fraction of the P present in the soil is available. This fraction seems not to be increased by P fertilization. Consequently, plants keep suffering from P deficiency also after fertilization. Contrary to P fertilization, irrigation greatly improved growth and increased P uptake, especially in the early stage of development (28 days after sowing) and during podding (99 days after sowing). Evidently, part of the P pool present in the soil becomes available under improved soil moisture conditions, probably due to improved P diffusion to the root. Apparently, soil moisture seems to be the factor regulating P availability rather than the soil P content, at least under the prevailing non-irrigated field conditions. Eliminating the most limiting factor, that is suboptimal soil moisture, improves plant P nutrition, but, based on the shoot P concentrations, being still critical. Although under the prevailing field conditions, soil moisture is assumed to be the factor regulating P availability and thus growth, water deficiency is not the first growth limiting factor. Under rainfed conditions, soil moisture indirectly affects growth by limiting P supply, leading to shoot P concentrations around or below the critical level. With water shortage as first growth limiting factor, plant growth will be inhibited by water stress, generally resulting in an accumulation of nutrients, inclusive P. This was not observed in the field trial. Nitrogen nutrition in both the pot and the field experiment seemed to be adequate, irrespective of the rate of P supply, addition of bean Rhizobium or soil moisture content. Nitrogen concentrations in the leaf blades were high in all treatments, indicating that N nutrition was sufficient (Tables 2.2 and 2.3). Rhizobium activity in the soil used seems to be high enough to meet the plant requirement. Adding extra inoculum to the soil seems to be not necessary and therefore this treatment was omitted in subsequent experiments. Phosphorus fertilization and soil moisture level did not affect plant N nutrition. Whether N2 fixation capacity will also be high enough under conditions where P is no longer limiting growth, and plant N requirement will be higher, is subject of a further study (Chapters 5 and 6). 18 Responseto phosphorus fertilization An important conclusion from the field research istha t the soil moisture content seemst ob e the important factor determining soil Pavailabilit y andP uptak e inthes erelativel y P-poor soils. Therefore Pfertilizatio n alone doesno t stimulate Pavailabilit y under dry soil conditions, but irrigation does,resultin g inincrease d Puptak e andhighe r DM.Thi s increase inP availability after irrigation, thoughrelativel y high andresultin g in significantly increased dry matterproductio n and seedyiel d isno ttha thig htha t itsupplie sth eplan twit h sufficient P. The shoot Pconcentratio n still remainsmarginal .Anothe r important finding istha t surface application of Pi sno t effective even under improved soil moisture conditions. Finally, asth e two selections used did not differ indr ymatte r production and showed the samedr y matter responset o added P,onl y one selection 'Diphiri Cream' wasuse d inth e subsequent experiments. 19 Factorsaffecting response to P fertilization Chapter3 Factors affecting the response of bambara groundnut (Vignasubterranea (L. )Verde. )t o phosphorus fertilization 3.1A compariso n ofth eP uptak e patterns ofbambar a groundnut, maize and pigeon pea 3.1.1 Introduction In the previous chapter, a positive response to P fertilization by bambara groundnut {Vigna subterranea) was found in apo t experiment butno t ina fiel d experiment. Lacko frespons e by bambara groundnutt oP fertilizatio n inou rfiel d experiment, evenwit ha see dyiel d of4.2 1 ha"1, implied that there was enough available soil P after irrigation to meet the requirement and probably alsot omee t therequiremen t ofeve na hig hP demandin g crop like acereal . Thepoo r responseo fbambar a groundnut toP fertilizatio n inth e field could also meantha t itha s alo wP requirement,possibl y combined witha hig h Pefficienc y (efficiency involves Pacquisitio n by the roots (uptake) and internal utilization by the plant). Efficiency in acquisition is frequently defined in terms of the total uptake per plant or specific uptake rate per unit root length and efficiency in internal utilization (internal use efficiency) as dry matter production per unit nutrient inth edr ymatte r(Marschner , 1997). Thesequestion shav ele dt oa po texperimen twher e theP requiremen t andP efficienc y (uptake and internal useefficiencie s of soilan dfertilize r P) of bambara groundnut were compared to those of maize (Zeamays) an d pigeon pea(Cajanus cajari).Maiz e isconsidere d to have ahig hP requiremen t (Arnon, 1975),whil epigeo n pea has ahig huptak e efficiency bymobilizin g soilP fraction s normallyno tavailabl et oothe r crops (Ae etal, 1990). 3.1.2 Materials and methods The experiment was conducted onth eground so fth e forestry nursery ofBotswan aColleg e of Agriculture, Sebele using a low P loamy sand (Ferric Luvisol, Joshua, 1991). The chemical characteristics ofth e soil are given inTabl e 3.1. Analytical methods are given inChapte r2 . The experiment ran for 78 days from the 4t h October to the 21 stDecembe r 1996.Th e plants wereharveste dwhe nthe y switched from vegetative growtht opo dfillin g stage.Maximu m and minimumtemperature svarie dbetwee n30.0-32. 7 °Can d 16.6-17.8°C ,respectively . Itha sbee n found thatunde rth eprevailin g daylengt han dtemperatur e ittake s for bambara groundnut about 45 days from sowing to flowering and another 31day s from flowering to podding,whil e after 21 Chapter3 Table 3.1. Selected chemical properties ofth e soilused . pH- P-Bray Organic Exchangeable cations CEC 1 1 CaCl2 (mgkg- ) carbon (cmol(+ )kg" ) (cmol (+)kg"') (wt%) Ca Mg K Na 4.6 6.2 0.4 1.3 0.93 0.60 0.04 4.69 podding vegetative growth stops (Brink, 1997; Chapter 2). The treatments were three crops, maize -Zea mays (L) .var . Kalahari Early Pearl),bambar a groundnut (Vignasubterranea (L. ) selection 'Diphiri Cream'- Sebeleprogeny) ,pigeo n pea(Cajanus cajan(L. )var . ICPL87105 - Sebele progeny) and five levels ofphosphoru s (0,70 , 140,28 0 and 560m g Ppot" 1;o n weight basiso f2 00000 0k gsoi lha" 'correspondin g to 0, 10,20,40 and 80k gP ha 1). The experiment was arranged ina completely randomized design withtreatment s replicated four times. White plastic bags (diameter 25 cm x 24 cm deep) holding 14k g of soil were used for growing the crops. Singlesuperphosphat e wascrushe d andthoroughl y mixedwit hth e soilbefor e planting. Three seeds were sown per bag and thinned to one seedling 12 days after sowing (DAS). To bambaragroundnu t andpigeo npe an omineral sothe rtha nP wer eapplied . Toth enon-N 2 fixing maizeplants ,i n additiont oP ,ure aa ta rat eo f35 0m gN pot" 1pe rapplicatio n (based ona weigh t basis of soilpe r hectare corresponding to 50k gN ha1), was dissolved in25 0 ml ofwate r and applied at 28, 42, and 56 DAS. The plants were sprayed with Kombat (active ingredient: Cypermethrin / pyrethroid 200 g L"1) at a rate of 5 ml per 20 litres of water to control grasshoppers. The soil moisture content was maintained at field capacity and the plants were exposed to natural daylight, temperature andda y length. Data recorded at final harvest (78 DAS) were dry weight of plant components (shoot, roots, nodulesan dpods )an dnumbe r ofnodules ,wher eapplicable .A tharvest ,maiz eha djus t started silking.Th eplan tpart s(shoo tan d roots)wer eoven-drie d at70° Cfo r2 4hours ,whil epod swer e air dried for two weeks. The ground dry shoot material was analysed for P as described in Chapter 2. Statistical analysis ofdat awa sa sdescribe d inChapte r 2. 3.1.3 Results Drymatter production Theeffect s ofP fertilization on shoot dry matter production are given inTabl e 3.2. Relative to the zero P treatment, maize had a significant dry matter increase of about 50% at all P rates, whilefo r bambaragroundnu t onlyP rate so f28 0m gP pot" 1 ormor e gave a significant response of30-60% .I npigeo npe aP fertilizatio n tendedt ogiv ea smal l increasei ndr ymatte r production, 22 Factorsaffecting response to P fertilization but this effect was not significant. Table 3.2 further shows the large differences in total dry matter production between the species.Th edifference s between the Presponse s ofmaiz e and bambaragroundnu t arefurthe r illustratedi nFigur e3.1 awher eth erelativ edr y matteryield s (as apercentag e ofth e maximum yield) are given. Phosphorus fertilization significantly increased root drymatte r weight of maizeonly .Th e root- shootratio sclearl y showtha tunde rth e experimental conditions bambara groundnut developed arelativel y smallroo t system (Fig.3.lb). Bambara groundnut has inal ltreatment s a root-shoot ratio ofabou t 0.2,whil e for maize andpigeo npe avalue s between 0.4 and 0.5 were found. Shootphosphorus status Theeffect s ofP fertilization onP nutritio no fth eexperimenta lplant sar egive ni nTabl e 3.3 and Fig. 3.2. Shoot Pconcentration s are hardly affected by P application except when the highest rates are used (Fig. 3.2). The maize shoot P concentrations are very low while the bambara groundnut shootP value sar etw ot otw o andhal ftime shighe r but still low. Pigeon pea had the highest shoot Pconcentrations . Table3.2 .Effec t ofphosphoru s fertilization on shootdr ymatte r ofmaize ,bambar a groundnut and pigeon peaa t 78day s after sowing. Phosphorusrat e(m g Shootdr y matter (gplant" 1) P pot1) Maize Bambara groundnut Pigeon pea 0 70.8 22.1 7.0 70 104.6 26.5 9.1 140 106.2 27.3 8.8 280 108.7 35.2 9.8 560 104.9 29.0 6.4 Mean 99.0 28.1 8.3 C.V. (%) 10.3 15.3 23.4 S.E. (15D.F.) 5.08 2.13 0.97 LSD005(Phosphorus ) 15.3 6.4 n.s. n.s. =no t significant 23 Chapter 3 (a) (b) 0.6 • ^ 0.5 • O) 0.4 - >-^ o 0.3 • 1— K—*~«^_ * no 0.2 ' CO 0.1 • o a: o • —I—I—I—I—I—I—I—h- 7014 021 028 035 042 049 056 0 70 14021 028 035 042 049 056 0 Phosphorus rate(m gP pot 1) Phosphorusrat e(m gP pot 1) Figure 3.1.Effec t of phosphorus fertilization on relative shoot dry weight (a) and root-shoot ratios (b), of maize (•), bambara groundnut (•) and pigeon pea (A) at 78 days after sowing (highest shoot DMyiel d = 100). Inmaiz ether ewa sa stead yincreas eo fshoo tP conten twit hincreasin g Prate s (Table 3.3). This effect isalread ypresen t atth elowes tP rat eo f7 0m g Ppot" 1an d isassociate d with adr y matter increase of about 50%(Tabl e 3.2). In bambara groundnut thispositiv e effect of P fertilization onshoo tP conten tan d shootdr ymatte rshow su pa ta fou r timeshighe rP rate .Fro m Tables 3.2 and 3.3 it can also be seen that maize growth heavily depends on additional fertilizer P, that bambaragroundnu t alsoneed s fertilizer Pfo rmaximu m shootdevelopment , buttha tpigeo n pea hardly relieso n additional fertilizer P. 3.1.4 Discussion Underth eprevailin g growthcondition srespons et oP fertilizatio n differed verymuc hamon gth e three species,reflecte d inbot h Putilizatio n and growth response. While maize and pigeon pea alreadyreache d 90%o fthei rmaximu m shoot drymatte rproductio n atth e lowest Prat e (70m g Ppot" 1),bambar a groundnut required afou r timeshighe r Prat et o realize a similar relative dry matter production. 24 Factorsaffecting response to P fertilization Table3.3 .Effec t ofphosphoru sfertilizatio n onshoo tP conten to fmaize ,bambar agroundnu tan d pigeon peaa t 78day s after sowing. Phosphorus rate (mg Shootphosphoru s content (mgP plant 1)* P pot') Maize Bambara groundnut Pigeonpe a 0 28.0 27.0 11.2 70 49.0(21.0) 28.5(1.5) 14.0 (2.8) 140 53.1 (25.1) 31.7(4.7) 14.8(3.6 ) 280 62.3 (34.3) 46.4 (19.4) 17.6(6.4) 560 99.0(71.0) 51.7(24.7) 18.3(7.1) Mean 58.3 37.0 15.2 C.V. (%) 12.3 18.5 17.7 S.E. (15D.F.) 3.59 3.42 1.3 LSD005 - (Phosph orus) 10.8 10.3 4.0 Figures between thebracket s represent the amount ofP originatin g from the P fertilizer. 0.3 •-S 0.25 c o (0 0.2 c o 0.15 c o o 0.1 Q. o 0.05 o JC V) 0 0 70 14021 028 035 042 049 056 0 Phosphorus rate (mg P pot1) Figure 3.2.Effec t ofphosphoru s fertilization on shoot Pconcentratio n ofmaiz e (•), bambara groundnut (•) and pigeon pea (A) at 78day s after sowing. 25 Chapter3 When analysing plant growth inrelatio n to Pnutrition , two major components havet ob e considered, that is(i )plan t Prequiremen t and (ii)plan t Putilization . Plant Prequirement , the minimum amount ofP t o betake n upt o achieve maximum drymatte r production, wasmuc h lower with the two leguminous speciestha n with maize.Wit h bambaragroundnu t and pigeon pea,maximu m shoot DMyield s of 35.2an d 9.8 gplant" 1wer e accompanied by shootP concentrations of 0.13% and 0.18%,respectively . Thisresult s in Prequirement s of46. 4 and 17.6m gP plant" 1,a tleas t ifdetermine d for agrowt hperio d of7 8day san drestricte d to shoots only. Withmaize ,th e experimental datahav et ob ehandle d with caution. While with pigeon peaan d bambara groundnut shoot Pconcentration s of0.1 3an d 0.18% correspond well withthei r minimum Pconcentration s (Reuter andRobinson , 1986;Chapte r 4),interna l P concentrations inmaiz eplant s were far belowth e minimum Pconcentration , indicated by Reuter andRobinso n (1986)t o vary between 0.10-0.15%.I n our experiment, only atth e highest Prat emaiz e approached a shoot Pconcentratio n of 0.10%wit h a shoot dry matter yield of 104.9g plant" 1.Base d onthes edata ,correspondin g Prequiremen t for maizewoul d be 99.0m g Pplant 1, buta level of 148.5m g Pplant" 1(base d on0.15 %P i n shoot) would probably bemor e realistic. ShootP concentration s inmaiz e far belowth e minimum concentration, asmeasure d for all Prate s except thehighes t one,illustrat e that at all lowerP rates maizewa s severely suffering from Pdeficiency . Extremely lowinterna l shoot P concentration (^0.05%P )measure d at 78DA Sca nonl y beth eresul t ofa stron g dilution of internal P,du et o prolonged biomassproductio n (photosynthesis) at stagnating Puptake .I n this experiment maize grown for 78day sa tdifferen t Prate s seemed to haveproduce d equal amounts ofbiomas s (except atzer o P),resultin g inincrease d internal Pdilutio n andthu s increasing Pdeficienc y with decreasing Prate . Though Prequiremen t ofmaiz ewa s at least two and five times ashig h astha t of bambara groundnut andpigeo npea ,respectively ,thes edat ahav et ob ehandle d withcar ei f transferred topractice . Only part ofth e growth cycle wascomplete d (78 DAS),calculation s arebase d only on above-ground parts and, ifextrapolate d from plant to crop level theplan t density in thefield ha st o be incorporated. Onth eothe rhand , ifdifference s inminimu m P concentrations among species canb eassume d tob e small,P requirement s will directly be related totota l dry matterproductio n perha .Roug h estimates from practice givetota l DM (excluding grain) data of7.0,4. 5 and 3.71ha" 1fo r maize (Sigunga, 1997),bambar a groundnut (Collinsonet ah, 1996;Chapte r 2)an dpigeo npe a (Aeet ah, 1990),respectively . This agrees withth ehig h Prequiremen t ofmaiz erelativ e toth e twolegume s asfoun d inthi s experiment. Phosphorus utilization, the second component involved inplan t Pnutrition , isregulate d by both Puptak e efficiency, and internal Pus e efficiency. Uptake efficiency is frequently defined astota l uptake perplan t ora suptak e peruni t root length (Marschner, 1997).I nthi s work total Pi n shoots at 78DA S wasassume d to represent total Ptake n uppe r plant, while finalroot/shoo t dry weight ratioso fth especie swer euse d asa paramete r characterizing the relative root density of aspecies .Maiz e showed tob ever y efficient inP uptake . Contrary to bambara groundnut andpigeo n pea,P uptak eb ymaiz ealmos t linearly increased with increasing Pfertilizatio n rate (Table 3.3).A relativ e high root density, reflected by ahig h root/shoot ratio,combine d with ahig h Prequiremen t and asmal l rooted soil volume (pot 26 Factors affectingresponse toP fertilization experiment) led to this linear response withmaize .Phosphoru s uptake withpigeo n peawa s lowwit h almost norespons e to Pfertilization , though root/shoot ratio was ashig h astha t of maize.Evidentl y Prequiremen t seemst o act asa sin k that indirectly controls Puptak e of pigeonpea .Phosphoru suptak epatter n ofbambar agroundnu twa squit edifferen t from that of maizean dpigeo n pea.U pt o 280m gP pot" 1no t much increase inP uptak e was observed, while from this level onP uptak e reached more or lessa maximu m ata leve ljus t in between maize and pigeon pea. Probably because ofa lo wroot/shoo t ratio (lowroo tdensity ) bambara groundnut seems not tob e ablet o benefit from fertilizer P,a tleas t ifdose s applied are below a certain critical level. Only atP rate s equal or greater than 280m g Ppot 1, soil P availability seems to behig h enought o supply poorly rooted bambara groundnut plantswit h adequate amounts of P. Increase inP uptak e wasaccompanie d byimprove d growth, indicating P uptake at lower Prate s was inadequate. Duet o high Puptak e and high Prequiremen t the major fraction ofP absorbe d by maize originated from fertilizer P(u pt o 71%),makin g fertilization with phosphorus necessary for maize grown under theprevailin g conditions.Jus t contrary, bambara groundnut and pigeon pea seemed to be ablet o cover upt o 80%an d 94%o fthei r Pb y Poriginatin g form soil, probably duet o their lowP requirement . Therefore, fertilization ofpigeo n peawit h P seems to bebeneficia l only at lowrates . Whetherthi s ispartl y related to additional plant specific P mobilizing strategies sucha s(i )roo t exudation oforgani cacids ,o r (ii)mycorrhiz a symbiosis, ispossibl e (Ae etal., 1990),bu tha s not been subject of study inthi sexperiment . Fertilization ofbambar a groundnut with P seemst o improve uptakean d growth only ifrate s exceed a certain level. This levelwil lprobabl y depend on soil type,soi l moisture status (Chapter 2) andplan t density. The internal Pus e efficiency, characterised byth e amount ofbiomas sproduce d per unit P absorbed (Janssen, 1998)differe d very much betweenth ethre e species.Maiz e showed the highest internal use efficiency (Table 3.4). Onaverag e maizeproduce d 1.89 gD Mpe r mgP taken up,whil evalue s for bambara groundnut andpigeo npe awer e only 0.79 and 0.56 gpe r mg,respectively . Because of severeP deficiency , extremely high Pus eefficiencie s of maize atlo w Prate shav et o beconsidere d asartefacts . Under thosecondition s theplant s would probably not have completed their growth cycle,inclusiv e seedproduction . More realistic values of internal Pus eefficiencie s will probably bethos e from treatments used to calculate P requirements. Then the values of 1.06,0.7 6 and 0.56 gD Mpe rm gP take n upwer e obtained for maize,bambar a groundnut andpigeo npea ,respectively . These data illustrate that maize isstil l the most efficient species,whil eth etw o legumesbehav e more or less similar. Finally, the resultso fthi s experiment clearly illustrate thatth e lack ofrespons e offield- grown bambara groundnut to Pfertilizatio n (Chapter 2)canno t beattribute d to exceptionally high Pus e efficiency ofthi scrop .I f compared with maize,a specie swel l knownt o require high Pfertilize r applications (Arnon, 1975),P uptak e and internal useefficienc y of bambara groundnut and pigeon peawer emuc h lower. Thepotentia l ofbambar a groundnut torealiz e 27 Chapter 3 highrelativ e yields on Ppoo r soils without fertilization asi nou r field trial has tob e attributed to itslo w Prequirement . However, asshow n inChapte r 2,unde r Botswanafield conditions relatively high Puptak e andrelativ e high seed yields (4.21 ha"1)coul d be achieved withbambar agroundnu tb y improvingth esoi l moisture levelb y irrigation, without P fertilization. Table 3.4. Effect ofphosphoru s fertilization on internal Pus e efficiency ofmaize ,bambar a groundnut and pigeon pea at7 8DAS . Phosphorus rate Phosphorus utilization efficiency (D M/ unit Ptake n up;g mg"1) (mgP pot 1) Maize Bambara groundnut Pigeon pea 0 2.53 0.82 0.63 70 2.13 0.93 0.65 140 2.00 0.86 0.59 280 1.74 0.76 0.56 560 1.06 0.56 0.35 Mean 1.89 0.79 0.56 Thequestion stha t stillremai n are(i )whethe r suchyield swithou t Pfertilizatio n canb eachieve d only on virgin soils and if so, (ii) what rates of P fertilization will be necessary to maintain adequate yields in permanent cultivation systems or to increase yields. In pot experiments however, soilP availabilit y seemst ob einsufficien t tocove rth erelativel y lowP requiremen t of bambara groundnut plants,du et oth e limited soil volume exploitable per plant. 