Résumé Considérée par la Vision 2020 du Rwanda comme premier moteur de croissance économique et de réduction de la pauvreté, lagriculture rwandaise se prépare à rentrer dans une phase de mutations profondes de transformation et de modernisation de ses systèmes de production de subsistance pour souvrir au marché. Pour passer dune agriculture de subsistance à une agriculture commerciale, beaucoup denjeux et de contraintes du secteur agricole, variant et ne se faisant pas sentir avec la même acuité, dune région à une autre, méritent dêtre identifiés et analysés avant de passer à laction. Pour appréhender le sujet une étude documentaire détaillée a été réalisée sur la zone et sur le contexte agricole rwandais en général suivie dune série denquêtes sur terrain. Celles-ci reposent sur un échantillonnage raisonné de 4 secteurs (deux de basse altitude et deux de haute altitude) représentatifs de la diversité agro-bioclimatique et socioéconomique de la région dans chacun des deux districts de la zone de GIKONGORO. Tenant compte du nombre dunités de production présentes dans chaque secteur, les ménages ont été aléatoirement choisis dans les 8 secteurs retenus de manière à atteindre un effectif de 210 unités de production, pour recueillir leurs opinions sur les contraintes rencontrées, leurs stratégies de production et leurs desiderata en lien avec ce que les services agricoles conduisaient. Les problèmes relevés sont nettement complexes et semblent avoir pour facteur-racine la rupture de léquilibre « ressources naturelles/population », principal risque pour la croissance économique à long terme et la lutte contre la pauvreté. Le premier constat de cette étude est que les mutations démographiques, économiques, sociales ou culturelles qui ont affecté, depuis des décennies la zone de GIKONGORO, nont pas pu parallèlement, induire une dynamique organisationnelle qui se traduirait aussi par le développement et le renforcement des capacités des paysans et des organisations interpellées par les nouveaux enjeux du développement rural. Sur le plan agronomique, la principale contrainte sur laquelle convergent toutes les préoccupations des agriculteurs, est celui dun appauvrissement critique des sols. Celui-ci est principalement dû à une surexploitation des terres liée aux faibles superficies disponibles, aux difficultés dextension des terres de culture et à lincapacité des agriculteurs qui, faute de moyens, ne peuvent restituer au sol ce que les cultures exportent. Les pâturages naturels nétant plus disponibles en plus de la faible production fourragère dans la zone, la taille des troupeaux a sensiblement diminué et la production du fumier est faible. Sur des sols naturellement pauvres, érodés et très acides de GIKONGORO, le facteur limitant de la production agricole est avant tout la quantité de la fumure organique disponible sans laquelle tout investissement dans lamélioration de la fertilité des terres ne constitue quun gaspillage économique. Sur le plan de fonctionnement socioéconomique, les résultats de ce travail révèlent que 39% des exploitations sont très pauvres, procurent des revenus relativement inférieurs au seuil de reproduction et au seuil de survie pour certaines dentre elles et présentent une tendance générale à la régression. Evaluées à 47%, les exploitations à dominance agriculture procurent des revenus légèrement supérieurs au seuil de reproduction et montrent une tendance à la stagnation et se reproduisent à lidentique. Seulement 14% des exploitations de GIKONGORO sont en développement. Soutenues par des ressources extérieurs, elles produisent des revenus nettement supérieurs au seuil daccumulation. Elles ont des capacités dinvestissement relativement importantes et présentent une tendance générale à lélargissement.

