This study was conducted in Kabale District in the highlands of south-western Uganda with farmers linked to urban markets. Kabale District covers an area of approximately 1800 square kilometres and lies between −1.20774˚S and 29.98588˚E at an elevation of 2076 meters above sea level at the Uganda-Rwanda boarder. The mountainous terrain of Kabale District is part of the northern slopes of the Mufumbiro Range which encompasses the tallest mountain, Muhavura (4127 m). The highlands of Kabale are composed of the Karagwe-Ankolean rock system which is largely of Precambrian age. The uplift to its present level occurred in association with the formation of the Western Rift Valley in East Africa. The highlands are dominated by steeply sloping ridges, hills and narrow valleys [7] . The highlands have steep long slopes that are intensively cultivated for potato under low input production systems [4] . Kabale has a population density of 144 persons km⁻² and it is considered among the most densely populated districts in Uganda [8] . The district has a montane type of climate with an annual rainfall of 800 - 1000 mm. The long rains occur between mid August-December, followed by short rains in mid March-May and a short dry spell between June and July [9] . Subsequently, Kabale is food self-sufficient, although the soils have over time been losing fertility due to over cultivation, soil erosion and lack of resources for soil fertility rejuvenation [4] . The soils on hill slopes are shallow dark brown with low productivity, while the valley bottoms have deep humose brown loam with moderate to high productivity. Fragmentation of land holdings is severe and the average landholding is about 0.32 ha per household [4] .

a) Selection of potato fields

Six parishes in Kamuganguzi Sub County were selected based on their common bio-physical features and characteristics (land use types, plot sizes, soil types and vegetation). Using farmer group research approach, potato fields located on different slope positions and relative distances from owners’ homesteads were selected. Using farmers’ perception on soil depth, relative position on hill-slopes and fertility levels, potato field typologies were determined. These were either confirmed or placed in the corrected field typologies (hilltop ± homestead, mid slope ± homestead and valley bottoms) during field transect walks with community members.

b) Soil sampling and laboratory analyses

Soils from different field typologies were sampled at depth of 0-15 cm and analyzed for chemical (soil pH, total organic matter, total N, extractable P, exchangeable K, Na, Mg and Ca) and physical (soil bulk density) properties [10] [11] .

c) Limiting nutrient and fertilizer response trials

Trials on limiting nutrients and fertilizer responses in potato on different field typologies were set up in six parishes in two rain seasons of 2017. Fertilizer treatments (Table 1 and Table 2) were randomly applied in plots of 4 m × 3 m which were planted with pre-sprouted potato variety (Victoria) at spacing of 75 cm × 30 cm. Each of the trial on limiting nutrients and fertilizer responses had 30 trials located on different field typologies. Analytical copper chloride (CuCl₂), zinc chloride (ZnCl₂), manganese chloride (MnCl₂) and Analytical Borax (Na₂B₄O₇) were sources for Cu, Zn, Mn and B respectively. Each of the micronutrients (Cu, B, Zn and Mn) was applied at the recommended rate of 1 kg ha⁻¹. Land preparation, planting and weeding were carried out using a hand hoe. At maturity (100 days after planting), the number of plants per plot were recorded before harvesting. Plants from the middle four rows in each plot were harvested and the number of tubers per plant and their grades recorded. The three grades were assessed by measuring the circumference of tubers. These were Grade A (>60 mm), Grade B (45 - 60 mm) and Grade C (<45 mm). Grades A and B are marketable

Fertilizer carriers were: Urea―source of N; Triple superphosphate (TSP)―source of P; Muriate of potash (MOP)―source of K; Single superphosphate (SSP)―source of P and S; Analytical magnesium oxide (MgO)―source of Mg; Analytical copper chloride (CuCl₂)―source of Cu; Analytical zinc chloride (ZnCl₂)―source of Zn; Analytical manganese chloride (MnCl₂)―source of Mn and Analytical Borax (Na₂B₄O₇)―source of B.

potato ware while grade C is considered seed size [4] .

