Thesis (PhD)--Stellenbosch University, 2004. / ENGLISH ABSTRACT: This project aimed at enhancing the net benefit in production systems. It took a holistic approach to evaluate the potential interactions of herbaceous legumes in relation to weed dynamics, soil fertility and livestock management in the crop-livestock system in Nigeria. The project was carried out between 2000 and 2002 in two localities. These were the National Animal Production Research Institute (NAPRI) at Zaria in the northern Guinea savannah and the International Institute of Tropical Agriculture (IITA) at Ibadan in the derived savannah. The main experiment was carried out in the northern Guinea savannah, while the secondary experiments were simultaneously conducted in the derived savannah and the northern Guinea savannah. The experimental design for the three experiments reported in this thesis is a splitsplit plot, fitted into randomised complete block design (RCBD), with four replications. Main plot treatments were herbaceous legumes, namely Vigna unguiculata, Arachis hypogaea, Glycine max, Aeschynomene bistrix, Centrosema pascuorum, Stylosanthes guianensis and natural vegetation. Sub-plot treatments were management systems (1) M1, 'residues left in the field'; (2) M2, 'residues taken out of the field' and (3) M3, 'residues fed to livestock, manure/urine/refused feeds returned'. Sub-plot treatments were administered in a sequence following rotational fallows of herbaceous legumes and natural vegetation. However, plots in the secondary experiments were not subdivided before the cropping of maize in 2002, and for logistical reasons only two sub-plot treatments, M1and M2, are featured in this experiment. Herbaceous legumes were established at the start of the rainy season, approximately in June, in 2000, 2001 and 2002. All herbaceous legumes received single super phosphate (SSP) at 20kg ha" P20S at planting, while minimum hand weeding was done to maintain pure legume stands during the establishment phase. Forage biomass was higher in the derived savannah than in the northern Guinea savannah. Similarly, higher forage yields were observed after two consecutive years of legume fallow and natural vegetation, compared to the first year plots. Grain yield for Glycine max was consistently higher than for the other two grain legumes in 2001 and 2002. Chemical analysis of herbaceous legumes and natural vegetation showed that crude protein values ranged between 11.2% to 17.3% for legumes; that was significantly (p<0.05) higher than the 8.6% value found for natural vegetation. Moreover, all herbaceous legumes and natural vegetation, except Arachis hypogaea, had dry matter digestibility values of more than 30%. Maize grain and stover yields on herbaceous legumes fallowed plots were evaluated and compared with those for natural vegetation. Results in 2001, i.e. after a one-year fallow with legumes, indicated that the dry matter of maize grain and stover yields in the Stylosanthes guianensis plots were higher among the forage legumes. Arachis hypogaea gave the highest grain and stover yields among the grain legumes in the northern Guinea savannah. Results in 2002, i.e. after a two-year fallow, also showed that the productivity of maize planted on Arachis hypogaea and Glycine max fallowed plots were consistently higher across the three management systems tested in the Centrosema pascuorum, Glycine max, Aeschynomene histrix, Vigna unguiculata and least weight gain was recorded for the natural vegetation. Objective functions in linear optimisation, or linear combinations in algebra, used to link dynamic processes in livestock production (Iiveweight gain) with the dynamic processes in soils (soil nitrogen), weeds (weed biomass), herbaceous legumes (legume biomass) and crop production (maize grain and stover yields) under varying management systems took the form: Management system 1, Y, = f (XI, Xz, x4); Yr= 0 Management system 2, Y, = f (x" x2); Yz = 0 Management system 3, Y, = f (XI, Xz, xl); Yz= f (XI, Xz, X4. x5) Where (1) (2) (3) Y1 = Crop in kg; Y2 = Livestock weight gain in kg; X1= Weed in kg; X2 = Soil N g kg-1; X3 = Livestock compost in kg; N= Herbaceous legumes in kg; X5 = Maize stover in kg, for the