3.2 Growth and development of bambara groundnut as affected by time of phosphorus application 3.2.1 Introduction Availability ofphosphoru s atearl y stages ofplan t growth hasbee n found tohav e a great effect ongrowt han dyiel d (Jonesan dWarren , 1954;Smi dan dBates , 1971;Arnon , 1975).Phosphoru s 28 Factors affectingresponse toP fertilization sources and methods ofapplicatio n have alsobee n found to differ inthei r effect on early plant growth (Reisenauer, 1963; Krigel, 1967;Smi d and Bates, 1971;Scot t andBlair , 1988). Bambaragroundnu tha sbee nreporte dt othriv ean dproduc esom eyiel dunde radvers econdition s such as limited water supply and low soil fertility (National Academy of Sciences, 1979; Collinson et ah, 1996). It has been found that bambara groundnut has a relatively low P requirementwit hmos to fth edr ymatte rproduce db ymakin gus eo fsoi lP (Chapte r3.1) .Th elo w P requirement partly explains why bambara groundnut did notrespon d toP fertilization under field conditionsbu tresponde di na po texperimen twher esoi lvolum eseeme dt olimi tth eamoun t ofP availabl e perplan t (Chapter2) . A positive response to Pfertilizatio n by bambara groundnut in the pot experiment, where the fertilizer was ground andthoroughl y mixed withth e soil,implie d also a"starte r Peffect" , e.g. the need for additional Pa tth e very early stage ofplant/roo t development (Chapter2) . This led to the hypothesis that additional P might only be needed at the very early stage of plant/root development. The assumption was based onth e idea that inth e pot experiment the rootso fyoun gbambar agroundnu tplant s"meet "th efertilize r Pdirectl y after germination,whil e in the field, whereth e non-ground fertilizer wasbroadcas t and ploughed under with a digging fork,thi shappen sonl ya ta late rstage .Thi smigh tb eparticularl y relevant forbambar agroundnu t becauseo fit spoo rroo t system (Chapter 3.1).Th e Prespons e inth epo texperimen t mightthus bea "starte reffect" . Tochec kthi shypothesis ,a po texperimen twa sconducte d inwhic h different times of Papplicatio n were compared. 3.2.2 Materials and methods The experiment was conducted onth e grounds ofth e forestry nursery of Botswana College of Agriculture with the same soil as used in Chapter 3.1 for a period of 78 days from the 25 th October 1996t o 11 thJanuar y 1997.Maximu man dminimu mtemperature svarie dbetwee n30.0 - 32.7° Can d 16.6-18.7 °C,respectively .Th etreatment sinclude da contro l (noP fertilization ) and three times of P fertilizer application (sowing, two and four weeks after sowing (WAS)). Phosphorus was applied at a rate of 420 mg P pot1, onweigh t basis of 2 000 000 kg soil ha1 corresponding to 60 kg Pha" 1.Whit eplasti cbag sholdin g 14k g of soilwer euse d for growing theplants .Thre e seedso fbambar agroundnu t selection 'Diphiri Cream' (Sebele progeny) were sown and thinned to one plant per pot 14DAS .Th etreatment s were arranged in a completely randomized design with five replicates. For treatment application at sowing, single superphosphate (SS) was crushed and thoroughly mixed with the soil. For the times of application attw oan d four WAS, thecrushe d SSwa sdissolve d in25 0m l ofwate r and applied toth e soil.N omineral sothe rtha n Pwer e applied. The soil moisture content was maintained at field capacity and the plants were exposed to natural light, temperature and day length. Data recorded atth efinal harves t(7 8DAS )wer edr yweigh to fplan tcomponent s (shoot,root ,nodule s and pods) and shoot P concentration. The ground dry shoot material was analysed for P as described in Chapter 2.Statistica l analysiso f datawa sa sdescribe d inChapte r 2. 29 Chapter 3 3.2.3 Results Shootdry matter production. Theincreas ei nshoo tDM ,relativ et ocontrol ,wa s70 %fo rP applicatio n atbot h sowingan dtw o WAS(Fig .3.3) .I fapplie d four WAS,th eincreas ewa sonl y 25%, being no longer significantly different from thezer oP treatment .Contrar yt o shoots,dr yweigh t ofroots ,nodul e number and nodule dry weight were not significantly affected by P fertilization, irrespective of time of application (Table 3.5).Th eroot:shoo t ratios (ondr yweigh t basis) did not differ significantly, butthe ytende dt ob ehighe rfo rtreatment scontro lan dP applicatio n atfou r weeks after sowing, compared to Papplicatio n withintw o weeks from sowing. Podyield Phosphorus fertilizer application significantly increasedpo ddr yweigh tpe rplant ,relativ e toth e zero Ptreatment , if supplied withintw o WAS.I f supplied during this period, pod dry weights wereincrease d bya facto r of sixt o seven (Fig. 3.4).I f applied four WAS,however ,po d yields weren olonge rsignificantl y affected. Withbot h shoot andpo d DMweights ,positiv e effects of P fertilization were clearly restricted to application during the first two weeks after sowing. Effects ofP fertilizatio n onpo d yieldwer e much strongertha no n shoot dry weights. ~ 30 S 2WA S 4WA S Treatments Figure 3.3.Effec t of time of phosphorus fertilization on shoot dry weight of 'Diphiri Cream' plants at 78 days after sowing. (OP= control; S = P at sowing; 2 WAS = P at 2 weeks after sowing ;4 WAS =P a t4 week s after sowing. 30 Factorsaffecting response toP fertilization Table3.5 .Effec t oftim eo fphosphoru sapplicatio n onshoo tan droo tdr yweight san droot:shoo t ratios of 'Diphiri Cream' plants at7 8day s after sowing. Treatments Shootdr y weight Rootdr y weight Root: shootratio s (gplant" 1) (gplant 1) (dry weight) Control 16.1 3.34 0.21 P at sowing 27.5 4.14 0.15 P at2 WAS * 27.3 3.88 0.14 P at4 WA S 20.1 3.70 0.20 C.V. (%) 15.1 19.3 27.8 S.E. (16D.F.) 1.54 0.32 0.02 LkIJo.05 4.61 n.s. n.s. *WAS= weeks after sowing;n.s .= no t significant. 2WA S 4WA S Treatments Figure 3.4.Effec t oftim eo fphosphoru s fertilization onpo ddr yweigh t of 'Diphiri Cream' at7 8 daysafte r sowing.(O P =control ; S= P a t sowing;2 WA S= P a t2 week s after sowing; 4WA S = Pa t4 week s after sowing. 31 Chapter 3 Shootphosphorus status Phosphorusfertilizatio nha dn osignifican t effect onP concentratio n ofth eshoo t(Tabl e3.6) .Th e shootP concentration swer ei nth esam erang ea si nth epreviou s experiment (Chapter 3.1),tha t is0.11-0.13% . ShootP contents ,reflectin g Puptake ,wer eaffecte d byP fertilizatio n in asimila r waya sshoo tdr yweight s(Tabl e3.5) .Plant ssupplie dwit hextr aminera lP durin gth einitia lstag e (sowingt otw oWAS )too ku pmuc hmor eP (50-70% )tha nunfertilize d plantso rplant sreceivin g mineral P at four WAS. Table3.6 .Effec t oftim eo fphosphoru sapplicatio no nshoo tP concentratio n andshoo tP conten t of 'Diphiri Cream' at 78day s after sowing. Treatments Shoot P Shoot Pconten t concentration (%) (mgP plant1) Control 0.12 19.6 P at sowing 0.12 34.7 Pa t2 WAS * 0.13 34.7 P at4 WA S 0.11 23.2 C.V. (%) 17.5 27.2 S.E. (16D.F.) 0.01 3.41 LSI-V05 n.s 10.2 "WAS= week s after sowing;n. s= no t significant 3.2.4 Discussion Theobserve d responseo fshoo tdr ymatte r production ofpotte d bambara groundnut plants toP given at sowing (ground Pfertilize r thoroughly mixed withth e soil at orbefor e planting) is in linewit hth eresult sobtaine d intw opreviou spo texperiments ,wher ei twa sdu et oa combinatio n oflo wP conten to fth esoi lan dlimite d soilvolum e(Chapter s2 an d 3.1).Th edat asho wtha t the response only occurs when P is applied (and well distributed in soil) at a very early stage of seedling development, inthi sexperimen twithi ntw oWAS .Phosphoru s application beyond this 32 Factorsaffecting response toP fertilization stage apparently does not improve shoot growth. This means that the critical time for P application with positive effect onplan t growth iswithi ntw o weeks after sowing. The fact that P application does not affect root development suggest that bambara groundnut seedlings dono textr ainves t inthi sorgans ,irrespectiv e ofP supply. Despiteth e fact that plants receiving Pa tfou r WAShav ea "normal "roo t system,n oextr aP i stake nu pdurin glate rstages . Thisma yb edu et oth eabsenc eo f anappropriat e sink. Itma ymea ntha t shoot growth (sink)o f bambara groundnut isdetermine d atearl y stages of growth,wit hP stress leadingt o arelativel y small sinktha tca nno tb ealtere d by ato o late increase in soilP availability. Phosphorus application hadn oeffec t onpo d numberbu ta ver y strongeffec t onpo dweight . For controlan dP applicatio na tfou rWA Streatment sth eaverag epo ddr yweigh tplant" 1 were0. 5an d 1.0 g,respectively ,whil e 4.0g fo r Papplicatio n atsowing .Thi sindicate stha tunde rP stres sth e aerialpart sdi dno t supply sufficient assimilates forpo dfilling. Reduce d early growth apparently hasgrea t consequences forpo dfilling. Thi sshow sth eimportanc e ofa "starte r P"o n production ofbambar a groundnut andtha t apoo r startca nno t bemad e good by a later P application. Finally,th efindings o fthi s experiment thatther e isa critica l time for Papplicatio n may partly explainth e lack of response to applied Pi nou rfield experimen t (Chapter 2). Inthi sfield trial , thefertilize r wasbroadcas t andworke d intoth e soilwit h adiggin g fork. This isdifferen t from thepo t experiments whereth efertilize r isthoroughl y mixedwit hth e soil,an dth e seedlings can makeus eo fi tdirectl y after germination.Unde rfield conditions ,probabl yth eseedling smee tth e fertilizer onlyafte r sometim ean dmisse dth ecritica ltim efo r Pi nbambar a groundnut. This can mean thatunde rfield condition s ifth e soil P status islow ,broadcastin g the fertilizer isno t the best way to increase P availability at early growth stages of bambara groundnut. Fertilizer placement inth erootin g zonebefor e planting may overcome this problem. 33 Phosphoruslevels inshoots Chapter4 Phosphorus levels in shoots of bambara groundnut (Vignasubterranea (L)Verde. ) 4.1 Critical shoot phosphorus concentration of bambara groundnut. A sand culture approach 4.1.1 Introduction Preceding fertilizer experiments with bambara groundnut gave positive responses to P application in pot experiments and no response in a field experiment (Chapters 2 and 3). Differences in rootable soil volume and in method of P application may well account for the differences inresponse .I nal lexperiment s reported ons ofa r the Pconcentration s of leaf blades and shoots were not affected byP treatments ,eve nwhe n shoot drymatte r was increased. This might mean that in all these experiments bambara groundnut plants maintained a shoot P concentration around the critical level by diluting theextr aP take n upint oth e extrabiomass . To check this idea a sand culture experiment with different P treatments was conducted to establishth ecritica lP level sfo rbambar agroundnut .I naccordanc ewit hEpstei n(1972 )a critica l level wasdefine d asth enutrien t concentration ofth etissu e associated with 5o r 10%reductio n inmaximu m growth duet o deficiency oftha t nutrient. 4.1.2 Materials and methods Theexperimen t wasconducte d ina greenhous e atBotswan a College ofAgriculture , Sebele for aperio d of7 8day s(podding ) from 20t hAugus t to 5t hNovembe r 1997.Th e average monthly temperatures ranged from 27.5t o 32.8 °Cdurin g the day and 15.0t o 20.5 °Ca t night. Washed coarserive r sandwa suse d for growingth eplants .Th emai nchemica l characteristics of the sand were:p H 7.4; PBra y 3.8 mgP kg' ; organic carbon 0.02 (wt.%);CE C 1.00 (cmol (+) kg1); exchangeable cations (cmol (+) kg ): Ca 1.35, Mg 0.30, K 0.04 and Na 0.04. The determinations of the chemical characteristics of the sand are as described in Chapter 2. The treatments, 12P levels,wer e 0.207, 0.414, 0.828, 1.656,3.312 , 6.625, 13.25,26.5 ,53.0 , 79.5, 1 106 and 159m g P plant" for the whole growing period. Phosphorus (NaH2P04 dissolved in water) was added to the plants at 12 different rates according to the relative addition rate technique (RAR; Ingestad, 1986). With this technique P was added during the entire growth 35 Chapter 4 period in accordance with the plants actual growth rate, the latter varying from maximum to severely limited.Pattern so fmaximu mgrowt hrat ean dcorrespondin g Puptak eversu stim ewer e obtained by constructing a standard curve based on growth and shoot P concentrations of a previous greenhouse pot experiment (Chapter 5).Phosphoru s was weekly addedt o thepot s at different rates,increasin g withtim eaccordin g toth egrowt hrat eo fth eplants ,varyin g from low tohig h onth e bases of intended growthan d corresponding P requirement. Theexperimen twa sarrange di na completel y randomized designwit htreatment sreplicate d four times. Three seeds of 'Diphiri Cream' were sown inwhit epot s (five litre) containing eight kg ofsan dan dthinne dt oon eseedlin g 13 daysafte r sowing(DAS) .A minu sP nutrien t solutionwa s prepared with deionized water and applied weekly starting from sowing. The nutrient solution 1 used was of the following composition (mmol L ): 3.75 NH4N03; 1.25 K2S04; 1.00 MgS04; 1 and trace elements (umol I/ ) 82F e (Fe-EDTA); 46 B (H3B03); 9 Mn (MnS04.H,0); 0.3 Cu (CuS04.5H20);0. 8Z n(ZnS0 4.7H20) and0.0 1M o (Mo03).Th e nutrient solution was applied according to the RAR. technique (Ingestad, 1986) based on the N requirement of bambara groundnut determined from astandar dgrowt hcurv ean dshoo tN concentration s from apreviou s experiment (as for P above). Contrary to P, all the pots received the same amount of nutrient solution which increased with time according to the growth rate of the plants. The smaller quantities ofnutrien t solutiona tearl y stages ofgrowt hwer e added oncea wee k whileth e large quantities at later growth stages were distributed throughout the week. The pH ofth e nutrient solutionwa smaintaine d between 5.7an d6.0 . Waterstres swa sprevente d byapplyin g deionized waterdail yt oeac hpo t inadditio nt oth enutrien tsolution .A tadvance dplan t growthth e saucers under thepot swer e filled up with deionized water ornutrien t solution. At final harvest (78 DAS), leaf number was recorded and plants separated into shoots,roots , nodules and pods. The leaf bladeswer e separated from the petioles to determine the total leaf areape rplant . The leafare ape rplan t wasdetermine d by stacking 20full y matured leaves and cutting throughthe mwit ha on ecentimetr e diametercor kborer .Th eweigh t ofth e 20lea f blade discsan dtota l weighto fth elea fblade spe rplan twer etaken .A facto r relatingweigh to fth e leaf blade discst othei r areawa s determined anduse dt o calculate the total leaf areape r plant (leaf areape rplan t= tota l leafweigh tpe rplan tx (are ao fth e2 0 leaf bladedisc s (15.71 cm2) divided byweigh t ofth ediscs)) .Th eplan tpart swer eoven-drie d at70° Cfo r 24hour san d weight taken. The shoots were ground and analysed for Pa s described in Chapter 2. Data analysis was the same as in previous chapters. The critical P concentration was determined from the relation between shootdr ymatte rproductio n andshoo tP concentratio n anddefine d asth e concentration giving 90%o f itsmaximu m shoot drymatte r(DM) . 36 Phosphoruslevels inshoots 4.1.3 Results Plantgrowth and dry matter production From Figure4.1a- c itca nb esee ntha tP rate so f26. 5t o 159 mgP plant" 1significantl y increased number of leaves, leaf area and shoot DM per plant, relative to the lowest P rate. Beyond a supply of 106m g Pplant" 1furthe r increases were nolonge r significant. Between Prate s0.20 7 and26. 