Résumé La plupart des régions équatoriales humides, dont fait partie le Rwanda, sont caractérisées par une altération avancée des roches, des températures relativement élevées, de fortes intensités des pluies qui se traduisent par des phénomènes de ruissellement et dérosion accompagnées par la lixiviation déléments nutritifs avec pour corollaires lacidification et la réduction de la fertilité des sols. Lun des moyens daméliorer la fertilité passe par la mise en uvre de techniques destinées à modifier les caractéristiques défavorables du sol. Ce travail a pour objectif dévaluer les effets de lapport dans le sol de biomasses végétales fraîches collectées dans le milieu naturel, combinées avec du travertin, sur lamélioration de la fertilité et de la productivité des sols acides du Rwanda. Afin datteindre ce but, deux types dessais ont été mis en place sur le terrain au niveau du site de Rubona en moyenne altitude et de Gakuta en haute altitude. Il sagit tout dabord dun essai portant sur la mise en culture en milieu naturel de deux espèces végétales Tithonia diversifolia (Hemsley) A Gray et Tephrosia vogelii (Hook F) et lévaluation de leur production en biomasse et en quantités déléments. Lautre essai, consiste en lévaluation des caractéristiques des sols et des rendements de la culture du maïs des parcelles fertilisées avec les biomasses de ces espèces et/ou de NPK chacun combiné ou non avec du travertin. Le choix de Tithonia diversifolia a été basé sur son potentiel fertilisant et de production en biomasse tandis que celui de Tephrosia vogelii sur son utilisation dans les exploitations agricoles sous forme de haie anti-érosive. Les résultats obtenus indiquent que ces espèces ont un potentiel élevé de production en biomasse et en accumulation déléments dans leurs milieux écologiques favorables. Elles donnent des rendements de 18,7 t.ha-1.an-1 pour Tithonia diversifolia à Rubona et 11,4 t.ha-1.an-1 pour Tephrosia vogelii à Gakuta. Au terme de quatre saisons culturales dexpérimentation, les traitements avec la biomasse de Tithonia diversifolia combinée au travertin ont significativement augmenté le pH, la CECE, le Veff et réduit le taux de saturation en aluminium échangeable dans les deux sites. A Rubona, laugmentation de 0,8 unité de pH a induit un écart positif de 43,9 % et 2,9 cmolc.kg-1 de sol respectivement pour Veff et de CECE et une réduction de 60 à 11 % pour mAl. A Gakuta, ces écarts sont respectivement de 20,5 % et 4,1 cmolc.kg-1 de sol et de 20 à 0,4 % pour une amélioration de 0,7 unité de pH. Ce même traitement et celui relatif à la biomasse de Tithonia diversifolia seule ont amélioré les teneurs en Cot des sols de 0,1 à 0,5 %. Les effets des traitements sur la disponibilité du phosphore sont les plus marqués avec lengrais minéral NPK utilisé seul ou en combinaison avec le travertin. Les teneurs correspondantes ont été multipliées par 2,8 et 2,7 à Rubona et 1,3 à Gakuta à la quatrième saison. Sagissant des paramètres microbiologiques du sol (azote microbien, carbone microbien, respiration basale et respiration induite par le substrat), une amélioration sensible a été observée dans les traitements avec la biomasse de Tithonia diversifolia ou de Tephrosia vogelii en combinaison avec le travertin ou de la biomasse de Tithonia diversifolia appliquée seule. Lanalyse des échantillons de terres par la méthode de spectroscopie proche infrarouge a permis de prédire avec la meilleure précision (0,8<r2<1,0 et 2,3<RPD<5,4) la plupart des paramètres chimiques et microbiologiques pris en compte dans cette étude (pH, Cot, Nt, CEC, CECE, Nmic, Cmic, RBAS, RIS). Pour le maïs, les rendements en biomasse et en grains sont positivement influencés par le traitement avec la biomasse de Tithonia diversifolia combinée avec du travertin. Ce traitement présente une meilleure efficience agronomique. A la quatrième saison, le rendement en grain pour ce traitement a été multiplié par 4,7 et 2,8 par rapport au témoin respectivement à Rubona et à Gakuta.

Diss. (sammanfattning) Uppsala : Sveriges lantbruksuniv., 2009. / Härtill 4 uppsatser.