three management systems considered in this experiment. Deductions from these equations showed that Aeschynomene histrix performed better under M1, i.e. when legumes residues were left on the field. Natural vegetation performed better than the herbaceous legumes under M2, i.e. when legumes residues were exported out of the field. However, the presence of manure in M3 enhanced soil fertility in the system and improved the overall productivity across all the legumes and natural vegetation. Overall rankings, conducted by pooling all components in the system, indicated that Glycine max performed best among the legumes, followed by Stylosanthes guianensis and Arachis hypogaea, which ranked second and viithird respectively. In relation to the specific legume groups, Stylosanthes guianensis performed better than the other two forage legumes, while Glycine max also performed better than the other two grain legumes tested. We can see from these on-station research results that there are indications of positive opportunities for improving overall productivity and resources. This can be done through integrating and complementing crop and livestock production, to provide a sustainable intensification of agriculture. / AFRIKAANSE OPSOMMING: Die studie is uitgevoer om die potensiële wisselwerking van kruidagtige peulplante met onkruiddinamika, grondvrugbaarheid en veebestuur te evalueer. Die hoofeksperiment in die "Northern Guinea Savannah" is vanaf 2000 tot 2002 uitgevoer by die "National Animal Production Research Institute" (NAPRI) in Zaria, Nigerië. Die twee sekondêre eksperimente is gelyktydig in Zaria en by die "International Institute of Tropical Agriculture" (liTA) naby Ibadan in die "Derived Savannah" uitgevoer. Die eksperimentele ontwerp was 'n dubbel gesplete perseelontwerp gepas in 'n volledig ewekansige blokontwerp met vier herhalings. Die hoofkomponente was die kruidagtige peulplante naamlik: Vigna unguiculata, Arachis hypogaea, Glycine max, Aeschynomene histrix, Centrosema pascuorum, Stylosanthes guianensis en natuurlike plantegroei. Die gekose peulplante is potensieel aangepas vir uiteenlopende omgewings en word dikwels na verwys as "'n mandjie van opsies". Subperseel behandelings was (1) peulplant gevestig en gelaat op die land - M1; (2) peulplant gevestig, geoes en weggeneem uit die land M2 en (3) peulplant gevestig, geoes, vir vee gevoer, mis/urine/vermorste voer terug na die land - M3. Die dubbel gesplete perseel behandelings is toegedien in 'n sekere volgorde nadat die peulplant rusoes/braak toegepas is. In die eerste jaar is die kruidagtige peulplante geplant op die hoofperseel van 25m by Sam. In die tweede jaar is die hoofperseel verdeel in twee persele waar onderskeidelik peulplante en mielies gevestig is terwyl daar in die derde jaar 'n verdere verdeling was wat gelei het tot verskillende gewasrotasiestelsels, nl. Peulplant-peulplant-mielies, peulplant-mielies-peulplant en peulplant-mielies-mielies. Parameters wat insluit planthoogte, -wydte en persentasie grondbedekking van die kruidagtige peulplante is tweeweekliks gemeet op vier 1M2 persele wat ewekansig oor elke hoofperseel versprei is in beide gebiede waar die studie uitgevoer is. Resultate het getoon dat die kruidagtige peulplante wat getoets is potensiële kandidate is vir insluitings in gewas/weiding rotasiesteiseis. Alhoewel voerproduksie hoër was in die "derived savannah" as in die "northern Guinea savannah", het die prestasie van die peulplante in die noordelike savannah gevarieer met die gewasproduksiestelsels. Hoër opbrengste is gerealiseer na twee opeenvolgende jare van oesrus met peulplante vergeleke met die eerste jaar waar daar geen residuele effek van die peulplante was nie, en die laaste jaar wat deur 'n mielie-oes voorafgegaan is. Biomassa opbrengste na twee jaar van aanhoudende verbouiing was die hoogste vir S. guianensis en die laagste vir A. hypogea. Ruproteien inhoud van die kruidagtige peulplante het gewissel van 170 g kg-1 DM in A. hypogea tot 62.4 g kg-1 DM in A histrix. Graanproduksie deur G. max was deurlopend hoër in 2001 en 2002 vergeleke met die twee ander graanproduserende peulplante nl. V. unguiculata en A. hypogea. Die voerproduksie in 2002 was heelwat hoër as die