The collected data were entered in Excel software for management and later analyzed using statistical software Genstat. Interactions of the different treatments were determined and Fisher’s LSD test was used to compare each pair of treatments analyzed using One-Way ANOVA (Analysis of variance).

Farmers were knowledgeable on the different levels of fertility and depths of potato fields located on the different slope positions (Table 3). Potato fields located on the upper slopes were often shallow with stony outcrops and extremely low fertility. This was an indicator of serious soil erosion and limited use of soil fertility improving innovations. The hill-slopes of the highlands of south-western Uganda were long and steep with high levels of erosivity [7] . Due to low soil fertility on upper slopes resulting into low potato yields, farmers often encroached and cultivated marginal areas such as wetlands as they searched for fertile soils. Soils in potato fields located in the valley bottoms were deep and fertile. These were extensively cultivated with limited use of soil fertility improving innovations such as fertilizers and manure that resulted into extensive soil nutrient mining.

Potato fields had soil extractable phosphorus (P) above the critical value of 15 mg/kg, below which potato growth could be negatively affected [12] . Soil nitrogen (N) and potassium (K) were inadequate and yet, they are heavily utilized by potato crop [12] . Among the soil parameters measures, only total soil carbon (SOC) and N that significantly varied with the farmer-market linkage types (p < 0.05) (Table 4). However, total SOC and N were below the critical value for optimum potato production [10] . Farmers under urban and non-urban market

*Critical bulk density above which, there is noticeable poor tuber development; **Critical soil chemical and nutrient values below which, potato growth and yields are negatively affected.

types applied soil nutrient below the optimum and yet potato is a heavy feed on N and K [12] . This indicated that linking farmers to urban markets did not lead to sustainable potato production. This was also revealed by the high bulk density under urban and non-urban market linkages that indicated exposure of the compacted sub soil as a result of soil erosion.

Other than, potato fields located on hilltops and far from homesteads (Hilltop − HH), there were no significant differences in tuber yields on different field typologies (p > 0.05) (Table 5). Soil potassium was the most limiting nutrient in potato fields located on hilltops and far from homesteads. Potassium follows soil nitrogen in quantity taken up by potato crop [12] . Potassium uptake influences tuber yields and size of tubers as well as a number of tuber quality factors including specific gravity, black-spot bruise susceptibility, chip and fry color, and storage quality. Potassium deficiencies reduce photosynthesis, reducing dry matter and starch formation [12] . Although, there were no significant differences in tuber yields on other field typologies other than Hilltop-HH, soil potassium was also the most limiting nutrients in potato fields located on mid-slopes and near homesteads (Midslope + HH) and valley bottoms. Potato fields located on hilltops and near homesteads (Hilltops + HH), micro-nutrients were the most limiting. Although soil P and N are macro-nutrients taken up in substantial amounts by potato, they were not the most limiting on the different field typologies. Similarly, sulfur (S) was not among the most limiting nutrients on the different field typologies.

Other than tuber yields of grade C (<45 mm), there were no significant different

1 = No fertilizer application; 2 = N + P + K + S + Mg + micronutrients; 3 = P + K + S + Mg + micronutrients; 4 = N + K + S + Mg + micronutrients; 5 = N + P + S + Mg + micronutrients; 6 = N + P + K + Mg + micronutrients; 7 = N + P + K + S + micronutrients; 8 = N + P + K + S + Mg.

NB: Other than the control, all treatments were given equal amounts of S, Mg and micro nutrients.

responses in tuber yields and number of tubers per plant on integrated soil nutrients (p > 0.05) (Table 6). During the study, the experiment was affected by bacterial wilt of potato caused by Ralstonia solanacearum that masked the effects of applied mineral nutrients on potato growth and yields. The disease was one the most important challenge in potato production in the highlands of south-western Uganda. It was known for reducing potato yields up to 100 % [13] .