Little is known regarding the effects of land use on soil organic matter and P status of South African soils. For that reason, the effects of the main agricultural land uses in the midlands region of KwaZulu-Natal [maize (Zea mays), sugarcane (Saccharum spp), annual ryegrass pasture (Lolium multiflorum), permanent kikuyu pasture (Pennisetum clandestnum), gum (Eucalyptus grandis) and pine (Pinus patula)] on soil organic matter content, microbial biomass C and P and inorganic and organic P pools derived from a modified Hedley P fractionation was investigated on two sites where the longterm history of land management was known. In comparison with undisturbed native grassland, permanent kikuyu pasture resulted in an increase in organic C, organic P and microbial biomass C and P. Maize and sugarcane production resulted in a decrease in organic C, organic P and microbial C and P. Under annual pasture, gum and pine forests, organic matter and microbial biomass concentrations remained similar to those under native grassland. Under native grassland, extractable organic P accounted for 50% or more of the total P content of soils but under agricultural management with regular applications of fertilizer P, there was an increase in the percentage of total P present as inorganic P. Agricultural management greatly affected the distribution of P among the various inorganic and organic P fractions. Resin-Pi and NaHC03-Pi (the potentially-available forms of Pi) showed similar trends with land use being greatly elevated under kikuyu pasture at both sites and sugarcane and maize at one site. This accumulated Pi was thought to have originated from recent fertilizer applications and possibly recently mineralized organic P. Trends for NaOH-Pi with land use differed greatly from those of the Resin- and NaHC03Pi fractions. Concentrations were notably high under maize and sugarcane production. Of the pools of soil organic P, the NaHC03-Po fraction was most greatly affected by land use, being elevated under kikuyu and decreased under maize and sugarcane. This supports the assertion that it is the NaHC03-Po fraction that is the most labile soil organic P pool. It was concluded that land use greatly affects soil organic C and P status, soil microbial biomass C and P contents, soil inorganic P concentrations and the distribution of P among the various P fractions. A short-term (8 weeks) laboratory incubation experiment was carried out to compare the effects of inorganic (KH2P04) and organic (cattle manure, poultry manure and maize crop residues) sources of P, applied at a rate equivalent to 30 kg P ha-1 , on soil inorganic and organic P fractions and the potential availability of soil P. Additional treatments consisted of lime [Ca(OHhl at 5 ton ha-1 and lime plus inorganic P. Applications of lime raised soil pH to a similar extent after 1, 4 and 8 weeks incubation. After 8 weeks, a small increase in soil pH was also noted for the cattle and poultry manure and maize residue treatments. For the inorganic P fractions, substantial treatment effects were observed only for the Resin-Pi fraction. The inorganic P source was more effective than the organic ones at increasing Resin-Pi after 1 and 4 weeks incubation and of the organic sources, cattle and poultry manure were more effective than maize residues. Resin-Pi concentrations generally increased between 1 and 4 weeks incubation but then declined rapidly between 4 and 8 weeks incubation. After 8 weeks incubation, treatment effects on Resin-Pi were small. Concentrations of NaHC03-Pi, dilute HCI-Pi and concentrated HCI-Pi all declined over the incubation period. There was no clear trend with incubation for NaOH-Pi although for the poultry manure and maize treatments, concentrations declined between 4 and 8 weeks incubation. In general, concentrations of NaHC03-Po were greater for organic than inorganic P sources after 8 weeks incubation suggesting microbial immobilization of P in these treatments. There were increases in NaHC03-Po and concentrated HCI-Po over the incubation period suggesting progressive immobilization of P from the Pi fractions that declined in concentration during the incubation. Concentrations of NaOH-Po were not greatly affected by incubation period. The lime treatments, however, had lower NaOH-Po concentrations than the others suggesting that liming may have stimulated microbial mineralization of Po. Residual-P concentrations increased over the incubation period. This was attributed to conversion of extractable Pi fractions into recalcitrant, non-extractable Pi forms and/or immobilization of Pinto intransigent organic forms. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2006.