5m gP plant 1,th eplant swer estunte d andshowe dchloroti cblotche san dbrownis h spots onolde rleaves ,an da sgrowt hprogressed ,th echloroti carea so nlea fedge sbecam enecrotic .Th e Pdeficienc y symptomswer eonl yo nolde rleaves ,whil elate rmaturin g leavesha dn osymptoms . 2500, '2000- ra o. U) > CO H 1- 50 100 150 200 50 100 150 200 1 Phosphorus rate(m gP plan t ) Phosphorus rate(m g Pplant 1) o 50 100 150 200 Phosphorus rate (mg P plant1) Figure 4.1. Effect ofleve l ofphosphoru s application onth e number of leaves (a),tota l leaf area (b), and shoot dry weight (c)o f 'Diphiri Cream' plants at 78day s after sowing. 37 Chapter 4 Traces ofP deficienc y were also noticed onplant s given 53.0m gP plant 1, but none atP rates of 79.5 to 159 mg P plant'. The P effect observed on shoot growth was similar for root and nodule growth and pod development; the number of days to flowering and podding was not affected by Prate . Shootphosphorus concentration Shoot P concentration significantly increased with increased level ofP fertilizatio n and varied between0.0 7an d 0.20%.Th eincrease si nP concentratio nwer esignifican t atP addition shighe r than 53 mg Pplant" ' (Fig.4.2) .I n agreement with theobserve d visual Pdeficienc y symptoms atP rate sbetwee n0.20 7an d26. 5m gP plant"' ,shoo tP concentration swithi nthi srang ewer elo w and varied between 0.07 and 0.09%. Nutrient calibration curvesrelatin g shootP concentration st oshoo tDM ,numbe r ofleave san d leaf area per plant are shown in Figures 4.3a-c. The critical shoot Pconcentration s associated with a 10%reductio n in maximum shoot DM, number of leaves plant"' and leaf area plant', determined visually from the graphs inFigure s 4.3a-c, were essentially the same for the three parametersbein g0.15% ,0.15 %an d0.16% , respectively.Th ecritica lnutrien trang efo rth ethre e parameters was from 0.15t o 0.20%. 0 50 100 150 200 Phosphorus rate (mgP plan t "1) Figure4.2 .Effec t ofth e level ofphosphoru s application onshoo t phosphorus concentration of 'Diphiri Cream' plants at7 8day s after sowing. 38 Phosphoruslevels in shoots 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0.225 Phosphorus concentration (%) Phosphorus concentration (%) 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0.225 Phosphorus concentration (%) Figure 4.3. Relationship between shoot P concentration and shoot dry weight (a), number of leaves (b),an d leaf area (c)o f 'Diphiri Cream' plants at 78day s after sowing. 4.1.4 Discussion The critical shoot Pconcentratio n andrang e for bambara groundnut determined visually from the graphs were 0.15% and 0.15-0.20%,respectively . Stunted growth seemed to beth e major symptom of P deficiency of Bambara groundnut and was observed in plants with shoot P concentrations less than 0.10%. This demonstrates the important role of P in growth and development of bambara groundnut. Thecritica l levelo fa nutrien tmean stha ta ttha tleve ltha telemen ti sstil lslightl y limitinggrowt h (Smith, 1962). A shoot P concentration of about 0.15 % seemed to be critical for bambara groundnut andtherefor e anindicatio ntha tplant stha tproduc e biomasswit htha t Pleve l inthei r shoots are at least marginally supplied with P. In our field experiment, at 78 DAS, shoot P 39 Chapter 4 concentrations were around the critical Plevel ,bein g 0.15an d 0.14% for rainfed and irrigated treatments,respectivel y (Chapter2) .Th eplant s inou rpo texperiment swer e also suffering from Pdeficienc y with shootP concentration s of0.11-0.13 %(Chapter s 3.1 and 3.2). Probably inou r field experiment, wherether ewa sn orespons et o Pfertilization , Psuppl ywa s stilllimite d and broadcasting of the fertilizer did not bring it close enough to the roots of the seedlings at the critical time (Chapter 3.2). The critical shoot P level of bambara groundnut determined in this experiment has, however, demonstrated thatth eP statu so fBambar agroundnu tplant si nal lou rpreviou sexperiment s (pot andfield) wa s marginal. Asshow n in Chapter 2, this seems to be due not onlyt o a low soil P status,bu tt oa combinatio n of lowsoi lP an d low soil moisture content, probably strengthened by arelativel y poor root development ofbambar a groundnut (Chapter 3.1). 4.2 Shoot phosphorus levels in bambara groundnut plants infarmers ' fields in Botswana 4.2.1 Introduction ShootP concentration s inth efiel d experiment (Chapter 2), inwhic hbambar agroundnu t did not respond to applied P, were around the critical P level as determined in the sand culture experiment (Chapter 4.1). This might mean that the content of available P in the soil was probably not high enought o meetth e Prequiremen t ofth ecrop . Asa ninadequat e Psuppl ywa sencountere d ona soi l cleared from awel l developed bush fallow vegetation (Chapter 2), it may be expected that a more outspoken P shortage can occur on unfertilized permanently (frequently) croppedfarmers ' field. Tochec kthi s assumption,a surve y was conducted to assess the P nutritional status of the crop using the critical concentrations established inth e sand cultureexperimen t asa reference. 4.2.2 Materials and methods Forthi sstudy ,te nfarmer s growing creamcoloure dbambar agroundnu t wererandoml y selected fromth eKgatlen gDistric ti nEaster nBotswana ,withi na radiu so f6 0k mfro m BotswanaColleg e of Agriculture, Sebele (24°33' S; 25°54' E, 994 m). The farmers were selected by extension workers on the basis ofthei r interest and many years ofexperienc e withth e crop.Th e farmers wererequeste d togro wth ecro pa susua l and keeprecord s ofplantin g date and yield ofpod sa t harvest. Where the farmer did not keep a record of the planting date, it was estimated with information givenb yth e farmer. The farmers werevisite d at least once amont h to monitor the 40 Phosphoruslevels inshoots farming operations and to collect information on the cropping activities. The nearest weather station to the farms was at Sebele and the rainfall data from this station is used to generally describe the climatic conditions inth e survey area (Fig.4.4) . Planting inth e farms was done between 8t hNovembe r and 20t h December 1996an dth e area under bambara groundnut per farm varied between 0.08 and 2.17 ha. The main criterion determiningth edat eo fplantin gwa sth eavailabilit y ofrains .Th eare aunde rbambar a groundnut was determined with a measuring tape. Three farmers (30%) had broadcast the seed and ploughed to cover it with soil,whil e the rest planted inrow swit h either atracto r or an animal drawn planter after ploughing. None ofth e farmers applied fertilizer duringth e survey period, buton efarme r (No.2 ;Tabl e4.1 )applie dP fertilize r inth e previous seasonwhil e another (No. 9) about eight years ago. In all the farms, the crop was a pure stand planted on the flat land withoutridges . Plant and soil samples from each farm weretake n around 78day s after sowing (DAS),whic h was50 %poddin g stage(Ramoleman aet al., 1997) .Th esamplin gperio dwa sestimate d from the planting date asprovide d byth efarmers . Asampl e of sixplant s 10m apar t wastake n from the middle ofth eare aunde rbambar agroundnu t startingfiv emetre sfrom th eedge .Th e soil sample collected perfar m wasa composit e of2 4subsample stake nwit ha nauge r (20m mdiamete r and 30 cm long) to a depth of 20 cm.A t each of thespot s where the sixplant s were sampled, four soil subsamples weretaken .Thre esoi l subsampleswer etake n ata radiu s of 50c m around each plant selected for sampling,wit hth e fourth onth e spot whereth eplan t was grown. Theplan tshoot swer eoven-drie d at70° Cfo r2 4hour san ddr yweigh ttaken .Th esi xplant swer e then groundi na plan t grinder with a2 m m sieve tomak e one composite sample, subsequently analyzed for N, P and K, as described in Chapter 2. The soil samples were air dried for two weeks and later analysed for mineral nutrients asdescribe d in Chapter 2. Podproductio n estimatesfro m theare aunde rbambar agroundnu t ineac hfarm , asobtaine d from thefarmers , were converted toyield spe rhectare .Dat ao nplan t densities ofbambar a groundnut inth efarm s involved inth e studywer eno tcollected . Correlations ofparameter s measured from both plant and soil samples weredetermine d using the SAS (1985) statistical programme. 4.2.3 Results Soilphosphorus status Thesoi l Plevel s(P-Bray ) ofth ete n farms ranged from 2.8t o 9.1 mgP kg" ' withth ehighes t at farm No.2 ,th efar m whereP fertilize r wasapplie dth epreviou sseaso n(Tabl e4.1) . Onaverage , P-Bray ofth efield s was4. 4m gP kg" 1soil .Th edifferenc e betweenothe rsoi lcharacteristic swer e small (Table 4.1).Th e pH levels ofth e soilswer e rather low, onaverag e4.6 . 41 Chapter 4 Table 4.1. Some chemical characteristics of soils from 10differen t bambara groundnut farms. Farm pH P-Bray Org. CEC Exchangeable cations (mg kg-') carbon (cmol (+) (cmol (+) kg-1) (wt. %) kg1) Ca Mg K Na 1.Masul e 5.06 3.33 0.17 3.96 1.49 0.74 0.33 0.04 2. Lekorwe 4.22 9.08 0.22 280 0.52 0.32 0.33 0.04 3. Mosekiemang 4.62 2.77 0.19 2.08 0.87 0.32 0.16 0.04 4. Motswasele 4.43 2.97 0.13 2.08 0.42 0.28 0.20 0.04 5. Thamage 4.50 2.88 0.16 2.04 0.92 0.24 0.20 0.04 6. Gare 4.67 3.33 0.25 2.12 0.70 0.26 0.16 0.04 7. Masupu 4.91 4.10 0.26 2.44 1.01 0.42 0.29 0.04 8.Ntlhwasan e 3.95 3.78 0.26 1.60 0.11 0.15 0.12 0.04 9. Tlhase 4.11 7.18 0.19 1.96 0.60 0.28 0.20 0.04 10. Maotsela 5.50 4.27 0.13 2.68 1.20 0.51 0.29 0.04 Average 4.6 4.4 0.20 2.40 0.78 0.35 0.23 0.04 Shootnutrient status Theshoo t Pconcentration s ranged from 0.10t o 0.16%wit hth ehighes t level again on farm No. 2,th efar m where Pfertilize r was applied thepreviou s season (Table4.2) .Tota l P inth e plants (represented by the shoot P content in mg Pplant 1) shows variations with a factor of 4. The variation inshoo t Pconten t was however duet o variation in shoot DM(Tabl e 4.2). Variations in shoot N and K concentrations were observed, but the lowest levels measured being respectively 2.14 and 1.56% areadequate ,a scompare d tomean so f2.7 0 and 1.13% atth e same growth stage from apreviou s field experiment (Chapter2) . Estimatedpo dyield svarie d between 635an d 875k gha -1(Tabl e4.2) .Thi svariatio nwa smuc h lesstha nth e observed variation in shoot DMplant -1. 42 Phosphoruslevels inshoots Table 4.2. Mineral nutrient status of plant samples collected at 78 days after sowing and pod yield estimates atfina l harvest from 10differen t bambara groundnut farms. Farm Shoot Shoot Shoot Shoot TotalP Yield DM(g P (%) N (%) K(%) (mg (kgha"' ) plant"') plant')2 1.Masul e 76.3 0.12 2.35 1.88 92 875 2. Lekorwe 23.8 0.16 2.60 2.26 38 778 3. Mosekiemang 54.8 0.15 2.33 2.25 82 806 4.Motswasel e 54.6 0.11 2.14 2.19 60 759 5.Thamage ' 19.1 0.12 2.30 2.11 23 n.a. 6. Gare 28.9 0.11 2.35 1.64 32 635 7. Masupu 47.5 0.13 2.60 1.96 62 875 8.Ntlhwasan e 29.4 0.11 2.62 1.78 32 673 9. Tlhase 36.7 0.10 2.38 1.56 37 740 10. Maotsela 25.2 0.13 2.09 1.85 33 700 Average 39.6 0.12 2.38 1.95 49 760 'The cropwa s later destroyed byporcupine s and ground squirrels. 2 Shoot Pconten t No correlations were found between soil P levels onth e one hand and shoot DM and shoot P concentration onth eothe rhand .Correlation s betweenshoo t Pconten t andshoo tDM ,shoo tD M and pod yield, soil organic matter content and shootN concentration were significant. 4.2.4 Discussion The average soil P content of 4.4 mg P kg"' found in the ten bambara groundnut farms in Botswanai sgenerall yconsidere dt ob elow ,compare dt oa minimu m ofP-Bra y of 10 mg P kg"1 required for sorghum (Beynon, 1991).Shoo t Pconcentration s were also low (Table4.2) . Only intw ofield s shootP concentration sreache dth eleve lo f0.1 5an d0.16 %whic hcorrespond s with the critical shoot P level for bambara groundnut. In all other fields, plants showed shoot P 43 Chapter 4 concentrationsbetwee n0.1 0 and0.13 % whichcorrespon d withmoderat et o severe P deficiency. This meanstha t inal lthos efields P suppl y was suboptimal. Theshoo tP conten to frainfe d plantsfrom ou rfield experimen t(Chapte r2 )wa s 109m gP plant"' , beingabou ttwic ea shig ha sth eaverag eo fplant sgrow no nth etraditiona l farms investigated in this study (Table4.2) .Plan tgrowt han d shoot Pconten t of only one farm (No.l) was similar to thato fou rrainfe d treatmentwit h ashoo tdr yweigh t of 76.3 gplant 1. Probably duringth e field experiment (1995/96)th e rainfall distribution was better than in the 1996/97 growing season, when the survey was carried out (Fig.4.4). Differences can not be explained on the basis of differences intota lrainfal l betweenth eyear s 1995/96an d 1996/97,bein g respectively 633 and 676mm . (a) (b) Figure 4.4. Rainfall distribution during the 1995/96 (a) and 1996/97 (b) planting seasons as recorded at Sebele Botswana Meteorology Station. (— bambara groundnut growthperiod) . Sinceth eshoo tP conten twa sdirectl yrelate dt oth e shoot dry matter weight (growth),th e latter probably acting as a sink for P uptake, differences in shoot P content between our field experiment and most of the farm-grown plants may also be due to differences in cultivation technique. Plant density, weed control and crop protection in our experimental plots were probably better than in farmers' fields. Finally, whether the low seed yields aspresente d inTabl e4. 2 are adirec t result of ato o lowP content of the soil is questionable. Probably without irrigation P availability was very low leading to lowP uptak e and Pdeficiencies . Adirec t water stress can also be involved. Besides Pdeficiency , also alo wplan tdensit y contributes to the lowyield s in farmers fields. Therefore, yield increasesca n only be achieved by irrigation incombinatio n with increased plant density, and subsequently followed by P fertilization. 44 Responseto nitrogen fertilizer Chapter5 Response of bambara groundnut (Vignasubterranea (L.) Verde.) to mineral nitrogen fertilizer application ifgrow n under non-limiting phosphorus conditions 5.1 Introduction Nitrogen fixation by legumes depends on different soil conditions such as the availability of mineral nutrients like nitrogen (Alios and Bartholomew, 1955;Hartfiel d etal., 1974;Graha m and Scott, 1984)an d phosphorus (Aguilar and Van Diest, 1981; Israel, 1987).I n our previous studieswit hbambar agroundnu t growni nP-poo rsoils ,however ,phosphoru s fertilization didno t affect nodulationno rshoo tN concentratio n (Chapters2 an d 3). ShootN concentration swer eno t affected even when shoot DM was increased by improved soil moisture conditions or by P application to the soil. Evidently under the prevailing conditions total N supply, consisting of N2fixed b yth eplant splu sminera lnitroge nfro m mineralizationwa ssufficien t tomee tth eplant s Nrequiremen t alsoa thighe rdr ymatte r(DM )production .However ,i nal lpreviou sexperiments , carried outbot hi nth efield an di npots ,bambar agroundnu t plantswer e growing undermargina l P conditions with shoot Pconcentration s around or belowth e critical level, irrespective ofth e treatment( Pfertilization , irrigation)o rD Mproduction .Thi sraise sth equestio nwhethe rth eplan t Nsuppl y will stillb eadequat e ifbambar a groundnut plants aregrow n under conditions where Pwil lb en olonge rmargina lo rdeficient . Undersuc hconditions ,highe rD Mproductio nwil lno t only increaseth eplan tN requirement , but also theN 2 fixation might behave differently. If Pi n the plant is directly involved inth eN 2fixation proces s (Israel, 1987),a lack of response ofN 2 fixation to P fertilization can be expected as long as the plant P status is not affected by P fertilization and remains at the deficient level. Such conditions prevailed in all previous experiments. Therefore, the main objectives of this experiment are to find out (i) whether with bambara groundnut the capacity ofth eN 2fixatio n process,possibl y supplemented with some native soil mineralN ,i shig henoug ht omee tth eplan tN requiremen t alsounde rnon-limitin g Pconditions , and(ii )whethe r atinterna l Pconcentration s exceeding the critical level,th e amount of fixed N2 increases with increasing internal P status. This might indicate the P-dependency of the N2 fixation process. Toanswe rth eabov equestions ,a po t experiment wascarrie dou twit hbambar a groundnutplant s grown in a P-poor soil fertilized at sowing with or without mineral N in combination with different rates of P,th e latter upt o levels leading to luxurious internal Plevels . 45 Chapter 5 5.2 Materials and methods The experiment was conducted from the 2n d July to 7t hNovembe r 1996i n agreenhous e at Botswana College of Agriculture with the same soil asdescribe d in Chapter 3.1. The average monthly temperatures varied between 30-34°C during theda y and 18-21°C at night. Thetreatment s composed oftw onitroge nlevel s( 0an d35 0m gN pot" 1; onweigh tbasi s of2 00 0 000k gsoi lha' 1correspondin gt o 100k gN ha 1)an dfive phosphoru s levels(0,105,210,31 5an d 420 mg P pot1 corresponding to 0, 30,60 ,9 0an d 120k g Pha" 1). Whitefive-litre pot s (25c m diameterx 2 5c mheight )holdin gseve nk go fsoi lwer euse d forgrowin gth eplants . Smallerpot s than inpreviou sexperiment swer euse dt ohav ea hig h root density and increased contact ofth e plantroot swit happlie dP .Thi swa sdon et oincreas eP uptak eb yth eplants .Th etreatment swer e arranged in acompletel y randomized design with threereplicates . Single superphosphate was crushed andthoroughl y mixedwit hth esoil .Ure awa sdissolve d in25 0m l ofwate r and applied evenly over the soil immediately after sowing. No mineral nutrients other than P and N were applied.Thre eseed so f 'Diphiri Cream' weresow npe rpo t andthinne dt oon e seedling 19 days after sowing(DAS) .Th eexperimen twa sstarte dwhil ei twa sstil lwinte rwhe nligh tintensit ywa s lowan dda ylengt hwa s sometimes lesstha n 12hours .T oachiev e optimal growth Philips light bulbs (75 W) were used for increasing light intensity and day length for the period between sowingan d 115DAS .Th ebulb swer e automatically switched onan d off at6:0 0 and at 18:00 hours,respectively . Thesoi lmoistur econten twa smaintaine d atfield capacity . From flowering onwards, water was poured into saucers to avoid damaging flowers. The plants were sprayed with Benlate (a.i.Benomyl) and Garden Ripcord (a.i.Cypermethrin) to control mildew disease and aphids,respectively . Earthing up(coverin g ofpod swit h soil)wa s done 70DAS . Sampleswer etake na tth evegetativ e growth stage (35DAS) ,50 % podding (76DAS ,whe nhal f ofth eplant sha d at leaston epo d which was 0.5 cm long),an d atth e final harvest when leaves turned yellow. Thefinal harves t ofplant s that received additional mineral nitrogen was at 115 DASwhil ethos ewithou tN fertilizatio n at 128DAS .Dat arecorde d wereshoot ,roo t andnodul e dryweights ,numbe r anddr yweigh t ofpod s and seed yield perplant . Root and nodule weights weretake n only at 76DAS .Th eplan t parts (shoot, root and nodules)wer e oven-dried at 70°C for 24hours ,whil epod swer e sun dried for two weeks.Th e dry shootswer e ground in aplan t grinder with a 2 mm sieve, and analysed for N and P as described in Chapter 2. Statistical analysis ofth edat awa s asdescribe d in Chapter 2. 