PhD. ThesisPoor soil nutrient supply is often overlooked as a contributing factor of increased cyanogenic glucoside accumulation in cassava produced in areas affected by konzo, despite the low soil fertility in these areas. High cyanogenic glucosides in cassava roots is one reason attributed to cassava cyanide intoxication that sometimes leads to a health disorder called konzo. Various studies were hence carried out to investigate the role of nutrient supply on cyanogenic glucoside production in cassava varieties grown in areas affected by konzo, using konzo-affected Mtwara region as the study area. The first study was a survey, carried out using a questionnaire. In this study farmers’ perceptions on the causes of cassava root bitterness were investigated, with the interest of finding out whether farmers had observed that soils in the region were a contributing factor of increased cassava toxicity and to learn from them the characteristics of these soils. The second study was carried out to investigate whether soil nutrient levels of soils in konzo-affected areas of Mtwara region was inadequate for cassava production and hence capable of causing increased cassava cyanogenic glucoside levels. The third study determined whether soil nutrient supply could equally influence cyanogenic glucoside production in cassava grown in these areas just like variety and moisture supply. A 2×3×4 factorial pot experiment which consisted of two cassava varieties, three soil moisture treatments and four fertiliser treatments and laid-out as a randomised complete block design was used to carry out this study. The fourth study was carried out as a field experiment using four cassava varieties and six N-P-K fertiliser treatments, it was arranged using a split-plot design and was repeated in two consecutive years. This study was carried out to investigate the beneficial effects of improved plant nutrition (with N-P-K use) on root HCN levels and yields of commonly grown cassava varieties in the region. About 14.2% of the farmers perceived that soils, particularly nutrient poor soils, were a contributing factor to cassava root bitterness and therefore probably of increased cassava root cyanide toxicity. In agreement with farmers, the survey results showed that soils in the region were severely inadequate for cassava production, with 99.1%, 34.8%, 84.3%, 13.9%, 84.3%, 63.5% and 93.0% of the cultivated soils in konzo-affected villages of Mtwara region being deficient in N, P, K, Ca, Mg, S and Zn, respectively. Adequate levels of most of the deficient nutrients in these soils are known to reduce cyanogenic glucoside production in cassava. Soil nutrient supply (p < 0.001) was also found to have an equally important influence on cyanogenic glucoside levels as variety (p < 0.001). On the other hand, water was found to have no influence (p = 0.080) and was only important for the variety Kiroba whose cyanogenic glucoside production was influenced differently by water supply depending on the soil nutrient conditions (p = 0.033). Leaf HCN levels in the two varieties used ranged from 44.4 to 310.4 mg/kg. The study finally revealed that effects of fertiliser use on the total root cyanide (HCN) content (a measure cyanogenic glucoside levels) of some commonly grown cassava varieties was greatly dependant on variety and was mainly non-responsive to fertiliser. Inconclusive results were obtained on the effects of N-P-K on yields. With N-P-K applied at the rate of 25-10-50 kg N-P-K/ha, the fertiliser responsive variety Kiroba, gave the lowest root HCN level of 43.6 mg/kg under water stress conditions and attained fresh root yields of 26.0 t/ha. The important role of soil fertility on cyanogenic glucoside accumulation was clearly demonstrated.