In the past, the indigenous knowledge of soils of small-scale farmers in South Africa has been largely ignored in scientific research. Hence the use of scientific approaches to land evaluation has often failed to improve land use in rural areas where understanding of the prescriptive scientific logic is lacking. Despite this, it is clear that local people and smallscale farmers have knowledge of their lands based on soil and land characteristics that remain largely unknown to the scientific community. It is therefore important for researchers to understand farmers’ knowledge of soil classification and management. To address this issue, a study was conducted in the uMbumbulu area of KwaZulu-Natal to investigate the use of indigenous knowledge as well as farmers’ perceptions and assessments of soil fertility. A preliminary questionnaire was designed to explore indigenous knowledge in a group interview that was conducted prior to the study. Another questionnaire was used to elicit indigenous knowledge from 59 randomly chosen homesteads representative of the population of Ezigeni and Ogagwini villages. Six homesteads were chosen for further detailed information on the cropping history, knowledge specific to the cultivated lands, detailed soil description and fertility. Soil samples were taken from these homesteads under different land uses (taro, fallow, veld and vegetable) at 0-30 and 30-60 cm depth for laboratory analysis. This was done to determine the effect of land use on soil physical and chemical properties and soil microbial activity. For scientific evaluation a general purpose free soil survey was conducted to produce land capability and suitability maps. Farmers identified ten soil types using soil morphological characteristics, mainly soil colour and texture. These soil properties were also used in the farmers’ land suitability assessment. In addition, slope position, natural vegetation and village location were used to indicate land suitability. The amount of topsoil was also used in land evaluation. However, slope position was considered the most important factor as it affects the pattern of soils and hence their suitability. Soils on the footslope were considered more suitable for crops than those found on the midslope and upslope. The yield difference observed between villages, which were higher in Ogagwini than Ezigeni, was also used as a criterion for evaluation. Farmers attributed these yield differences for various crops to the effect of soil type on productivity. In support, scientific evaluation found that Ezigeni village had a number of soils with a heavy textured, pedocutanic B horizon and hence a relatively shallow effective rooting depth. Moreover, the Ezigeni village land suitability was limited in places by poor drainage and stoniness. These limitations were rarely found for the Ogagwini village soils. Farmers had a total of six comprehensive and well defined soil fertility indicators, namely crop yield, crop appearance, natural vegetation, soil texture, soil colour and presence of mesofauna. Results showed that farmers’ fertility perceptions are more holistic than those of researchers. However, despite this, their assessment correlated with soil analysis. There was a close relationship between scientific and indigenous suitability evaluation for three commonly cultivated crops (taro, maize and dry beans). This was further substantiated by yield measurements which were significantly higher for Ogagwini as rated by both farmers and scientific evaluation as the more suitable. The significant agreements between the scientific and indigenous approaches imply that there are fundamental similarities between them. Recognizing this and subsequently integrating the two approaches will produce land use plans relevant and profitable for both small-scale farmers and scientists. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.