5.3 Results Drymatter production Shootdr ymatte r(DM )afte r Pfertilizatio n (withoutadditiona lN )wa sa tal lsamplin g datestwic e ashig ha sth econtro l (-P;- Ntreatment )(Tabl e 5.1).A nincreas ei nP rat ebeyon d 105m gP pot"1 46 Response tonitrogen fertilizer did not further increase shoot DM at all sampling periods. At early podding (76 DAS), N fertilization significantly increased shoot DMo f Pfertilize d plants by 15-32%,relativ e to -N. Therewer en o significant differences inshoo t DMa t3 5DA San d final harvest between -Nan d +Nplant sneithe r with orwithou t Pfertilization . Theplant sgive nextr aminera lN mature d early (115 DAS),compare d to -Nplant s (128DAS) . Table 5.1.Effec t ofphosphoru s andnitroge napplicatio n onshoo tdr yweigh to f 'Diphiri Cream' plants atdifferen t growth stages (days after sowing;DAS) . Phosphorus Shoot dryweigh t (gplant"' ) (mgP pot') 35DA S 76DA S 115/128 DAS* -N +N -N +N -N +N 0 1.18 1.44 9.23 9.82 12.46 12.83 105 1.99 1.95 20.20 23.23 22.18 24.87 210 1.97 1.98 21.34 25.63 24.94 24.76 315 1.97 2.16 19.9 26.34 23.32 24.74 420 2.27 2.29 21.30 26.91 24.24 26.31 Mean 1.88 1.96 18.41 22.39 21.43 22.70 C.V. (%) 13.1 12.0 11.7 S.E.(2 0 D.F.) 0.10 1.0 1.05 L^Doos 0.30 2.95 3.10 +N plantsharveste d at 115,whil e -Na t 128DAS . Relative toth e control, increases indr yweigh t after Pfertilizatio n (-N) were low for roots (about 25%),bu tth e nodule dry weight was increased two and halftime s after P fertilization (Table 5.2). Effects ofP fertilization onroo t growthwer e significant withmaximu m root dry weight at 105m g Ppot" 1an dthe n decreasing with further increase inP rate .Wit hth e exception ofth e zero Ptreatment , nodule dryweight sa tdifferen t Prate swer eno t significantly different. Nitrogen application had no significant effect onroo t andnodul e dry weights. 47 Chapter 5 Table 5.2.Effec t ofphosphoru s and nitrogen application onroot s andnodul e dryweigh t of 'Diphiri Cream' plants at7 6day s after sowing. Phosphorus Dryweight s ({ I plant"1) 1 (mgP pot ) Roots Modules -N +N -N +N 0 5.20 4.38 0.69 0.48 105 7.81 7.85 1.77 1.42 210 6.58 7.11 1.79 1.68 315 5.89 5.73 1.56 1.50 420 5.66 7.14 1.83 1.56 Mean 6.23 6.44 1.53 1.33 C.V. (%) 15.4 19.8 S.E. (20 D.F.) 0.40 0.12 LkUo.05 1.17 0.34 Podand seed yield Number ofpod san d pod dryweigh t after Pfertilizatio n (-N) were on average increased three and four times,relativ e to control,respectivel y (Table 5.3).Phosphoru s fertilization beyond a rate of 105m g Ppot" 1di d not further increase thenumbe r ofpod san dpo d dry weight. Seed yield after Pfertilizatio n (-N) was onaverag e increased four andhal ftimes , relativet o control, with nodifferenc e between Prates .Nitroge n application had no significant effect on number ofpods ,po d dryweigh t and seed yield (Table 5.3). 48 Responseto nitrogen fertilizer Table 5.3.Effec t ofphosphoru s and nitrogen application onnumbe r ofpods ,po d and seed dry weight of 'Diphiri Cream' atfinal harves t (115an d 128 DAS)*. Phosphorus Number ofpod splant" 1 Podwt .(g plant"1) Seed wt. (g plant1) (mgP pot"' ) -N +N -N +N -N +N 0 31.7 24.7 11.9 6.0 8.8 4.5 105 91.0 74.7 48.4 44.3 40.1 36.9 210 92.7 81.0 45.0 46.0 37.8 38.4 315 89.0 105.0 50.2 49.6 42.0 41.6 420 90.0 71.3 45.9 38.6 38.7 32.5 Mean 79.0 71.3 40.3 36.9 33.5 30.8 C.V. (%) 24.1 21.0 22.7 S.E. (20 D.F.) 7.39 3.30 2.97 LSD005 21.8 9.74 8.77 * +Nplant s harvested at 115,whil e -N at 128DA S Shootnutrient status Phosphorus The increases in shoot Pconcentratio n duet o Pfertilizatio n (-N) werever y significant at all threeharvest s (Table 5.4).Th eeffect s ofth evariou s Prate s were significantly different from control at all sampling periods,wit h the highest rate giving thehighes t shoot P concentration. The -Nplant s atth ehighes t Prat edevelope d necrosis onlea f edgeswhic h started at flowering and continued until early podding stage,whic hwa s suspected tob e duet o exceptionally high shoot Pconcentrations . Thehig h shoot Pconcentration s for thistreatmen t at 35 and 76DA S confirm this.Fo rth ecorrespondin g +Nplants , shoot Pconcentration s were equally higho r lower and plantsdi dno t showlea fnecrosi s (Table 5.4).Ther ewa sa stron g decrease in shoot Pconcentratio n between 76DA San dfinal harvest . Shoot Pconten t followed a similar trend asth e shoot Pconcentration^'wit hth e highest shoot P concentration giving thehighes t shoot Pcontent . Phosphorus fertilization at aleve l 1105 mg Ppot -1 49 Chapter 5 increased the shoot Pconcentration s beyond thecritica l shoot Prang e (0.15-0.20%) for bambaragroundnu t asdetermine d from asan d culture at 78DA S (Chapter 4). Table 5.4. Effect ofphosphoru s andnitroge n application on shootphosphoru s concentration of 'Diphiri Cream' plants at different growth stages (days after sowing, DAS) Phosphorus Shoot Pconcentratio n (%) (mgP pot') 35DA S 76DAS ' 115/128DAS 2 -N +N -N +N -N +N 0 0.09 0.12 0.11 0.12 0.07 0.08 105 0.21 0.19 0.20 0.17 0.06 0.06 210 0.28 0.27 0.21 0.20 0.09 0.08 315 0.35 0.31 0.27 0.20 0.10 0.10 420 0.39 0.40 0.32 0.25 0.13 0.14 Mean 0.26 0.26 0.22 0.19 0.09 0.09 C.V. (%) 11.8 11.7 15.9 S.E. (20 D.F.) 0.013 0.010 0.006 I^DQOS 0.12 0.03 0.02 1N applicatio n significant only at 76DAS . 2+ Nplant s harvested at 115,whil e -N at 128DAS . Nitrogen Phosphorus fertilization (-N) significantly increased shootN concentration at 76DAS , relative tocontro l (Table 5.5).A n increase inP rat e beyond 105m g Ppot" 1di d not further increase shootN concentrations . At final harvest, Pfertilizatio n had no significant effect on shootN concentrations. At 76 DAS,N fertilizer depressed the shootN concentration s especially when Pfertilize r was applied (Table 5.5). Also atth efinal harvest , Nfertilizatio n inth e presence of P decreased 50 Responseto nitrogen fertilizer shootN concentration, withth ehighes t Prat egivin gth e lowestN concentration. The lower shootN concentrations at7 6DA S andfinal harves t in+ Nplant s corresponds withth e yellowing of leavesa sobserve d betweenflowering an dearl ypoddin g stages,whic h intensified atth epo dfilling stag e and ledt o early maturity. Table 5.5.Effec t ofphosphoru s andnitroge n application on shootnitroge n concentration of 'Diphiri Cream' plantsa tdifferen t growth stages (days after sowing-DAS) Phosphorus ShootN concentration (%) 1 (mgP pot ) 35DA S 76 DAS1 115/128DAS 2 -N +N -N +N -N +N 0 3.11 3.16 2.18 1.60 1.72 1.97 105 2.92 3.38 2.89 2.03 1.51 1.52 210 3.12 3.06 2.47 1.87 1.67 1.40 315 2.81 3.03 2.77 1.77 1.58 1.43 420 2.84 3.08 2.72 1.85 1.78 1.37 Mean 2.96 3.20 2.61 1.83 1.65 1.54 C.V. (%) 11.1 11.6 8.3 S.E. (20 D.F.) 0.14 0.10 0.05 LoD0o5 n.s. 0.31 0.23 1N application significant at7 6DA S andfinal harvest . 2+ Nplant sharveste d at 115,whil e -Na t 128DAS .n.s .= no t significant. The increases in shootN conten t after Pfertilizatio n followed asimila rtren d as shootN concentrations. Average increases inshoo tN contents after Pfertilizatio n (-N) were 30,36 0 and 172m gN plant" 1,relativ e to control, at 35, 76DA San d final harvest, respectively (Table 5.6). At 35an d 76 DAS,a furthe r increase inP rat ehighe rtha n 105m gP pot" 1di d not change the shootN content. Atfinal harvest , therewa s asligh t increase in shootN content with increase inP rate . Contrary to shoot DM,N fertilization significantly decreased the shootN content at7 6DA Sb y 18%,relativ e to -N(Tabl e 5.6). This decrease in shootN content isi n agreement withth e decrease in shootN concentratio n at 76DA S for +Nplant s (Table 5.5). 51 Chapter 5 Atfinal harvest ,N fertilization hadn o significant effect on shootN content. However, it has to be considered thatN inth epod sha s not been included inth e shootN content s aspresente d inTabl e 5.6.Thi sca nquantitativel y bever y important atth efinal harvest . Table 5.6. Effect ofphosphoru s and nitrogen application onshoo t nitrogen content of 'Diphiri Cream' plants atdifferen t growth stages (days after sowing;DAS) . Phosphorus ShootN conten t(m gN plant" 1) 1 (mgP pot" ) 35DA S 76DAS ' 115/128DAS 2 -N +N -N +N -N +N 0 31.0 45.5 201.2 155.9 213.3 252.8 105 59.1 65.8 584.5 472.5 337.7 379.7 210 61.7 60.5 537.9 482.4 414.9 343.5 315 55.2 71.0 551.0 467.5 367.8 354.3 420 64.1 71.4 577.6 498.4 431.3 362.3 Mean 55.5 62.9 490.4 415.4 353.0 338.5 C.V. (%) 20.0 19.3 13.8 S.E. (20 D.F.) 4.8 35.7 19.4 '-,^>^-'005 14.3 105.3 57.3 1N application significant. 2+ Nplant sharveste d at 115,whil e -Na t 128DAS . 5.4 Discussion Thepositiv e response of shoot,roo t andnodul e dry weights and shoot Pan dN concentrations to Pfertilization , asfoun d inthi s study, agrees withth e critical Pleve l for bambara groundnut (Chapters 4).I n previous experiments shoot DMan d seedyield s were increased byP application onlyunde rpo tcondition swit h shootP concentration s remaining around or below the critical Plevel . The shoot Pconcentration s inthi s study reached levels aboveth e critical P range for bambara groundnut (Chapter 4),whic h meanstha tth e plants weren o longer 52 Response tonitrogen fertilizer suffering from Pdeficiency . Consequently, growth wassignificantl y increased byP fertilization. Inthi s study, improved Pavailabilit y increased noduledr yweight san d shootN concentrations (Table 5.2 and 5.5). Thismigh t pointt o apositiv e effect of improved P nutrition onN 2 fixation. Plants fertilized with Pbu tno t withN showed optimum growth and seed yield at shoot Pan dN concentration s atearl y podding stage of0.20 %an d 2.89%, respectively (Tables 5.1,5.2 , 5.3,5.5) .However , shootN concentrations in ourpreviou sfield experiment (Chapter 2)wer e similar, being 2.70%a t7 8DA Sbot hwit h and without irrigation. This meanstha t plants intha tfield experimen t were also adequately supplied with N, though shoot Pconcentration s were around orbelo wth e critical level (0.15%P). Evidently phosphorus seemed nott o have limitedN nutritio n intha tfield experiment . Onth e other hand, shootN concentrations inpreviou spo t experiments (Chapter 3)wer e much lower and varied between 1.68an d 2.03%,wit h shoot Pconcentration s always varying around or below the critical level. Thismigh t pointt o areduce dN 2fixation a tsuboptima l Psupply . Evidently, results ofth e variousexperiment s are not inagreemen t with each otherwit hrespec t to the possible role of Pnutritio n inth eplan tN 2fixation capacit y and further research onthi s subject is needed. Application ofN at sowing under non-limiting Pcondition s did not affect shoot, rootan d nodule dry weights norN nutritio n ofbambar a groundnut. This meanstha t under the prevailing conditions, bambara groundnut plantswer e ablet o meetthei rN requirements through N2 fixation, inadditio n to mineralN from the soil.I n fact mineral N fertilizer had a negative effect which ledt oearl ymaturit y andlo wnumbe r ofpod san d lowsee d yields (Table 5.3). Plants given mineral fertilizer Na t sowing seemed to dobette r atth e initial phase of growth and were during thatperio d probably well supplied withN originatin g from the soil N, from N fertilizer andt o alimite d extent from N2 fixation. This isshow nb y the higher shoot DM withthi streatmen t at 76DA S(Tabl e 5.1). However, later onminera lN inth e soil seemed to be exhausted andplant s apparently did nothav eth epotentia l to switch onth e very shortter m toN 2 fixation. This ledt o growth stagnation of+N plants.Tha tthi sproces s has already been started at7 6DA S isshow nb yth e lowershoo tN concentrations andlowe r shoot N content of+N plantsa ttha ttim e (Tables 5.5 and 5.6).Plant s fed with extraminera l N fertilizer changed from aluxuriou sN t opoo rN supply with acorrespondin g early maturity. WithN 2fixing plant s (-N plants)N supply steadily goeson ,wit h acontinuou s growth and retarded maturity. Thoughnodul e dryweight so fplant swit h and without extraminera lN fertilizer were similar at 76DAS ,th e lowshoo tN concentration and low shootN content of the former indicate thatN 2 fixation capacity wasreduce d after N fertilization. Further, N fertilization seemedt ohav ea negativ e effect onpo d set(Tabl e 5.3).Th elo wnumbe r ofpod s and lowpo d yield for the+ Ntreatmen t inth e absence ofP coul d beresponsibl e for the high shootN concentration atfinal harves t (Table 5.5).Th ehig h shootN concentratio n for this treatment may bea n indication thatth epoo r pod set limited thereallocatio n of internal N from the shoott oth epods .Therefore , mineralN fertilization tobambar a groundnut seemed nott o benecessary , also not under non-limiting Pconditions . 53 Chapter 5 About thecontributio n of soil mineralN t oth eN nutrition ,mor edetaile d work hasbee n carried out (Chapter 6). However, based onth e loworgani c matter content ofth e soil used (Chapter 2),N delivered toth eplan t viamineralizatio n can beexpecte d to below . Consequently, N inplant swithou t Nfertilizatio n will mostly originate from N2 fixed by the plant. Finally, inbambar agroundnu t grownunde r non-limiting Pconditions ,nitroge n fixation possibly supplemented with somenativ e soil mineral Nseeme dt ob e sufficient to meetth eN requirements for optimal growth.Unde rth eexperimenta l conditions,applicatio n ofminera l N fertilizer tobambar a groundnut plantsa t sowing didno t improveth eplan tN nutrition nor growth.I tonl yresulte d inearl ymaturin go fplants . 54 Effectof time of nitrogen application Chapter6 Response of bambara groundnut (Vignasubterranea (L.) Verde.) to mineral nitrogen fertilizer. Effect oftim eo f application 6.1 Introduction In the previous pot experiment (Chapter 5) it was found that under non-limiting phosphorus conditions growth of bambara groundnut did not respond to mineral N fertilizer if applied at sowing time. In the literature, however, there is evidence that during certain developmental stagesbambar agroundnu ti slikel yt oexperienc ea shortag eo fN .On eo fthos estage si sth eonse t ofpodding ,whe na majo r shift inassimilat epartitionin gtake splac efro mvegetativ epart st opod s (Brink, 1998; Chapter 2). This is likely to have implications for the supply of nodules with assimilatesan dthi sma ylimi tnodul edevelopmen tan dit sN 2fixin g capacity.I tha sbee nreporte d for other leguminous crops that N2 fixation can not always meet the plant N requirement for optimal yield, andtha t with low soilN levelsth eyiel dpotentia l canb e limited (Harper, 1974; Bhangooan dAlbritton , 1976;Ingestad , 1980).Flowerin g might beanothe r stagedurin g which Nsuppl y isimportant .Bot hi ncowpe aan dsoybea nN applicatio na tflowerin g hasbee nreporte d to retard leaf senescence leading to an extended vegetative growth period and increased seed yield (Elowad and Hall, 1987;Isfa , 1991). These considerations prompted the ideat o investigate the effect ofN fertilization on bambara groundnut if applied at the following development stages: at sowing (as in the previous experiment),a tearl yvegetativ egrowth ,a t(50% )flowerin g and(50% )podding .T o separateth e contributionso f soilminera l Nan dN 2 fixation toth eN nutritio n ofbambar a groundnutplants , treatments with sterilized and unsterilized soil were incorporated in the experiment. For the sterilized soiltreatmen t itwa s assumed thatth eplant swoul d becompletel y dependent on soil mineralN fo r growth.Therefore ,thi sstud ywa sconducte dt odetermin eth epotentia lo fbambar a groundnut to utilize mineral N fertilizer applied at different growth stages, and to assess the contribution of soil mineralN t oN nutritio n ofbambar a groundnutplants . 6.2 Materials and methods Apo texperimen twa sconducte d from 26t hNovembe r 1996t o 3r dApri l 1997o nth e grounds ofth eforestr y nurseryo fBotswan aColleg eo fAgriculture ,wit ha simila rsoi la suse di nChapte r 3.1. Maximum and minimum temperatures varied between 26.5-32.4 °C and 16.3-18.7 °C, respectively. 55 Chapter6 There were six treatments as indicated in Table 6.1. The treatments were arranged in a completely randomized designwit hfive replicates . Forth evegetativ e stage,N fertilization was at2 8day safte r sowing(DAS) ;50 %flowering stag ewa sdefine d aswhe n50 %o f theplant sha d atleas ton eope nflowe r andthi swa sa t4 5DAS ;50 %poddin g stagewa swhe nhal fo fth eplant s hada tleas ton epo dwit ha lengt ho fabou t0. 