A nutrient omission trial according to “A double pot technique for rapid soil testing” (JANSSEN, 1974) with soybean (Glycine max) as test plant has been conducted on a Rwandan research station to investigate the nutrient status of four Rwandan soils. These soils were identified as low responding to P-fertilisation and inoculation, when growing soybean and climbing bean (Phaseolus vulgaris), respectively. The soil samples were taken across Rwanda from three different provinces (1x Northern province, 1x Eastern province, 2x Southern province), representing three different Agro-Ecological-Zones (AEZ). Every macronutrient apart from Ca (N, P, K, Mg, S) was tested separately; micronutrients were tested combined in one treatment. Lime treatments were integrated as a third experimental factor for two acid soils (pH < 5). Following indicators were used to evaluate the performance and growth of soybean: Visual observations, aboveground biomass, stem height, Sufficiency Quotient (derived from Relative-Growth-Rate), nodulation, activity of nodules and plant tissue analysis. Every measurement was taken at three growth stages: 14 days after emergence (DAE), 26 DAE and 34 DAE.
Plants grown in every soil showed lean performance and minor growth when K was omitted. Clear potassium deficiency symptoms were detected on every -K treatment. Aboveground biomass, stem height and nodulation were significantly reduced already in early growth stages. On average the final biomass of -K treatments was reduced by 71.4 % compared to control treatments with full nutrient supply (nut no N) and by 91.9 % compared to N-fertilised treatments. A laboratory analysis classified all soils to have a “low” potassium content. These results demand an increased attention to potassium fertilisation and prevention of K losses.
In general N-fertilised treatments had an unexpected high biomass production. Their growth exceeded the growth of all other treatments for many times. Even treatments with complete nutrient supply (but no N)and seed inoculation, creating favourable conditions for biological nitrogen fixation (BNF), could not compete with fertiliser-N-supplied treatments. A coherent explanation for this unexpected difference in biomass production is still missing.