Soybean is emerging as a very important food, market and oil crop in Ethiopia. However, its productivity in Western Ethiopia is constrained by several production constraints, of which soil acidity is one of the most important ones. On acidic soils the availability of several plant nutrients is limited; among which phosphorus is the least available. Thus, development of high yielding and low P tolerant soybean varieties need to be among the top priorities in areas with such problematic soils. Therefore, the objectives of the study were to: 1) conduct a Participatory Rural Appraisal (PRA) study to assess farmers’ perception on various soil fertility, soybean consumption and marketing issues, 2) evaluate soybean genotypes under low and high P regimes, and 3) conduct genetic analysis of soybean performance under low and high P conditions. The PRA was conducted to assess farmers’ perception on various soil fertility, soybean consumptions and market issues. A total of 186 soybean producing farmers across three locations of Western Ethiopia were interviewed using a semi-structured questionnaire. Results from the study indicated that the use of soybean for crop rotation and soil fertility improvement was more important to the farmers than household consumption and marketing of the crop. The study also revealed poor demand for soybean compared to other crops on the local market. The majority of respondent farmers’ recognized that soil fertility has been declining over time and obtaining inorganic fertilizers on time was difficult; mainly due to high price of fertilizer. Though farmers’ cooperative was identified as the major supplier of fertilizer, farmers rated the quality of its service in supplying fertilizer as poor. With deteriorating soil fertility and limited capacity to use inorganic fertilizers, farmers are producing soybean under low soil fertility conditions. Thus, breeding programs need to develop varieties that perform well under low fertility soil. Screening soybean genotypes for response to different P regimes was performed in a field experiment using a split plot design, where the main plots were three levels of applied P (0, 100 and 200 kg ha-1 P), and the sub plots were 36 soybean genotypes (G) planted across three locations (L) with two replications. The extent of genetic variation of the 36 soybean genotypes was assessed under low (0 kg ha-1) and high P (100 kg ha-1) conditions. The analysis of variance revealed significant differences among genotypes for all the traits, except pod number at low P; while all the traits, except root volume, pod number, and number of seeds per pod showed significant differences at high P. Plant fresh weight, root fresh weight and root volume exhibited high genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) under both P conditions. Both principal component and cluster analyses revealed variation in the population. The 100-seed weight, plant height, roots and plant fresh weight combined high heritability and genetic advance estimates indicating that the inheritance of such traits is controlled by additive gene action under both P conditions. In general, the study revealed high genetic variation in the population, which can be exploited to improve performance under both high and low P conditions. The analysis of variance revealed significant genotype X phosphorus (GXP) interaction for number of nodules and total nodule weight at Jimma, and Assossa, and for root weight and root volume at Mettu. Though the GXP and GXPXL interactions showed non-significant difference for across locations analysis, the genotypes displayed significant difference for root fresh weight, root volume, tap root length, and weight of effective nodule. Genotypes: Pr-142 (26), AGS-3-1, SCS-1, AGS 234, and H 3 were identified among the best for root and nodulation characteristics. Yield and yield related traits were also assessed separately in the screening program. The results revealed significant GXP interactions for grain yield only at one site; while the genotypes exhibited highly significant differences for most of the traits in all the sites. G and GXL interaction were significantly different for most the traits. Essex 1, IAC 11, and AGS-3-1 were the best performing genotypes at high P; while genotypes IAC 11, AA 7138, G 9945 and AGS-7-1 displayed tolerance to low P. Genotypes AA-7138, PR-142 (26) and H3 exhibited stable performance across the three P levels. These genotypes have paramount significance in breeding soybean for low P tolerance and stable performance in varying P conditions for resource poor subsistence farmers.The genetic control mechanism for the major quantitative traits for performance under high and low P condition was studied in a nine parent half diallel cross. The results revealed that the GCA effects were highly significant for grain yield, pod length, days to maturity and plant height under low-P conditions. GCA effects were highly significant for grain yield, 100-seed weight, days to maturity, plant height, pod number, and pod length under high P. GCA effects were also significant for number of seeds per pod under high P condition. In addition, the relative contribution of GCA was higher than SCA under both P conditions, except for 100-seed weight at low P. Variety Hardee-1 was the best general combiner for most of the quantitative traits under both P conditions, indicating that it can be used in breeding programs to improve soybean for better genetic response to low and high P. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.

Diss. (sammanfattning) Umeå : Sveriges lantbruksuniv., 2007. / Härtill 3 uppsatser.

Diss. (sammanfattning). Umeå : Sveriges lantbruksuniv. / Härtill 4 uppsatser.

Lic.-avh. (sammanfattning) Umeå : Sveriges lantbruksuniv. / Härtill 2 uppsatser.

Diss. (sammanfattning) Uppsala : Sveriges lantbruksuniv., 2007. / Härtill 4 uppsatser.