5cm ,an dthi swa sa t6 9DAS .Fo rth esterilize d soil treatment, the soil was sterilized bywettin g and heating at 80° Cfo r two and half hoursusin g aCamplex -Electri c Soil Sterilizer, Cat.No .HD5112 .Whit eten-lite rpot s (diameter 35c m x3 5 cmdepth )holdin g 14k go fsoi lwer euse dfo rgrowin gth eplants .Crushe d single superphosphate applieda ta rat eo f42 0m gP pot" 1(o nweigh tbasi s of200000 0k g soil ha"1correspondin g to 60 kgP ha' 1)wa sthoroughl y mixed withth e soil.Ure a ata rat eo f 525m gN pot" 1 (corresponding to 75 kg N ha"1) was dissolved in 250 ml water and evenly applied over the soil at different growth stages.N o mineral nutrients other than Pan dN were applied. Three seeds of 'Diphiri Cream' weresow npe rpo tan dthinne dt oon eseedlin g at 17 DAS.Th esoi lmoistur econten twa s maintained atfiel d capacity.Afte r flowering, water waspoure d into saucerst o avoid damaging the flowers. Samples were taken at different stages of growth as described in Table 6.1. The final harvest included allth etreatment san dtoo kplac e 128DA Swhe nth eleave swer eyellow .Dat arecorde d weredr yweigh to fplan tcomponent s (shoot,roots ,nodule san dpod swher eapplicable) ,numbe r of nodules and pods. Shoots,root s and nodules were oven-dried at 70 °C for 24 hours, while pods were air dried for two weeks.Th e shoots were ground in aplan t grinder with atw o mm sieve,an d analysed for Na sdescribe d in Chapter 2. Statistical analysiso fdat awa s carried out asdescribe d in Chapter 2. Table 6.1. Description ofth e different treatments andtime s ofplan t sampling. SI =vegetativ e stage;S 2= 50%flowerin g stage;S 3= 50% podding stage;S 4= fina l harvest. Treatment Samplingtime 1 Tl (noN fertilization; control) SI; S2; S3; S4 T2 (noN fertilization; soil sterilization) S1; S2 ;S3 ; S4 T3( N fertilization at sowing) S1; S2;S3 ; S4 T4( N fertilization atvegetativ e stage;2 8DAS ) S2;S3 ;S 4 T5 (N fertilization at 50%flowering ; 45DAS ) S3;S 4 T6( N fertilization at 50%podding ; 69DAS ) S4 ' Withtreatmen t T2 sampling S2an d S3wer e delayed withtw o weeks compared to other treatments. 56 Effectof timeof nitrogen application 6.3 Results Drymatter production Compared toth e control treatment (Tl), nitrogen fertilization did not significantly affect shoot dry matter (DM),irrespectiv e oftim e ofapplicatio n (Table 6.2). However,N application at sowing (T3)tende d to giveth ehighes t andN applicatio n at 50% flowering (T5)th e lowest shootdr y weight atth efinal harvest . Ingenera l theplant s given extra mineral N at 50%flowerin g (T5)mature d earlier thanplant swit h othertreatments . Though sterilization (T2) increased shoot DMb y 89%a tth e 50%flowering stage ,a t allothe r stages sterilization had no significant effect on shoot DM(Tabl e 6.2). Onth e otherhand , sterilization significantly affected flower initiation. The 50%flowering an d also the50 % podding stages were delayed withabou ttw o weeksb y sterilization. Thismigh t explain the high shoot DM at 50%flowering wit h the sterilization treatment. Nodulation Soil sterilization significantly reduced nodule formation (number and biomass),bu t did not inhibit itcompletel y (Table 6.3).I n alltreatment s nodule number andnodul e dry weight per plant increased withtime ,bu t in sterilized soil the increase wasmuc h less.A t 50% flowering, nodule number and nodule dryweigh t after sterilization weretw o and ahal ftime s lower than that ofth econtrol .A t 50%poddin g therewa s still aclea r negative effect on nodulation. Plants grown in sterilized soil had onaverag e 170nodule s andth e control plants 400nodules . Datao n nodulation atfina l harvest aremissin g because mostnodule swer e decomposed at that time. Effects ofN fertilization onnodulatio n differed withplan t age.A t 50% flowering stage,nodulatio n was not significantly affected byN fertilization. However, mineral N fertilizer tended to reduce nodulation atth e 50% poddin g stage,bu t only ifapplie d late during the growing season. Podand seed yields Atfinal harvest , plants fertilized withN a t sowing (T3)gav e asignificantl y lower number of podspe rplan t (Fig. 6.1a), while soil sterilization did not affect pod number. Total pod dry weight perplan t (shell plus seed)a tfinal harves t gavea similar pattern aspo d number (Fig. 6.lb) . Again plants fertilized withN atsowin g yielded the lowestpo d DM,bu t duet o agrea t variability inpo d and seed dry weights,difference s between treatments were not statistically significant (P<;0.05). 57 Chapter 6 Table6.2 . Effect ofa singl edos eo fnitroge n applied atdifferen t growth stages and soil sterilization on shootdr yweigh t of 'Diphiri Cream' plants atvegetativ e (28 DAS),50 % flowering (45DAS) ,50 %poddin g (69DAS )an dfinal harves t (128DAS) . Treatment Shoot dryweigh t (gplant 1) Vegetative 50% 50% Final stage flowering podding harvest Control (Tl) 1.03 6.56 18.3 22.9 Sterilized soil (T2) 0.79 12.31 15.51 27.0 N at sowing (T3) 1.32 7.01 20.1 34.1 Na tvegetativ e stage (T4) - 7.70 20.4 28.8 N at 50%flowerin g (T5) - - 21.1 18.4 N at 50%poddin g (T6) - - - 23.6 C.V. (%) 25.4 33.0 21.4 30.8 S.E. 0.12 1.24 1.83 3.56 L0D005 0.38 3.73 n.s. n.s. n 15 20 25 30 1Ther ewa sa 2 week sdela y inharvestin gthes etreatment s at 50%flowerin g and 50% podding, n.s.: not significant; w: number ofobservations . ShootN status Effects oftim eo fN applicatio no nshoo tN concentratio n arepresente d inTabl e 6.4.A tth e50 % flowering stage,N fertilization gavea significantl y highershoo tN concentratio n ifapplie d atth e vegetative stage (T4). At the final harvest, shoot N concentrations were increased by N fertilization if applied at sowing (T3). Soil sterilization (T2) significantly affected shoot N concentration only at the 50%flowerin g stage. At this stage, shoot N concentration of plants grown in the sterilized soil was 60%o f the control. Furthermore, with all treatments shoot N concentrationsdecrease dwit hplan tag efro m levelsaroun d3-3.5 %a tth evegetativ egrowt hstag e downt o 1-1.5%a tth e final harvest. 58 Effectof timeof nitrogen application Table 6.3. Effect of a single dose of nitrogen applied at different growth stages and soil sterilization on number ofnodule san dnodul e dry weight of 'Diphiri Cream' atvegetativ e (28 DAS), 50%flowerin g (45 DAS) and 50% podding (69DAS )stages . Treatment Vegetative stage 50% flowering 50%poddin g Numberof nodules per plant Control (Tl) 34 142 397 Sterilized soil (T2) 3 56' 1721 N at sowing (T3) 29 170 562 N atvegetativ e stage(T4 ) - 149 479 N at 50% flowering (T5) - - 262 C.V. (%) 63.2 54.7 31.1 S.E. 6.26 31.6 52.1 ^*DQQ5 19 n.s. 154 Noduledry weight (m g plant1) Control (Tl) 144 600 1970 Sterilized soil (T2) 1 2401 7701 N at sowing (T3) 8 600 1870 N atvegetativ e stag = (T4) - 280 1550 N at 50% flowering (T5) - - 1110 C.V. (%) 292 51.4 29.2 S.E. 0.07 0.10 0.19 LSD(0.05) n.s. 300 560 n 15 20 25 n.s.: not significant; n:numbe r of observations. 'There was a2 week s delay inharvestin g thesetreatment s at 50%flowerin g and50 % podding. 59 Chapter 6 (a) (b) Figure 6.1. Effect of nitrogen fertilization atdifferen t growth stagesan d soil sterilization on number ofpod s (a)an dpo d dry weight (b)o f 'Diphiri Cream' atfina l harvest. (Treatments: Tl=control; T2=sterilized soil;T3 =N fertilization at sowing;T4= N fertilization at vegetative;T5= N fertilization at 50%flowering ; T6=N fertilization at 50%podding) . Soil sterilization did not affect the shootN conten t ofbambar agroundnu t plants.However , it hast ob econsidere d thatplant s growni nsterilize d soildevelope d more slowly andreache d the 50%flowerin g and 50%poddin g stageabou t 2week s latertha nplant s withth e other treatments (Table 6.5). Consequently, harvest ofth e T2treate d plantswa s also delayed with two weeks.Nitroge n fertilization significantly affected shootN contents .A tth e vegetative stage,shoo tN content wasincrease d ifplant swer e fertilized withN as sowing (T3).Thoug h at laterharvest s effects ofN fertilization on shootN conten t weremostl y not significant, N fertilization tended to increase shootN content irrespective oftim eo f application. The N contents atth e final harvest given inTabl e 6.5 represent shootN content s and dono t include Ni n thepods .Thi s quantity canb ever yimportan t atth efinal harvest . 60 Effectof timeof nitrogen application Table 6.4. Effect of asingl e dose ofnitroge n applied atdifferen t growth stages and soil sterilization on shoot nitrogen concentration of 'Diphiri Cream' plantsa tvegetativ e (28 DAS), 50%flowerin g (45DAS) ,50 %poddin g (69DAS )an dfina l harvest (128DAS) . Treatment ShootN concentratio n (%) Vegetative 50% 50% Final stage flowering podding harvest Control (Tl) 2.77 1.98 1.64 1.10 Sterilized soil (T2) 3.05 1.201 1.351 1.00 N at sowing (T3) 3.23 2.03 1.91 1.37 N atvegetativ e stage (T4) - 2.55 1.62 1.04 N at 50%flowerin g (T5) - - 1.88 0.88 N at 50%poddin g (T6) - - - 1.08 C.V. (%) 23.0 12.2 18.7 15.5 S.E. 0.31 0.11 0.14 0.08 LIMJ005 n.s 0.32 n.s. 0.22 n 15 20 25 30 n.s.:no t significant; n:numbe ro fobservations . 1Ther ewa sa 2 week sdela y inharvestin gthes etreatment s at 50%flowerin g and 50%podding . 6.4 Discussion Resultso fthi swor kclearl y illustrateth eredundanc y ofa supplementar y fertilization ofbambar a groundnut withminera lnitrogen , atleas tunde rth eprevailin g conditions.Nitroge n application at flowering did not extend the vegetative growth period and not increase seed yield, as hypothesized inth e Introduction ofthi s chapter.Nitroge n applied atsowin g evenreduce d pod formation and seedyield .A lac ko frespons et oN fertilization , asobserved , isa n indication that this leguminous crop can meet itsN requirement byN 2 fixation plus mineral N present in the soil. 61 Chapter 6 Table 6.5. Effect of a single dose of nitrogen applied at different growth stages and soil sterilization on shoot nitrogen content of 'Diphiri Cream' plants at vegetative (28 DAS),50 % flowering (45DAS) ,50 %poddin g (69DAS ) andfinal harves t (128DAS) . Treatment ShootN content (mgN plant1) Vegetative 50% 50% Final stage flowering podding harvest2 Control (Tl) 28 130 298 250 Sterilized soil (T2) 26 1481 214' 278 Na t sowing (T3) 42 142 382 468 N atvegetativ e stage(T4 ) - 194 326 310 N at 50%flowerin g (T5) - - 396 160 Na t 50%poddin g (T6) - - - 256 C.V. (%) 29.6 23.2 22.8 39.9 S.E. 4.2 15.9 32.9 51.2 l^Doos 13.1 n.s. 97 149 n 15 20 25 30 n.s.: not significant; n:numbe r of observations. 1Ther ewa sa 2 week sdela y inharvestin g thesetreatment s at 50%flowerin g and 50%podding . 2Th e amount doesno t includeN inth e seed. Nitrogenfertilization The absence of a positive response of bambara groundnut toN fertilization corresponds with shoot N concentrations as measured (Table 6.4). At all four sampling times, shoot N concentrations of control plants are almost similar to those of N-fertilized plants. Moreover, measured shootN concentrations incontro lplant so f2.77 %N atvegetativ e and 1.64%N atth e 50%poddin g stagear ea nindicatio n ofadequat eN supply (Reuteran dRobinson , 1986;Chapte r 5). Evidently, relatively high nodulation rates of control plants (Table 6.3) seem to be 62 Effectof timeof nitrogen application accompanied by aRhizobia activit y high enought o supply plants with sufficient nitrogen. The contribution of soil mineralN wil lb e discussed later. As hypothesized (Introduction) and observed with other leguminous crops (Elowad and Hall, 1987;Isfa , 1991)lat eapplicatio no fminera lN migh texten dth ephotosyntheti c activeperio dan d improvepo dan d seedyiel do fbambar a groundnut. However, resultso fthi s experiment clearly demonstratetha ta supplementa lapplicatio no fminera lN a t50 %flowerin g stagedi dno tincreas e podan d seedyiel dwit hbambar agroundnut .Jus tcontrary ,afte r Napplicatio n at 50% flowering stage shootdevelopmen twa scompletel ystoppe d (Table 6.2), andth efina l pod and seed yields were similar to those which did notreceiv eextr aN . Whenextr aN wa sapplie d at sowing,po d and seed yieldswer e evenreduced . Whenapplie d at that time plants took up more N (Tables 6.4 and 6.5), and showed a higher final shoot DM production. However, the higher final shoot dry matter production was completely realized duringth epoddin gstag e(Tabl e6.2) .Evidently ,extr aN take nu pprolonge d theperio d of shoot biomassproductio nb ymaintainin ga re-allocatio no fnitroge nan dcarbohydrate s toth e shoota t theexpens eo fpo d andsee dproductio n (Fig.6.1) .Lowe rfina l pod andsee dafte r N fertilization atsowin gar ei nagreemen t with findings ofa previou sexperimen t (Chapter 5).However , while in both experiments N fertilization at sowing resulted in lower pod and seed yields, yield reductions inth e two experiments wereeffecte d differently. While inth e previous experiment early maturing was the main reason for reduced yield, in the present experiment it was just extended vegetative growth after Napplicatio n atsowin gtha tle dt oreduce dpo d andsee d yield. There is no real experimental evidence to explain differences in response to N fertilization at sowingbetwee nth etw o experiments.Onl yamount so fN an dP fertilize r applied were different inth etw o experiments. Nodulationand soil sterilization During the period of vegetative growth, nitrogen nutrition after soil sterilization was adequate despite a severereductio n in root nodulation. Evidently mineralN present in soil, originating from mineralization,wa ssufficien t to meet theplant sN requiremen t duringth e initial phase of growth.Als oa t50 %flowering , whennodulatio nwa sstil lpoor ,plant sseeme dt ocontai n enough Nt omaintai na shoo tdr ymatte rproductio nequa lt otha to fcontro lplants .However ,i fcompare d withth econtrol ,averag e shootN concentratio n attha tmomen twa slowere db y soil sterilization to aleve lto olo wfo r anorma lgenerativ e development, expressed by adela y offlowerin g with about two weeks. Evidently, at the flowering stage soil mineral N was no longer sufficient. However, in this experiment soil sterilization was not completely successful and from the beginning on some nodulation was observed in sterilized soil (Table 6.3). At 50% flowering stage,nodulatio n seemedt ob eto o lowt o supply thebambar a groundnut plantswit h enoughN to maintainadequat einterna lN concentrations . Sometim e later,however , nitrogen nutrition of plants grown in sterilized wasn o longer different from controls. After atemporar y lack of N2 63 Chapter 6 fixation, Rhizobium activity seemst ohav ebee nrecovere d andt ohav ebecom e active enough to supplyth eplant swit henoug hN , evena tlo wnodulation .Thi simplie s anextraordinar y highN 2 fixation capacity ofnodule spresen t in sterilized soil,o rexces snodulatio n incontrols . From this experiment it was not possible to quantify the contribution of soil mineral N to N nutrition ofbambar agroundnu t becauseth e soil sterilization method asuse d seemed tob eno t 100%effective . Resultsa t least indicatetha t during the vegetative phase soil mineral N clearly contributed to the nitrogen nutrition ofbambar a groundnut, but soon afterwards it became of minor importance,probabl y duet o depletion. Finally,fertilizatio n ofbambar agroundnu twit hnitroge nseeme dt ob eno tbeneficial , irrespective oftim e of application. IfN fertilization affected pod and seed yield, effects were negative. Soilminera lN seem st ocontribut et oth eN nutritio no fbambar agroundnut , but only during the vegetative phase. A further quantification of this contribution has to be done, with special emphasis toth e soil organic matter content ofth e soilsused . 64 Generaldiscussion Chapter7 General discussion Inthi s chapter,th eresult so f a study onminera l nutrition ofbambar agroundnu t are discussed inrelatio nt oth e generalobjective s aspresente d inFigur e 1.1 (Chapter 1). Factors such assoi l P content, soil moisture content, time and rate of P fertilization, and root development are discussed in relation to P nutrition of bambara groundnut and its possible role on other plant processes such asnitroge n fixation. TheN 2 fixation processpe r se isno t investigated, but the effects ofminera lfertilize r Po nnodulatio n(numbe ro fnodule san dnodul edr yweight )an dshoo t N accumulationar econsidere da srepresentin gth erol eo fP i nnitroge n fixation. Also discussed isth enee dfo r extraminera lN an dit seffect s onnitroge nfixation . Thechapte r isconclude d with the implications ofth eresult sfo r bambara groundnutfarmer s in Botswana. Phosphorus nutrition Like other crops, bambara groundnut requires P for growth and development. Under low P conditions, shootP concentration s arekep t at orbelo wth ecritica l shoot Pleve l and dry matter production islimited .Th eobserve d response of shootdr ymatte r production of potted bambara groundnut toP fertilizatio n inthi s study was duet o acombinatio n of low Pconten t ofth e soil and limited soil volume (Chapter 2). Although bambara groundnut did not respond to P fertilization inth efiel d ata soilP conten t of6. 2 mgP kg" 1(P-Bray )ther ewa sa n increase inP uptakewit h increasei nsoi l moisture availability irrespective ofP fertilization . Under improved soil moisture conditions, soilP availabilit y wasincrease d but itwa sno t sufficient for optimal growtha sshow nb ya shoo tP concentratio n whichwa saroun dth ecritica l level.Thi smean stha t alsoafte r irrigationsoi lP availabilit y limitedplan tgrowt han dsee dyield .Therefore , seedyield s of more than 4.2 t ha"1 in Botswana soils can be expected only if the soil moisture content is adequate and soilP i s increased. Oneo fth eexplanation s for the lacko frespons e to Pfertilizatio n bybambar a groundnut inou r field experiment with a soil apparently low in P,wa stha t itha s alo w Prequiremen t (Chapter 3.1).Abou t 80%o fth eP requiremen t ofbambar agroundnu t couldb eme tb yP originatin g from the soil.However , the internal Pus e efficiency was low.Probabl y the low Prequiremen t may explain whybambar a groundnut has beenreporte d to growwel l inlo w fertility soils (National Research Council, 1979, Wassermann et ah, 1984). In potted bambara groundnut plants the positive response only occurswhe n P is applied (and well mixed withth e soil) ata ver y early stage of seedling development (Chapter 3.2). Phosphorus application beyond two weeks after sowing did not improve shoot growth and seed yield. The importance of adequate levels of available Pi nsoil sdurin gth e seedling stageha sbee nreporte d alsofo rmaiz e (Arnon, 1975)an d tomato(Jone san dWarren , 1954).Th efac ttha tther ei sa critica ltim efo rP i nbambar agroundnu t 65 Chapter 7 canals opartl y explainth elac ko frespons et oP fertilizatio n inou r field experiment (Chapter 2) where fertilizer Pwa s broadcast and worked intoth e soil with adiggin g fork. This apparently didno tbrin gi tclos eenoug ht oth eseedlin ga tth ecritica ltime .Furthermore ,bambar agroundnu t wasshow nt ohav ea relativel y smallroo t system,tw otime s lowertha ntha to fmaiz ean dpigeo n pea(Chapte r3.1) .A root-shoo trati oo f0. 