vorige jaar. Rotasie effekte op mieliegraan en oesreste na peulplante is vergelyk oor die dubbel gesplete persele (areas met verskillende oesruslengtes) om die ximplikasies van hulle residuele effek op grondvrugbaarheid verbetering en onkruid dinamika te bepaal. Mielie-opbrengs na een jaar van rusoes toon hoër waardes op persele wat onder kruidagtige peulplante was vergeleke met natuurlike plantegroei. Net so was die opbrengs deurlopend hoër op persele waar daar graan peulplante was as waar daar voer peulplante was. Oor die algemeen was die waardes hoër vir G. max, gevolg deur A. hypogea, A. histrix, C. pascuorum, V. unguiculata, S. guianensis en die laagste vir natuurlike plantegroei. In terme van die bestuurstelsels, het persele wat kompos ontvang het (M3) beter as die ander twee bestuurstelsels presteer (2.6 Mg ha" mieliegraan). Produksie van mieliegraan en oesreste na twee opeenvolgende jare van mielieverbouing was die hoogste na G. max (7.2 Mg ha" mieliegraan), gevolg deur die A. histrix perseel en die laagste op S. guianensis persele. Algemene waarnemings oor die twee subpersele wat met mielies beplant was in 2002 het getoon dat mielies beter presteer het op persele wat twee opeenvolgende jare met peulplante beplant was. Onkruidsamestelling en verspreiding is in beide die peulplante en natuurlike plantegroei gemeet. Grond vir saadbankontledings is op diagonale transekte in 0.5m by 0.5m kwadrate gemonster. Grondmonsters is ge-analiseer vir pH, totale stikstof, organiese koolstof, fosfor, kalsium en magnesium. Hierdie parameters is gebruik in 'n meervoudige regressie ontleding om hulle effek op onkruidspesievoorkoms te bepaal. Onkruidgetalle in lande na In rusoes het In deurlopende hoër vlak van besmetting getoon op die natuurlike plantegroei persele as op die peulplant persele. Die prestasie in terme van onkruidonderdrukking was in dalende volgorde: S. guianensis, C. pascuorum, V. ungucuilata, G. max, A. hypogaea en A. histrix. Die onkruidsamestelling het verskilonder die verskillende behandelings en dit het ook met tyd verander in dieselfde behandelings. In Bykomende eksperiment met die peulplante is uitgevoer om hulle effek op inname en groei van skape, asook die effek op kompos wat gemaak is van vermorste materiaal en uitskeidings van die skape, te bepaal. Droë materiaal verteerbaarheid was hoog vir S guianensis, G. max en A. histrix terwyl die laagste syfer verkry is by A. hypogea (177.6 g kg-1 DM). Ramme wat met kruidagtige peulplante gevoer is het beter presteer as die wat met natuurlike plantegroei gevoer is. Ramme wat met A. hypogaea gevoer is, het In gemiddelde daaglikse toename (GOT) getoon van 85.7 g daq", gevolg deur S. guianensis, C. pascuorum, G. max, A. histrix, V. unguiculata en laaste natuurlike plantegroei. Bykomende ontledings was gemik daarop om objektiewe funksies af te lei om dinamiese prosesse in vee (massatoename) met dinamiese prosesse in grond (grond N), onkruid (onkruidmassa), kruidagtige peulplante (peulplantmassa) en mielies (mieliegraan en oesreste massas) onder verskillende bestuurstelsels te verbind. In In poging om objektiewe funksies xiivan die verskillende komponente van die studie te bepaal, is die volgende lineêre funksies vir die drie bestuurstelsels oorweeg nl. Bestuurstelsel1. Yl = f (Xl, xz, x.); Yz= 0 Bestuurstelsel 2, Yl = f (Xl, Xz); Yz= 0 Bestuurstelsel3, Yl = f (Xl, X2, Xl); Yz= f (Xl. Xz. x.. les) Waar (1) (2) (3) Y1 = Oesopbrengs in kg; Y2 = Daaglikse massatoename in kg; X1 = Onkruidmassa in kg; X2 = Grond N in g kg-1 ; X3 = Kompos in kg; X4 = Kruidagtige peulplante in kg; X5 = Mieliereste in kg. Onder bestuurstelsel 1, het A. histrix beter as die ander gewasse presteer, terwyl natuurlike veld beter presteer het onder bestuurstelsel 2. Dit is 'n aanduiding dat bestuurstelsel 2 nie volhoubaar is nie. Die derde bestuurstelsel verteenwoordig volle integrasie van gewas en vee produksiestelsels. Die teenwoordigheid van mis in die stelsel het grondvrugbaarheid verbeter en algemene produktiwiteit verbeter. Hierdie resultate dui aan dat daar geleenthede is om algemene produktiwiteit te verbeter deur integrasie en komplementering van gewas- en veeproduksiestelsels om volhoubare intensifikasie van landbou te bereik.