2a sfoun d inthi sstud y iscomparabl e totha t reported for bambaragroundnu t byCollinso n etal., (1996).Th epoo r root system ofbambar a groundnut mayals ohav elimite dth eP availabilit yt oth e seedlinga tth ecritica l timean dhav e contributed to the poor P response under field conditions. A poor root system and a critical time for P application inbambar a groundnut meanstha t for positive results,fertilize r P should be placed asclos ea spossibl e toth e seed at sowing. The critical shoot P level (0.15% P) as determined in a sand culture experiment (Chapter 4) showed that bambara groundnut plants in our experiments were grown under P limiting conditions irrespectiveo fth etreatment , Pfertilizatio n andirrigatio n (Chapter2 and3) . A seed yieldo f 4.21ha" 1from ou rfiel d experimentwit ha shoo tP concentratio naroun dth ecritica lleve l is an indication of the potential of bambara groundnut to produce well under P limiting conditions ifth esoi lmoistur econten t issufficient . This isals o in linewit h literature that it can produceconsiderabl eyield si nlo wfertilit y soils(Nationa l ResearchCouncil , 1979;Wasserman n etal., 1984).Bambar a groundnut farmers inBotswan ad ono t apply fertilizer ormanur e toth e crop (Brink et al., 1996), and the shoot P concentrations of bambara groundnut plants in the farmers' fields as found in this study were around or below the critical P level (Chapter 4). Farmers attribute low yields to poor rainfall (Brink etal,. 1996) and in this study it has been shown that soil moisture content is important for P uptake. Probably the P status of bambara groundnut plantsfro m farmers field waslo wbecaus eo fa combinatio n oflo wsoi l Pconten t and lowsoi lmoistur e content.Th eresult s showtha tth eshoo tP concentration , inadditio nt oth e soil analysis, can be a useful guide for monitoring the P status and P requirement of bambara groundnut. In other leguminous crops Pfertilizatio n increased shootN concentration s which isthough t to haveresulte dfrom improve d N2 fixation (Kang andNangju , 1983; Israel, 1987;Othma n et al., 1991).Thi s effect wasno t found inou r experiments when shoot Pconcentration s were around or below the critical P level. It might mean that, thanks to an efficient and possibly low P demandingN 2fixin g mechanism,bambar agroundnu tplant so nP-poo r soilsi nBotswan aar eno t suffering from N deficiency. This view is supported byth ehig h shootN concentrations found infarmers ' fields (Chapter4) . Inou rexperiment s Pnutritio n didno t affect thedevelopmen t of bambara groundnut plants,e.g . the time it takes from sowing to flowering, podding and seed maturity. As long as internal P concentrations were around orbelo wth e critical level addition ofP resulte d in increased shoot drymatte rbu tdi dno taffec t thenumbe ro fpods .Thi s improvedpo d filling and increased seed weight. With P shortage yield is thus not sink-limited but mainly source-limited. Lack of soil moistureha sbee nreporte d todecreas epo dnumber s(Ameya wan dDoku , 1983;Collinso n etal., 1996;Chapte r 2). Whenthi s occurs inP-poo r soils itha s intw o ways anegativ e effect on seed 66 Generaldiscussion yield,a direc ton eo npo dnumber san da nindirec ton eo npo dfilling throug hreduce d availability of soilP . Nitrogen nutrition In this study bambara groundnut did not respond to nitrogen fertilization (Chapters 5 and 6), making it obviously different from other leguminous crops that respond positively to N fertilization (Alios andBartholomew , 1955;Hartfiel d etal, 1974;Minchi n etal., 1981; Elowad and Hall, 1987;Isfa , 1991).Nitroge n fertilizer applied to bambara groundnut at sowing even reducedpo d formation andsee dyield . Application ofminera lN fertilizer atrate s more than 20 kg N ha"1 has been reported to suppress nodulation in cowpea and other leguminous crops (Richardson et al., 1957;Alio s and Bartholomew, 1955;Graha m and Scott, 1984). Contrary, nodulation (number ofnodule san dnodul edr yweight ) inbambar a groundnut continues even at mineralN fertilize r rateso fu pt o the equivalent of 100k gN ha" 1.However , application ofN at flowering (49DAS' )seeme dt ob eth eonl ytim ewhe nN fertilizatio n ofbambar a groundnut had negative effects on nodulation (number of nodules and nodule dry weight) (chapter 6). This seems to suggest that the flowering stage isa critica l time for nodulation which isi n line with the observation of Dakora et al., (1992) that the negative effect of applied N fertilizer on nodulation and N2 fixation of bambara groundnut was severe between 43 and 56 DAS. This shows that application of mineral N can have negative effects on N nutrition of bambara groundnut. Ourexperimenta l resultsindicat etha tth e N2fixin g mechanism ofbambar agroundnu t plays an important role in meeting the N requirement of thecrop .Th e results in chapter 6 showed that poor nodulation, e.g. in the treatment with sterilized soil, resulted in delay in flowering and podding andthi sunderline sth e importance ofhavin geffectiv e Rhizobiai nth e soil. Fortunately bambaragroundnu t seemst onodulat efreel y withindigenou srhizobi ai nBotswan asoil sa s found inou rstud y(Chapte r2) .Guey ean dBordelea u (1988)i nSenega lals ofoun d thatbot hindigenou s and introduced Rhizobiumstrain s nodulated all 24 bambara groundnut genotypes tested. The ability to nodulate with the indigenous Rhizobiaan dth e capacity to meetN requiremen t from N2 fixation and soil N indeed makes bambara groundnut an important crop for resource poor farmers. Implicationsof the study Water isa scarc e resource inBotswan a andunde r theprevailin g conditions itappear s to beth e most important factor inth e production ofbambar a groundnut. Soil moisture shortage restricts P availability to bambara groundnut plants and decreases pod production. Under Botswana conditions,wit h its short growing season, iti simportan t to sowbambar a groundnut as soon as theearl y rainsar e steady.B ydoin gthis ,plant sca nbenefi t directly after planting from increased P availability which is a prerequisite for realizing at least part of its potential yield. To avoid moisture stress at a later stage, wide spacing of bambara groundnut plants is a good strategy 67 Chapter 7 which many farmers seem to adopt. It ishoweve r cleartha t for reliable high yields,firs t of all adequate soilmoistur e isneede d followed by optimal plant densities and timely application of Pfertilizer . Inou rfiel d experiment, irrigated fields witha plan t population of 5400 0plant sha' 1 produced ayiel d of about 4.2 tonnes seed ha"1.Assumin g that not only the pods butth e entire plants are harvested, thenth etota l amount of Premove d from thefield i so fth e order of2 5k g ha"'. Due to the fact that the experimental plants were grown on a soil cleared from a well developed fallow vegetation this quantity ofP coul d betake n up from the soil.Howeve r when bambara groundnut is grown on more or less permanently cropped soil additional P will be needed of the same order of magnitude as given above. Tentatively it can be concluded that Botswana farmers growing rainfed bambara groundnut cannotb e advised to use P fertilizers. If howeverthe yca nirrigat ethei rcro pan dplan ta thig hdensity ,P fertilizatio n ata rat eo fabou t 30 kg P ha"1 can be recommended whereby P fertilizers should be placed in the soil close to the seeds. Determining shoot P concentrations at the early podding stage might be a suitable diagnostic means to monitor the P nutrition of the crop and to refine this rather general P fertilizer recommendation. As to nitrogen, our experimental results suggest that bambara groundnut can meet its requirements by N2 fixation and supplementary N uptake from the soil, implying that N fertilization isno t needed. 68 References References Ae, N., Arihara, J., Okada, K., Yoshika, T. and Johansen, C, 1990. Phosphorus uptake by pigeon pea and itsrol e incroppin g systemso fth e Indian Subcontinent. Science, 248:477-480 . Aguilar, A. and Van Diest, A., 1981. Rock-phosphate mobilization induced by the alkaline uptakepatter no flegume sutilizin g symbiotically fixed nitrogen. Plant and Soil, 61: 27- 42. Alios, H.F. and Bartholomew, W.V., 1955. Replacement of symbiotic fixation by available nitrogen. Soil Science, 87:61-66 . Ameyaw,C.E.G . and Doku,E.V. , 1983. The effects of soilmoistur e stress onth e reproductive efficiency and yield ofth e bambaragroundnu t (Voandzeia subterranea). Tropical Grain Legume Bulletin, 28: 23-29. Anon., 1962.Othe r legumes. In:Annua l Report ofth e Department ofAgricultur e for the year 1960/61 (Part 2). Nyasaland Protectorate, Government Printer, Zomba, Nyasaland, pp.47-49 . Anon., 1970. Other legumes. In: Annual Report ofth eDepartmen t of Agriculture (Part 2) for the year 1964/65.Malaw i Government, Government Printer, Zomba, Malawi,p .30 . Anon., 1971.Othe rlegumes .In :Annua lRepor to fth eDepartmen t ofAgricultur e (Part2 )fo r the year 1966/67.Malaw i Government, Government Printer, Zomba, Malawi,p .38 . Anon., 1973.Groun d beans. In: Annual Report of the Department of Agriculture (Part 2) for the year 1969/70.Malaw i Government, Government Printer, Zomba, Malawi,p . 122. Anon., 1975.Groun dbeans .In :Annua lRepor to fth eAgricultura l Research Council of Malawi 1974/75. Thondwe,Malawi ,pp .33-34 . Anon., 1979. Jugo beans. In: Annual Report of the Agricultural Research Division 1977/78. Swaziland Government, Malkerns, Swaziland, pp. 107-112. Anon., 1980.Mino r legumes.In :Annua lRepor t ofth eDepartmen t ofAgricultura l Research for theyea r 1974/75.Malaw i Government, Ministry ofAgricultur e andNatura l Resources, pp. 112-113. Anon., 1989.Bambar a groundnut. In: Status Report onAgronom y Research in the Wetlands. Adaptive Research Planning Team. Mongu, Western Province,Zambia ,p .x . Anon., 1997.Evaluatin g thepotentia l for bambara groundnut asfoo d crop in semi-arid Africa. An Approach for Assessing the Yield Potential and Ecological Requirements of an Under-Utilised Crop.Europea n UnionLif e Sciences and Technologies for Developing Countries, STD3Fina l Report, University ofNottingham , U.K. Anon., 1998. Meteorological data collected from the Department of Meteorological Services, Ministry of Worksan d Communications, Gaborone, Botswana. Arnon, I., 1975.Minera l Nutrition of Maize,Internationa l Potash Institute,Berne , Switzerland. 69 Beynon, S., 1991. The response of sorghum to phosphate fertilizer in Botswana. The Bulletin ofAgricultura l Research inBotswana , (9)61069 . Bhangoo,M.S .an d Albritton, D.J., 1976.Nodulatin g and non-nodulating lee soybean isolines response to applied nitrogen. Agronomy Journal, 68: 642-645. Brink, M., Collinson, S.T. and Wigglesworth, D.J., 1996.Bambar aGroundnu t(Vigna subterranea) cultivation inBotswana . Report offarmers ' survey. University of Nottingham, U.K. and Wageningen Agricultural University, TheNetherlands , p.52 . Brink, M., 1998. Matching crops and environments: quantifying photothermal influences on reproductive development inbambar a groundnut {Vignasubterranea (L.) Verde). PhDThesis ,Wageninge n Agricultural University, TheNetherlands , p. 161. Brooks, C.B.,Dadson , R.B. and Green, B.M., 1988.Evaluatio n of symbiotic effectiveness of elite and wild strains of Bradyrhizobium on cultivars of Voandzeia subterranea (L.) Thours. Tropical Agriculture, 65(1):61-63 . Collinson, ST., Azam-Ali, S.N., Chavula, K.M.an d Hodson, D.A., 1996.Growth , development and yield of bambara groundnut (Vignasubterranea) i n response to soil moisture.Journa l of Agricultural Science, 126:307-318 . Coudert, M.J., 1982.Ni€b6 etvoandzou; uneperspective pour led&veloppement du commerce regionalen Afrique deVouest. FAOmicrofich e XF8332635 .Internationa l TradeCentr e UNCTAD/GATT, Geneva. Cumberland, J., 1978. Jugo beans. Research Report No. 19.Agricultura l Research Division, Ministry ofAgriculture . Malkerns Research Station, Swaziland,p. l1 . Dadson,R.B .an dBrooks ,C.B. , 1989.Response so fbambar agroundnu t (Voanzeiasubterranea (L.)Thours )t oapplie dnitroge n in SouthernTogo .Tropica l Agriculture, 66(2): 169-175. Dakora, F.D.,Atkins , C.A. and Pate,J.S. , 1992.Effec t ofN0 3 onN 2 fixation and nitrogenous solutes of xylem in two nodulated West African geocarpic legumes, Kerting's bean (Macrotylomageocarpum L. )an dbambar agroundnu t( Vignasubterranea (L.) .Plan t and Soil, 140: 255-262. Dennis,E.A. , 1977. Nodulation andnitroge nfixatio n inlegume si nGhana . In:Ayanaba , A.an d Dart, P.J.(eds): Biological Nitrogen Fixation in Farming Systems of the Tropics. John Wiley and Sons,Chichester , U.K., pp.217-232 . Elowad,H.O .an dHall ,A.E. , 1987. Influences ofearl yan dlat enitroge nfertilizatio n onyiel dan d nitrogen fixation of cowpea under well-watered and dry field conditions. Field Crops Research, 15(3/4):229-244 . Epstein, E., 1972. Mineral Nutrition of Plants: Principles and Perspectives. John Wiley and Sons,Inc .Ne w York. Goli, A.E. and Ng,N.Q. , 1987. Bambara groundnut multi-location yield trial. In: Genetic Resources Unit. Annual Report 1987,IITA ,Ibadan ,Nigeria , pp. 11-15. Graham, R.A. and Scott, T.W., 1984.Respons e of cowpea (Vignaunguiculata (L.) Walp.)t o nitrogen and inoculation inTrinidad . Tropical Agriculture, 61(1):56-58 . Gueye, M. and Bordeleau, L.M., 1988.Nitroge n fixation in bambara groundnut, Voandzeia subterranea(L. ) Thouars.MIRCE N J., 4:365-375 . Gueye,M. , 1992.Effec t ofRhizobium andRhizobium/Glomus inoculation onnitroge n fixation inbambar agroundnut .In :Mulungoy ,K. ,Gueye ,M .an dSpencer ,D .C.(eds) : Biological 70 References Nitrogen Fixation and Sustainability of Tropical Agriculture. John Wiley and Sons, Chichester, U.K., pp.283-287 . Harper, J.E., 1974.Soi l and symbiotic nitrogen requirements for optimum soybean production. Crop Science, 14:255-260 . Harris,D .an dAzam-Ali ,S.N. , 1993.Implication so fdaylengt hsensitivit y inbambar agroundnu t (Vigna subterranea) for production in Botswana. Journal ofAgricultura l Science, 120: 75-78. Hartfield, J.L., Egli, D.B.,Leggett , J.E. and Peaslee,D.E. , 1974.Effec t of applied nitrogen on the nodulation and early growth of soybeans {Glycinemax (L.) Merr.). Agronomy Journal, 66: 112-114. Ingestad,T. , 1980.Growth ,nutrition ,an dnitroge nfixatio n ingre yalde ra tvarie drat eo fnitroge n addition. Physiologia Plantarum, 50:353-364 . Ingestad, T. and Lund, A.B., 1986. Review Article. Theory and techniques for steady state mineral nutrition and growth of plants. Scandinavian Journal of Forestry Research, 1:439-453. Isfa, D., 1991. Fertilizer nitrogen uptake by soybean asrelate d to cultivars and time of application using 15Ntechnique .Journa l of PlantNutrition , 14(12) : 1369-1380. Israel, D.W., 1987. Investigation of the role of phosphorus in symbiotic dinitrogen fixation. Plant Physiology, 84:835-840 . Janssen, B.H., 1998. Efficient use of nutrients: an art of balancing. Field Crops Research, 56: 197-201. Jones, L.G. and Warren, G.F., 1954.Th eefficienc y ofvariou smethod s of application of P for tomatoes.Proc . Amer. Soc. Hort. Sci., 63: 309-319. Joshua, W.D., 1991.Physica l properties, moisture regime and crop production: A case study on a Luvisol in Sebele. FAO/UNDP/Botswana Government Project BOT/85/011. Final Document No. 36.Gaboron e Botswana. Kamstra,A.S. , 1995.Phosphoru s inbambar agroundnut . Effects ofconcentratio n and period of Psupply .Ms cThesis ,Department s ofAgronom y and Soil Science and Plant Nutrition, Wageningen Agricultural University, TheNetherlands , p.44 . Kang, B.T. and Nangju, D., 1983. Phosphorus response of cowpea. Tropical Grain Legume Bulletin, 27: 11-16. Krigel,I. , 1967.Th eearl yrequiremen t fornutrient sb ysubterranea n clover seedlings {Trifolium subterraneum). Aust. Journal ofAgricultura l Research, 18:879-886 . Linnemann, A.R., 1986. Bambara groundnut {Vignasubterranea (L. ) Verde): an annotated bibliography. Tropical Crops Communication No. 11. Wageningen Agricultural University. Department ofTropica l CropScience ,Wageningen , TheNetherlands , p.39 . Linnemann, A.R. and Azam-Ali, S.N., 1993. Bambara groundnut {Vignasubterranea). In: Williams,J.T .(Ed.) : Pulsesan dvegetables .Chapma nan dHall ,London ,U.K. ,pp . 13-58. Linnemann, A.R., 1994.Photo-therma l regulation ofphenologica l development and growth in bambara groundnut {Vignasubterranea (L.)Verde) . PhDThesis ,Departmen t of Agronomy, Wageningen Agricultural University, TheNetherlands , p.123 . 71 Lungu, D.M. and Mbewe, D.M.N., 1986. Characterisation of yield and nutritional quality of bambaragroundnu t{Voanzeia subterranea). ProgressRepor tSACCAR ,Researc hGran t SGR-03. University ofZambia , Lusaka,Zambia ,p . 12. Mafongoya, P.L., 1988.Symbioti cnitroge n fixation ofbambar a groundnut {Vignasubterranea L. Verde). MSc Thesis, Department of Biochemistry and Biological Sciences, Wyne College,Universit y of London. Marschner, H., 1997.Minera lNutritio n inHighe r Plants.Institut e of Plant Nutrition, University of Hohenheim, Federal Republic of Germany. Academic Press,Harcourt BraceJovanovich Publishers. Mazhani, L.M. and Appa Rao, S., 1985. Groundnut, bambara groundnut and finger-millet germplasm collection in Botswana. Department of Agricultural Research, Ministry of Agriculture. Crops Research BulletinNo . 3.p .6 . Messiaen, CM., 1975. {Voanzeiasubterranea). In: Le potages tropical, Tome 2; Cultures speciales. Presses Universitaires deFrance ,Paris ,France , p.340 . Minchin, F.R., Summerfield, R.J. and Neves, M.C.P., 1981.Nitroge n nutrition of cowpeas {Vigna uguiculata): effects oftimin g of inorganic nitrogen applications on nodulation, plant growth and seed yield. Tropical Agriculture, 58(1): 1-12. Mukurumbira,L.M., 1985.Effect s ofrat eo f fertilizer nitrogen andpreviou s grain legume crop onmaiz e yiels.Zimbabw e Agric.Journal , 82:177-179 . Mulila-Mitti,J.M . and Kanenga,K. , 1997.Toward s improved bambara groundnut production inZambia . In:Proceeding s ofth eInternationa l BambaraGroundnu t Symposium held at the University ofNottingham , U.K., 23-25Jul y 1996,pp . 193-200. Musonda, M.V., 1988.Respons e ofbambar a groundnut {Vignasubterranea) to levelso f nitrogen, phosphorus and potassium, andth e useo fvisua l diagnostic approach to identify mineral deficiencies, BSc.Thesis ,Universit y ofZambia ,Lusaka ,Zambia ,p .34 . National Research Council, 1979.Bambr agroundnut . In:Tropica l Legumes,Resource s for the Future.Nationa l Academy of Sciences,Washingto n D.C.,pp .47-53 . NationalInstituu tvoo rLandbouwstudi e inBelgisch-Cong o(NILCO),1960 .Jaarversla gvoo r net dienstjaar 1958.Clarenc e Denis,Brussels , Belgium, pp.245,250,362 . Nnadi, L.A., Balasubramanian, V. and Singh, L., 1976. Yield and mineral composition of grain legumes. GhanaJourna l ofAgricultura l Science 9(3): 211-215. Nnadi, L.A., Singh, L. and Balasubramanian, V., 1981. Effect of grain legumes and sorghum onth e soil nitrogen status and yield of subsequent maize crop. SamaruJourna l of Agricultural Research, 1(2): 183-190. OthmanWan , W.M.,Lie ,T.A. , 't Mannetje, L.an d Wassink, G.Y., 1991. Low level of phosphorus supply affecting nodulation, N2 fixation and growth of cowpea {Vigna unguiculata L. Walp). Plant and Soil, 135: 67-74. Rachie, K.O.an d Silvestre,P. , 1977.Grai n legumes.In :Leakey , C.L.A.an d Wiilis,J.B . (eds): Food Crops ofth e Lowland Tropics.,Oxfor d University Press,U.K. , pp.41-74 . Ramolemana, G.M., Maphanyane, G.S.an d Wessel, M., 1997.Respons e of bambara groundnut tophosphoru s with andwithou t irrigation. In:Proceeding s ofth e International Bambara Groundnut Symposium held at the University of Nottingham, U.K., 23-25 July 1996,pp .61-70 . 72 References Reisenauer, H.M., 1963. Relative efficiency of seed and soil applied molybdenum fertilizer. Agronomy Journal, 55:459-460 . Reuter, D.J. and Robinson, J.B., 1986.(eds) .Plan tAnalysis :A n Interpretation Manual. Inkata Press,Melbourne , Sydney. Richardson, D.A., Jordan, D.C. and Garrard, E.H., 1957.Th e influence of combined nitrogen onnodulatio n and nitrogen fixation byRhizobium meliloti dangeard. Canadian Journal ofPlan t Science, 37:205-214 . SAS, 1985.Statistical Analysis Systems. User'sGuide. SASInstitute , Inc.,Cary ,NC ,USA . Scott,J.M .an dBlair ,G.J. , 1988. Phosphorussee dcoating sfo rpastur especies .I .Effec t ofsourc e and rate of phosphorus on emergence and early growtho fphalari s (Phalaris aquatica) andlucern e(Medicago sativaL.) .Aust . Journal ofAgricultura l Research, 39: 437-445. Sesay, A., Saboleh, S. and Yarmah,A. , 1997.Farmer s knowledge and cultivation of bambara groundnut in Sierra Leone. In: Proceedings of the International Bambara Groundnut Symposium helda tth eUniversit y ofNottingham , U.K., 23-25 July 1996,pp . 119-132. Sigunga, D.O., 1997. Fertilizer nitrogen use efficiency and nutrient uptake by maize (Zea maysL. ) inKenya . PhDThesis ,Departmen t of Soil Sciencean d Plant Nutrition, Wageningen Agricultural University, TheNetherlands , p.207 . Sims, D. 1981. Agroclimatological information, croprequirement s and agricultural zones for Botswana. Land Utilization Division, Ministry ofAgriculture ,Privat e Bag003 , Gaborone. Smid,A.E . and Bates,T.E. , 1971.Respons e of cornt o small amounts of fertilizer placed with the seed.V . Seed coating compared with banding.Agronom y Journal, 63: 380-384. Smith,P.F. , 1962.Minera l analysis ofplan ttissues .Ann .Rev .Plan t Physiol., 13: 81-108. Somasegaran,P. ,Abaidoo ,R.C .an dKumaga ,F. , 1990.Host-Bradyrhizobium relationships and nitrogen-fixation inth ebambar a groundnut {Voandzeiasubterranea (L.) Thouarsnom . cons.).Tropica l Agriculture, 67:137-142 . Stanton, W.R., Doughty, J., Orraca-Tetteh, R. and Steele,W. , 1966.Grai nLegume s in Africa. FAO,Rome . Stegman, E.C.,Musick , J.T. and Stewart, J.I., 1983.Irrigatio n water management. In: Jensen, M. E. (ed):Desig n and Operation of FarmIrrigatio n Systems. American Society of Agricultural Engineers. 2950Nile sRoad , St.Joseph , Michigan 49085 USA,pp .763-816 . Tanimu, B. and Yayock, J.Y., 1989.Bambar a agronomy. In: Cropping Scheme Report 1989, Legumes and Oilseeds Research Programme. Ahmadu Bello University, Institute of Agricultural Research, Zaira,Nigeria ,pp . 136-140. Tanimu, B.an d Yayock, J.Y., 1990. Bambara agronomy. In: Cropping Scheme Report 1990, Legumes and Oilseeds Research Programme. Ahmadu Bello University, Institute of Agricultural Research, Zaria,Nigeria . Tardieu, M.,1958. Les cultures d'appoint dans la zone d'action du CRA Bambey. [Secondary crops inth e action zoneo fth e Agronomic Research Centre,Bambey] .Bulleti n Agronomique Ministere del aFranc e d'OutreMer , 17: 5-54. 73 Tayo, T.O., 1981. Studies on the effects of foliar spray of nutrients on the performance of cowpeas(Vigna unguiculata (L.)Walp.) .Journa l ofAgricultura l Science, 96:375-388 . Tayo, T.O., 1986.Th e response ofcowpe a (Vignaunguiculata (L.) Wailp.)t o foliar sprays of nitrogendurin gth ereproductiv estag eo fdevelopment .Fiel dCrop sResearch , 14(2): 181- 191. Thompson,E .J .an dDennis ,E.A. , 1977.Studie so nnodulatio n andnitroge nfixation b y selected legumes. In: Proceedings of thejoin t University of Ghana Council for Scientific and Industrial Research Symposium onGrai nLegume s inGhana .Cro p Science Department, University of Ghana,Legon ,Ghana ,pp .85-102 . Walkley, A. and Black, I.A. , 1934.A n examination ofth e Degtjareff method for determining soil organicmatte r andpropose d modification ofth echromi c acid titration method. Soil Science, 37: 29-38. Wassermann, V.D., Kruger, A. J. and Hyens, G., 1984. The response of Bambara groundnut (Voandzeia subterranea) and pigeon pea (Cajanuscajari) to applications of lime, P and K. South Africa Journal ofPlan t and Soil, 1 (1): 4-8. 74 Summary Summary Bambara groundnut (Vignasubterranea (L. ) Verdec.) is a legume crop grown especially by small farmers mainly in semi-arid parts ofAfric a both inmixe d cultivation and pure stands.I t is considered as ahard y crop because ofit sdrough t tolerance,resistanc e to pests and diseases andabilit yt oyiel do nchemicall y poorsoils .Th ecro pproduce sedibl eseed swhic hca nb e eaten unripe or stored asdrie d pulse for later consumption. InBotswan a iti s grownunde r semi-arid conditions by small farmers asa mino rcrop .Attempt shav e beenmad et o increase the bambara groundnut yield through application of nutrients and improvement of the nitrogen-fixing capacity. Thisresearc h hasno t been undertaken systematically and especially the responses to Pan dN fertilizers didno t showa clea rpattern .Give nth elo wavailabl e Plevel s and the limited N supply in Botswana soils,a researc h programme was undertaken to investigate the Pan d N nutrition of bambara groundnut. The main objectives of the study were:(i) to investigate the effects ofapplie d PandN o ngrowt han ddevelopmen t ofbambar agroundnu t inBotswan asoils , and (ii) to investigate the P and N uptake, shoot concentrations at different growth and development stages of bambara groundnut and (iii) how uptake, internal concentration and growth arerelated . Theexperimenta l programme consisted of five pot experiments and alarg e field experiment with Pan dirrigatio na streatment s inwhic h bambara groundnut was grown on arepresentativ e lowP soil .I nadditio na san dcultur eexperimen t wasconducte d to establish the critical P concentration in bambara groundnut shoots. The critical P levels were subsequently usedt o determine the Pstatu s ofbambar a groundnut plants in farmers' fields. InChapte r 2th equestio nwa swhethe rbambar agroundnu twil lrespon d toP fertilizatio n ina P - poorsoi li nth eabsenc ean dpresenc eo fadde dRhizobium inoculum,whethe rbambar a groundnut selections are different with respect to Pnutrition , and whether soil moisture level under field conditions limits Puptake .I na preliminar y pot experiment, which wasterminate d at flowering (51 days after sowing; DAS),th e response of bambara groundnut shoot dry matter (DM) to P fertilization (fertilizer ground andthoroughl y mixed withth e soil)wa spositiv e and linear.Ther e was no effect of P fertilization on nodulation both in the absence and presence of added inoculum. Leaf blade P and N concentrations were not affected by P fertilization. The two selectionsshowe dbasicall yth esam erespons et oP fertilization . Theindigenou srhizobi a seemed to be sufficiently effective in nodulating bambara groundnut. In the field, P fertilization (fertilizer broadcast after ploughing andworke d intoth e soil with adiggin g fork) had no effect onbambar a groundnut plants,whil eirrigatio nincrease d allplan tgrowt hparameter s except root DM. Total seed yield was 2.8 t and 4.2 t ha"1fo r rainfed and irrigated treatments, respectively. Leaf blade P and N concentrations were not affected neither by P fertilization nor by soil moisture content. However, irrigation increased the uptake of P. Evidently, part of the Ppoo l present inth e soil becomesavailabl e under improved soil moisture conditions,probabl y duet o improved Pdiffusio n toth eroots .Th elac ko frespons et oapplie dP i nth efiel d experiment with a lowP soi l mightimpl ytha twhe nroo t growth isno trestricte d by soil volume as isth e case in potexperiments ,availabl e soil Pwa senoug ht omee tth eP requirement s ofbambar a groundnut. 75 This could mean that compared to other crops like cereals bambara groundnut has a low P requirement possibly combined with ahig h P efficiency. Thishypothesi s wasteste d in apo t experiment (reported in Chapter 3.1) inwhic hth e Puptak e and the internal Pus e efficiency of bambara groundnut were compared to those of maize and pigeonpea .Maiz ei sa cro pwit ha hig hP requiremen t andpigeo npe aha sth ecapacit yt ous esoi l Pnormall y notavailabl et oothe rcrops .Th erespons eo fshoo tdr ymatte rt o added Pwa shighes t formaize ,intermediat efo rbambar agroundnu t andlowes tfo rpigeo npea .Compare dt oth eothe r two crops bambara groundnut had aver y lowroot-shoo t ratio.Th e Puptak e of maizewa stw o andfive time shighe rtha ntha to fbambar a groundnut andpigeo npea , respectively. Contrary to maize, most of the Prequiremen t ofth eothe r two crops could beme t with Poriginatin g from soil. Maize had the highest internal Pus e efficiency, onaverag eproducin g 1.89g DMpe rm g P taken up,agains t values of 0.79 for bambara groundnut and 0.56 for pigeon pea. Therefore, lack of response by bambara groundnut to Pfertilizatio n inth efield canno t be attributed to an exceptionally high Pefficiency , butprobabl y to itslo wP requirement. Anotherreaso nfo rth edifferenc e inrespons eo fbambar agroundnu tt oapplie dP a sfoun d inpot - and field experiment could be the mode of fertilizer application. In the pot experiment, P fertilizer wasthoroughl y mixedwit hth esoi lbefor e plantingwhil ei nth efield experimen t Pwa s broadcastan dworke dint oth etop-soi lwit ha diggin g fork. Thismean stha ti npot sth eP fertilizer is directly after germination available to the plant and in the field at a later stage or not at all duringth egrowin gperiod .Thi side ale dt oa po t experiment described in Chapter 3.2, in which different timing of P application was studied. The effects of application at sowing,tw o weeks andfou r weeksafte r sowingwer e studied. Itwa s found that large significant DMresponse s are onlyfoun d whenP i sapplie dwithi ntw oweek safte r sowingan dtha tthi searl yapplicatio n isals o needed for ahig h seedyield .Thi sindicate stha tunde rfield condition s broadcasting of fertilizer atsowin gi sno t aneffectiv e wayt o supply bambaragroundnu t plants intim e (e.g.directl y after germination) with additional P. Inth efield an dpo t experiments (Chapters 2an d 3),P fertilizatio n had noeffec t onth e shootP concentration of bambara groundnut plants irrespective of the treatment (P rate, time of P application andirrigation) .Thi sle dt oth equestio nwhethe rinterna l shootP concentration s found in bambara groundnut plants in previous experiments were adequate or not. To answer this question, the critical shoot P concentration of bambara groundnut was determined in a sand culture experiment using 12P rate s varying between 0.21 and 159m g Pplant" 1fo r the whole experimental period. The critical P value, determined graphically and associated with 10% reduction inmaximu m shootdr yweight ,wa sabou t0.15 %wit ha critica lrang eo f0.1 5t o0.20% . Thecritica lP valu eshowe dtha tbambar agroundnu tplant si nou rpreviou sexperiment s (potsan d field) were growing under suboptimal Plevels . Subsequently afar m survey wasconducte d on ten farms to assessth e soil P status ofbambar a groundnut fields in Botswana under farmers' conditions, andt o compare the Pnutritiona l status ofthos e field-grown plantswit hth e critical shootP concentratio n determined from oursan dcultur eexperimen t (Chapter4.2) . Soilan dplan t samplesfrom eac ho fth ete n farms selected for the survey were collected atabou t 78DA S and analysed for P.Th e soil Plevel s from 80%o f the farms werebelo wfive milligram s Pkg" '(P - 76 Summary Bray). On two farms, shoot P concentrations were atth e critical level and on the other farms belowth ecritica l level. In all previous experiments, bambara groundnut plants were growing under marginal P conditions with shoot Pconcentration s around or belowth e critical level, irrespective ofth e P fertilization and irrigation treatments, while the N nutrition was adequate. This raises the questionwhethe rth eplan tN suppl ywil lstil lb eadequat e ifbambar agroundnu tplant sar egrow n under conditionswher eP wil lb en olonge rmargina l ordeficient . Undersuc hconditions , higher DM production will not only increase the plantN requirement, but also the N2 fixation might behave differently. To answer these questions, a pot experiment described in Chapter 5 was conducted tofind ou twhethe r (i)th ecapacit y ofth eN 2fixatio n process,possibl y supplemented with somenativ esoi lminera lN i shig h enought omee tth eplan tN requiremen t alsounde r non- limitingP conditions ,an d(ii )a tinterna lP concentration sexceedin gth ecritica llevel ,th eamoun t of fixed N increases with increasing internal P. Inth e experiment, bambara groundnut plants were fertilized with different rates of P with or without additional mineral N. Phosphorus fertilization increased shoot, root, and nodule dry weights, seed yield, shoot P and shoot N concentrations and shoot Pan d shootN content s atal l sampling periods.Nitroge n fertilization on the other hand had no effect on the parameters measured but influenced the time to plant maturity. Under non-limiting P conditions bambara groundnut seems to be able to meet itsN requirement from N2 fixation and soilminera lN . Theresult s indicate that no mineral fertilizer N should be supplied at sowing, but this does not rule out the possibility that at a later stage additionalN ma yb eneeded .On eo fthos emigh t beth e onset ofpodding ,whe n amajo r shift in assimilate partitioning from vegetative growtht opo d filling takesplace .T o investigate this,th e effect ofN fertilizatio n atdifferen t development stages ofbambar a groundnut: at sowing (as in thepreviou sexperiment) ,a tth eearl yvegetativ e growth,a t(50% )flowerin g and(50% )podding , was investigated in apo t experiment (Chapter 6).T o separateth e contributions of soil mineral Nan dN 2fixation t oth eN nutritio n ofbambar a groundnutplants ,treatment s with sterilized and unsterilized soil (control) were incorporated in the experiment. Nitrogen fertilization did not affect shootdr yweigh tan dsee d yield. OnlyN fertilization at 50%flowerin g decreased nodule number andnodul edr yweight , shootN concentratio n and shootN content. This canmea n that the flowering stage iscritica l for nodule formation inbambar a groundnut. Sterilization of soil decreased nodule number and nodule dry weight. Shoot N concentration of plants grown in sterilized soilwa sdecrease d onlya tth e50 % flowering stage,an dpoo rN nutritio na tearl y stages ofgrowt hdelaye dth eflowerin g andpoddin g stagesb ytw oweeks .Shoo tN concentratio n ofth e sterilized soil treatment recovered despite the low nodule number and nodule dry weight, implying a high N2 fixation efficiency of nodules present. It was not possible to quantify the contribution ofth esoi lminera lN t oth eN nutritio no fbambar agroundnu t becausenodule swer e present inth e sterilized soiltreatmen tthroughou tth eexperimenta l period.Nitroge n fertilization, irrespective of time of application or growth stage was not beneficial to plant growth and reproduction. Finally in Chapter 7 a general discussion of the main results on the responses of bambara groundnut to Pan dN fertilizatio n inBotswan asoil san dth e implications tobambar a groundnut farmers is presented. It is concluded that P is important for growth and seed yield of bambara 77 groundnut,an dth elo wP requiremen t mayb eresponsibl e for itsabilit yt othriv e in chemically poorsoils .Bu tfo r apositiv erespons e toP fertilizatio n to occur,th e soil moisture content must beadequat ean dth e fertilizer shouldb eavailabl et oth eseedlin gwithi ntw o weeks after sowing. Bambaragroundnu t canmee tit snitroge nrequiremen t from N2 fixation and soil mineral N,an d therei sn onee dfo r supplementary mineralN fertilizer. The shootP concentrations atth e early podding stage seem to be a suitable guide for monitoring the P status and P requirement of bambara groundnut. 78 Samenvatting Samenvatting Bambaraaardnoo t(Vigna subterranea (L. )Verdec. )i see nvlinderbloemi g gewasda tveela ldoo r kleine boeren in semi-aride gebieden vanAfrik a wordt geteeld, zoweli n gemengde teelten als inmonocultures .Toleranti e tegen droogte,resistenti e tegen ziekten enplage n enhe tvermoge n ombi j lagechemisch ebodemvruchtbaarhei dno g eenoogs tt egeve nmake nhe tgewa s geschikt voor teelt onder marginale omstandigheden. Het gewas produceert eetbare zaden die onrijp kunnenworde n geconsumeerd,maa rwaarva nd epeule nn adroge noo kkunne nworde nbewaar d voorlater econsumptie .I nBotswan a wordt het gewasveela l doorklein eboere n geteeld als een bijgewas.Poginge nzij nondernome no md eopbengs tva nd ebambar aaardnoo tt everhoge ndoo r hetgebrui kva nmeststoffe n enverbeterin gva nhe t stikstofbindend vermogenva nhe tgewas .Di t onderzoek is echter niet systematisch uitgevoerd en vooral de respons op een toediening van stikstof en fosfaat meststoffen gaf geeneenduidi g beeld. Metal suitgangspunte n eenlag efosfaa t beschikbaarheid enee n laagstiksto f naleverend vermogenva nbodem si nBotswan a isee nonderzoe kgestar tme tal sdoe ld eP e nN voedin gva n debambar a aardnoot nader tebestuderen . Debelangrijkst e doelstelling van dit onderzoek was hetbestudere n van: (i)d eeffecte n vanee n bemesting met Pe nN o p degroei , ontwikkeling en productie vand ebambar aaardnoo t inBotswana , (ii)he t verloop van deopnam eaa n Pe nN en degehalte naa nd ebeid enutriente n ind e spruit tijdens deteelt , en (iii) de samenhangtusse n de opname en gehalten aan debeid e nutrienten in de spruit enerzijds en de groei van de bambara aardnoot anderzijds. Het onderzoek omvatte een vijftal potproeven evenals een uitgebreid veldexperiment met fosfaatbemesting enirrigati eal sbehandelingen ,waarbi j deteel tplaatsvon do pee nP-arm ebode m representatiefvoo rhe tmerendee lva nd ebodem si nBotswana .Bovendie nwer dee nzand-cultur e experiment uitgevoerd omhe tkritisc hP gehalt evoo rbambar aaardnoo tt ebepalen .Di tkritisch e Pgehalt ewer dvervolgen s gebruikto md eP toestan dva nbambar aaardnoo ti nvelde nva nboere n vastt e stellen. Debelangrijkst evrage ni nHoofdstu k 2zij no fd ebambar aaardnoo treageer to pee nP -bemestin g bij een teelt op een P-arme grond eno pee n aanvullende enting van debode mme tee n externe Rhizobium stam.Verde r isnagegaa n ofverschillend e bambara aardnoot cultivars verschillend reagereno pP-bemestin g end eP voedin gva nhe tgewa sword tbeinvloe d door de vochttoestand van de bodem. In een orienterende potproef diewer d beeindigd bij het begin van de bloei (45 dagen na zaaien),blee k de droge stof (DS) van de spruit positief en lineair te reageren op P- bemesting (deP meststo fwa sgemale ne nintensie fme td egron d gemengd).E rwer d geen effect 79 van P-bemesting gevonden op het aantal wortelknolletjes, ook niet na een enting met extern Rhizobium. Bemestingme tP ha dgee neffec t opd egehalte n aanP e nN i nd ebladeren . Detwe e cultivars vertoonden eenzelfde responso p P-bemesting. Dereed s ind e bodem aanwezige rhizobia activiteit bleek voldoende te zijn voor een adequate nodulatie van het bambaraaardnoo t gewas. Inhe tvel d hadee nP-bemestin g (breedwerpigetoedienin gn aploege nva n Pmeststo fkorrei sdi e vervolgens metee nvor k werden ingewerkt) geeneffec t opd egroe iva nd ebambar a aardnoot, terwijl irrigatieall eplan tparameters ,me tuitzonderin g van deworte l DS,positie f beinvloedde. Detotal ezaadopbrengste n indi tveldexperimen t warenrespectievelij k 4.2 en2. 8ton/h ame t en zonder irrigatie. De gehalten aan P en N in de bladeren werden niet beinvloed noch door P- bemesting,noc hdoo rirrigatie .Echte ral sgevol g van irrigatie waswe ld etotal eP opnam edoo r hetgewa sverhoogd .Kennelij k komtbi jee nverbeterd evochttoestan dva nd ebode mee ndee lva n de P voorraad in de bodem beschikbaar voor de plant, waarschijnlijk als gevolg van een verhoogde diffusie van Pnaa rd ewortel .He t ontbreken van eenrespon s opP-bemestin g inhe t veldexperiment met een P arme bodem duidt wellicht op betere omstandigheden voor bewortelingi nhet vel d dani npotproeven . Alsgevol g vanee ngrote rbewortel d bodemvolume in het veld is de totale hoeveelheid beschikbare P voor het gewas onder de gegeven omstandigheden kennelijk voldoende en reageert het gewas dientengevolge niet op een P- bemesting. Dit kan er op duiden dat in vergelijk met andere gewassen (bijvoorbeeld granen) de bambara aardnoot een lageP behoeft e heeft, mogelijk incombinati e met eenhog e Pbenutting . Deze hypothese werd door middel van een potproef getoetst in Hoofdstuk 3.1 waarbij de P opname en interne Pbenuttin g van debambar a aardnoot werd vergeleken met dieva n maise n duivenboon. Mais isee ngewa sbeken d omz' nhog eP behoefte , terwijl van duivenboon bekend isda the tgebrui k kan makenva n bepaalde Pvoorrade n ind ebode m die voor andere gewassen nietbeschikbaa rzijn .Mai svertoond ed egrootst erespon so pP-bemesting ,gevolg d doorbambar a aardnoote nduivenboon .I nvergelij k metd etwe eander esoorte nvertoond ed ebambar aaardnoo t een zeer lage wortel-spruit verhouding (op DS basis). De P opname bij mais was twee, respectievelijk vijfmaa lz ohoo gal sdi eva nd ebambar aaardnoo te nduivenboon .I ntegenstellin g totwa tbi j maishe tgeva lwas ,blee kbi j detwe eander e gewassend evooraa d aanP i n de bodem toereikend. Mais vertoonde dehoogst e interne Pbenutting , met een gemiddelde DS productie van 1.89 g per mg opgenomen P. Deze waarden bedroegen voor de bambara aardnoot en duivenboon respectievelijk 0.79 en 0.56. Uit dit onderdeel kan geconcludeerd worden dat het ontbreken van een respons van de bambara aardnoot op P-bemesting in het veld niet toegeschreven kan worden aan een uitzonderlijk hoge Pbenutting ,maa r waarschijnlijk aan z'n lage P behoefte. Eenander everklarin gvoo rhe tverschi l inrespon so pP-bemestin g ind epot -e nveldproeve nka n berusten opee nverschi l in dewijz e waaropd emeststo f werd toegediend. Terwijl demeststo f ind epotproeve n wasgemale n enhomogee n werd verdeeld over het gehele potvolume voordat werd gezaaid, werd in de veldproef de meststof in korrelvorm breedwerpig toegediend en vervolgens met een vork ingewerkt. Dit heeft tot gevolg dat in potten het fosfaat meteen na kieming voord eplan t beschikbaar is.I nhe t veld zalhe t waarschijnlijk veel latervoo r de plant beschikbaar komen.Di twa saanleidin gto the topzette n van eenpotexperimen t (Hoofdstuk 3.2) 80 Samenvatting waarind eP meststo f opverschillend etijdstippe n werdtoegediend . Heteffec t vantoedienin g bij zaaien, twee weken na zaaien en vier weken na zaaien werd bestudeerd. De resultaten gaven duidelijk aanda tP-bemestin g slechtsda nto tee nsignificant etoenam ei nD Sproducti ee nto tee n verhoogde zaadproductie leidt, wanneer de meststof binnen een periode van twee weken na zaaien wordt toegediend. Dit bevestigt de eerdere veronderstelling dat een breedwerpige toediening van P meststof inhe t veld geen effectieve manier iso md ebambar a aardnoot plant tijdig (direct nakieming ) vanvoldoend e Pt e voorzien. In develd - enpotproeve n (Hoofdstukken 2e n3 )blee kee n P-bemesting geen effect tehebbe n opd eP gehalte ni nd ebladere nva nd ebambar aaardnoot , ongeacht deaar d van de behandeling (nivo van de P gift, tijd van Ptoedienin g en irrigatie). Dit gaf aanleiding tot de vraag of deP gehalteni nd esprui tva nd ebambar aaardnoot ,zoal sgevonde ni nd evoorafgaand e experimenten, voldoende waren of niet. Om deze vraag te kunnen beantwoorden werd het kritisch P gehalte voord esprui tva n debambar a aardnootbepaald . Daartoewer d inee n zand-culture experiment fosfaat op 12 nivo's aangeboden, varierend van 0.21 tot 159 mg P/plant voor de gehele experimentele duur. Het kritisch P gehalte, grafisch bepaald en corresponderend met een P gehaltei nd esprui twaarbi j90 %va nd emaximal esprui tD Sproducti eword tverkregen ,bedroe g ongeveer 0.15% met een kritisch traject van 0.15-0.20%. Het kritisch P gehalte geeft aan dat bambara aardnoot planten in alle voorafgaande experimenten (pot- en veldproeven) onder suboptimale Pvoorzienin g groeiden. Vervolgens (Hoofdstuk 4.2)wer dee nonderzoe k uitgevoerd ind epraktij k (op deboerderij) . Bij eentienta lboerderi je ni nBotswan awer dd eP toestan dgemete nva ngronde no ppercele nwaaro p onder'praktijkomstandigheden 'bambar aaardnoo tword tgeteeld . Ookhe tgewa so pd e percelen werd bemonsterd enn a chemische analyse werden deP gehalten inhe t gewas vergeleken met het kritische P gehalte zoals gevonden in het zand-culture experiment (Hoofdstuk 4.1). De bodem-e ngewasmonster so pel kva nd etie nboerderije n werdenverzamel dongevee r 78dage n nazaaien .D egehalte naa nP i nd ebode mva n 80%va nd eboerderije n warenlage rda n5 mgP/k g (P-Bray). Slechts op twee bedrijven werden P gehalten in de spruit gemeten gelijk aan het kritisch nivo,terwij l op alle andere bedrijven de Pgehalte n in het gewas lager waren dan het kritisch nivo. In alle voorafgaande experimenten groeiden de bambara aardnoot planten bij marginale P voeding, gekarakteriseerd door P gehalten in de spruit op of onder het nivo van het kritisch gehalte, ongeacht bemesting met P of irrigatie. De N voeding van de planten in deze experimenten bleeksteed svoldoend et ezijn . Ditheef t geleid tot devraa g ofth eN voedin g nog welvoldoend e isal sd egroe iva nd ebambar aaardnoo tnie t langermee rgerem dword t door een suboptimale Pvoorziening . Onderdergelijk e omstandigheden zalnie t alleen als gevolg van een verhoogde droge stofproductie de behoefte aanN toenemen, maar gedraagt mogelijk ook het procesva nd estikstofbindin g zichanders .O mee nantwoor d te kunnen geven op dezevrage n is een potproef uitgevoerd (Hoofdstuk 5) om na te gaan (i) of de capaciteit van het stikstofbindingsproces, mogelijk aangevuld met opname van minerale stikstof uit de bodem, toereikend iso m deplan t van voldoende Nt e voorzien ookonde r omstandigheden waarbij de 81 P voeding niet langer meer groeilimiterend is,e n (ii) ofbi j P gehalten van de spruit hoger dan hetkritisc hniv od ehoeveelhei dN 2di egebonde nwordt ,toeneem t metee n verdere toenameva n hetP gehalt ei nhe tblad .I ndi texperimen t werdenbambar aaardnoo t planten bij zaaien bemest met oplopende hoeveelheden P, al dan niet in combinatie met een gift aan minerale N. Een bemesting met P resulteerde niet alleen in een verhoogde DS productie van spruit, wortel en wortelknolletjes, maaroo ki nverhoogd ezaa dopbrengs te nhoger eopnam ee ngehalte nva nP e n N ind espuit .Ee nbemestin g metmineral eN ha d geen significant effect opd e bovengenoemde plantparameters ,maa rbeinvloedd e (vervroegde)we ld everouderin gva nd eplanten . Ookonde r conditieswaa rP nie tlange rmee rgroeilimiteren d is,schijn td eN 2bindin gdoo rrhizobiu msame n met minerale N in de bodem in de N behoefte van bambara aardnoot te kunnen voorzien. De resultaten makenduidelij k datee nN bemestin gbi j zaaien overbodig is,maa rdi t sluitnie tui tda t eenbemestin g metN opee nlate rtijdsti p welpositiev e gevolgen kanhebbe n voord e groei van de bambara aardnoot. Een dergelijk tijdstip zou kunnen zijn het begin van peulvorming, een momentwaaro pee ngrot everschuivin gplaatsvind t ind eassimilate n verdeling binnen deplant , n.l. van vegetatieve (spruit) naar generatieve delen (peulen). Om dit te onderzoeken is een potproef met de bambara aardnoot uitgevoerd waarbij minerale Nwer d toegediend: bij zaaien (zoals in het vorige experiment), vroeg tijdens de vegetatieve fase, bij 50% bloei en op het moment dat 50%va nd eplante npeule nha d(Hoofdstu k 6).O md ebijdrage n vand eN 2 binding en van minerale N uit de bodem aan de totale N voeding van de bambara aardnoot te kunnen ontkoppelen,ware nbehandelinge n metgesteriliseerd e en ongesteriliseerde potgrond (controle) in de proef opgenomen. Een bemesting met minerale N bleek noch de DS productie van de spruit,noc hd ezaa dopbrengs t tebeinvloeden . EenN toedienin g ophe tmomen t van 50%bloe i resulteerde in een afname van het aantal en het totale gewicht van de wortelknolletjes en verlaagde deN opname enhe tN gehalt e ind ebladeren . Dit kan er opduide n dat nodulatie bij de bambara aardnoot tijdens deze fase van de groei gevoelig isvoo r eenverhoog d aanbod aan mineraleN . Sterilisatieva nd egron dresulteerd e inee nafnam e vanzowe lhe taanta l alshe ttotal e gewichtva nd ewortelknolletjes . HetN gehalt ei nd esprui tva nplante n geteeld in gesteriliseerde grond was lager op het moment van 50%bloei , waarbij tevens sprake was van een vertraagde bloeie npeulvormin gn asterilisatie ,n.l .me tongevee r2 weken .D eaanvanklij k lagereN gehalte n in de spruit van planten op gesteriliseerde grond herstelden zich later tijdens de groei ondanks de lage nodulatie. Dit duidt op een hoge efficientie van de aanwezige wortelknolletjes in de gesteriliseerde grond.Aangezie n ern a sterilisatie toch nog wortelknolletjes aanwezig bleken te zijn washe tnie tmogelij k omd ebijdrag e vanmineral e Nui t debode m in detotal eN voeding van debambar a aardnoot te kwantificeren. Duidelijk is evenwel dat een bemesting met stikstof voor de groei en zaadproductie van de bambara aardnoot geenmeeropbrengs t oplevert, ongeacht hetmomen t van toedienen. Tenslotte volgt inHoofdstu k 7 een algemene discussie van debelangrijkst e resultaten van dit onderzoek, met speciale aandacht voor derespon s van debambar a aardnoot op een bemesting met P en N en de betekenis hiervan voor deteler s van bambara aardnoot in Botswana. Het is duidelijk dat Pbelangrij k is voord e groei enzaadproducti e vand ebambar a aardnoot. Uit een praktijkonderzoek bij boereni nBotswan ablee k defosfaa t voeding van debambar a aardnoot in de meeste gevallen groeilimiterend te zijn. Dit kon worden aangetoond met behulp van het kritisch P gehalte voor bambara aardnoot verkregen uit de zand-culture proef. Dat bambara aardnoot in de praktijk bij een lage P-toestand van de bodem toch nog kan produceren is waarschijnlijk het gevolg van derelatie f geringeP behoeft e van het gewas. 82 Samenvatting Omd eproducti eva nhe tgewa st everhoge ni see nverbeterin gva nd eP voedin gee nvoorwaarde . Echter onder de heersende klimaatsomstandigheden in Botswana belemmert vooral een vochttekort de aanvoer van voldoende P naar de wortel, ook wanneer door bemesting het P gehalte wordt verhoogd. Bemesting met Pheef t dan ook pas zin nadat het vochtgehalte in de bodem is verbeterd, of door irrigatie of door een beperking van de teelt tot de seizoenen of strekenme tvoldoend e neeslag.Ui the tonderzoe k blijkt verder dat debambar aaardnoo t slechts dan reageert op P indien de meststof binnen een periode van 2 weken na zaaien wordt toegediend. Ind epraktij k dientda n ookbi j zaaien demeststo f zodich tmogelij k bij het zaad te worden aangebacht, zodatd eplant(enwortel ) meteenn akiemin gerove rka n beschikken. Een bemesting van de bambara aardnoot met minerale N leidt niet tot verhoogde groei en opbengst, ongeacht het moment waarop demeststo f wordt toegediend. Kennelijk kan deplan t via N2 binding aangevuld met minerale Nui t debode m overvoldoend eN beschikken. Uit dit onderzoek blijkt dat een bemesting metN eerder leidt tot opbrengst derving danto t opbrengst verhoging. 83 Curriculum vitae GaebeweModirw aRamoleman awa sbor no n2 3Augus t 1956i nSerowe ,Botswana . In 1982h e completed aB.Sc . inGenera lAgricultur e atth eUniversit y ofBotswan a and Swaziland. In June 1982 hejoine d the Ministry ofAgricultur e in Botswana and worked asa n Assistant Regional AgriculturalOffice r inth eDepartmen t ofAgricultura lExtensio n Servicesunti lSeptembe r 1983. From October 1983t o July 1986h eworke d asa Research-Extensio n Liaison Officer ina join t project of Department- of Agricultural Research and Agricultural Extension Services in the Ministry ofAgriculture . In August 1986h e went to USAt o study for M.Sc. in agronomy and completed inJun e 1988.H econtinue d withth epositio n of Research-Extension Liaison Officer untilMarc h 1990.Fro mApri l 1990t o September 1991h eworke da sa nAgronomis t responsible for arable commercial farms (Pandamatenga Commercial Farms) under the Division of Crop Production of the Ministry of Agriculture. In October 1991, still in the Division of Crop Production,h ewa sre-assigne da sSenio rAgricultura lOffice r Food Security.Fro m October 1992 hejoine d Botswana Collegeo fAgricultur e asa Lecture r insoil san dsoi l fertility. In 1994h ego t aPh. Dsand-witc h fellowshipfrom th e Wageningen Agricultural University partly supported by Botswana College ofAgriculture . 85 Appendix Appendix 1.Map so fBotswan ashowin gth elocatio nwher eth estud ywa sconducte d (Gaborone) andth e annual rainfall pattern. REPUBLIC OFBOTSWAN A C nepueuc OF BOTSWANA BOTSWANA 87 Appendix 2.Long-ter mweathe rdat aGaboron e(24°40'S ,25°'E ;altitud e 100m) .R.H .= Relativ e Humidity; P.E.T. =Potentia l Evapotranspiration (Pennman). Month Av. Max. Min. Av. Av. R.H. R.H. P.E.T. Rainfall Rainfall Rainfall max. min. 08.00 14.00 (mm) (mm) (mm) (mm) temp temp (%) (%) (°C) (°C) 1) 1) 1) 2) 2) 3) 3) 4) Jan 97 324 10 32.6 19.6 70 42 161 Feb 84 334 6 31.8 19.3 73 44 132 Mar 69 311 3 31.7 17.4 74 43 124 Apr 41 267 0 30.4 13.4 78 41 85 May 13 82 0 27.6 8.3 77 34 63 Jun 5 66 0 22.3 4.7 77 32 44 Jul 3 74 0 22.6 4.4 74 30 48 Aug 4 34 0 25.5 7.3 64 27 80 Sep 15 111 0 29.7 12.3 55 26 122 Oct 44 142 2 31.1 16.1 58 31 155 Nov 69 192 2 31.5 17.9 64 36 156 Dec 87 222 1 32.3 18.8 66 39 166 Annual 531 924 224 28.6 13.3 69 35 1336 1)1925 - 1997 1) 1961- 1996 3) 1963-1996 4) 1958-1975 Source:Anon. , 1998 88 Appendix 3. Thedistributio n ofrainfal l andth etim ean d quantity of irrigation from sowing on the 18t h December 1995 until final harvest on the 22 nd April 1996 in a field experiment at Sebele Botswana. Days from sowing Rainfall (mm) Irrigation (mm) Total (mm) 0-10 24.0 - 24.0 11-20 0.5 9.0 9.5 21-30 48.0 9.0 57.0 31-40 116.5 - 116.5 41-50 14.5 - 14.5 51-60 27.8 - 27.8 61-70 - 12.0 12.0 71-80 69.0 - 69.0 81-90 - 12.0 12.0 91-100 - 12.0 12.0 101-110 42.0 - 42.0 111-120 - - - 121-130 18.0 - 18.0 Total 360.3 54.0 414.3 89