PhD. ThesisMicronutrient and protein deficiencies are among the major public-health concerns in Tanzania. These deficiencies mainly affect children below five years of age and pregnant women. Malnutrition is a preventable problem if nutrition education and food-based approach strategies are effectively emphasised from the village to the national levels. The banana-based farming system of Kagera region of Tanzania has good edible food diversity. However, households mainly consume monotonous diets, which are energy dense and have inadequate content of micronutrients. Inadequate dietary diversity contributes to unacceptable high levels of under-nutrition, particularly protein energy malnutrition (PEM) and low intake and risks of deficiency of essential micronutrients including vitamin A and iron, particularly in children below five years of age. Popular dishes consumed by the population in Kagera region, are low in vitamin A and iron. Participatory dietary diversification is one of the strategies to address nutrient deficiencies and fight malnutrition. This study modified the local popular recipes to improve the nutrient levels of the foods consumed especially by children below five years of age. The study is a cross sectional design and the objectives included; (i) modify local dishes fed to children below five years of age to improve vitamin A, iron and protein intake; determine iron, vitamin A and protein content of the modified/improved recipes for children below five years; (iii) determine in vitro bio-accessibility of vitamin A and iron of the modified/improved recipes for children below five years; (iv) assess acceptability and preference of the modified recipes for children below five years. To enrich the preferred energy-rich diets, this study modified the local recipes to improve content of vitamin A, iron and protein to meet requirements for children of age between 6 to 23 months. Fifty mothers were randomly selected from Izimbya ward, Bukoba district participated in a recipe development exercise. Five recipes of banana-based from East African highland bananas (EAHB) ‘nshakala’ and triploid hybrid of Musa acuminata and Musa balbisiana (AAB) ‘bira’ banana varieties and three maize-based porridges were formulated in combination with other ingredients. Other locally available foods included beans, amaranths, red palm oil, pumpkin, groundnut, maize and orange-fleshed sweet potato. The recipe names included 1N (‘nshakala’ :EAHB, dry red kidney beans, amaranths, palm oil); 2N (‘nshakala: EAHB’, fresh red kidney beans, pumpkin leaves, sunflower oil); 3B (‘bira’’ :AAB, dry red kidney beans, amaranths, palm oil); 4B (‘bira’;AAB’, fresh red kidney beans, pumpkin leaves, sunflower oil); 5N (nshakala’, pumpkin fruit, groundnuts flour); 6OFSP (Fermented maize flour, orange fleshed sweet potatoes, groundnut flour); 7B (fermented maize, ‘bira’, groundnut flour); 8E (Fermented maize flour, egg, red kidney beans). Diet modification started with calculation of vitamin A, iron and protein contents of foods by using the Tanzania Food Composition Table (TFCT). The conversion factor of 12:1 for beta-carotene was used. Red palm oil, orange fleshed sweet potato (OFSP), pumpkin fruit and leaves were used to increase vitamin A content in the modified recipes. Furthermore, red kidney beans (Phaseolus vulgaris) and groundnut were included to improve protein and iron content of the formulated recipes. Red kidney beans formed a good source of iron, protein and energy when incorporated in complementary foods. The family dishes and children’s local popular dishes were made into a purée to improve the consistency (viscosity and density) to increase food intake and therefore vitamin A, iron and protein. The developed recipes contained more than 100% recommended dietary allowance (RDA) for vitamin A and protein. The RDA for iron in the modified diets ranged from 61% to 99%. Vitamin A, iron, protein and energy content of modified recipes ranged from 108-2768 Retinol Activity Equivalent (RAE), 6-17 mg, 28-56 g and 697-1635 kcal, in 500 g consumption portion, respectively. These levels meet the RDAs for breastfed and non-breastfed children. Furthermore, the modified recipes were subjected to laboratory analysis for nutrient contents and bio-accessibility tests. All samples were analysed in triplicates for vitamin A, iron and protein contents. For vitamin A; three carotenes; all-trans α-carotene, 13-cis-β-carotene and all-trans β-carotene were determined by High Performance Liquid Chromatography. Bio-accessibility was assessed using in-vitro bio-accessibility model in three phases; simulated gastrointestinal system oral, gastric and intestinal. The analysed provitamin A Carotenoids (pVACs) were converted into ‘Retinol Activity Equivalents’ (RAE). Total RAE of the modified diets ranged from 8.8 to 137.4 μg/100g, and after in-vitro digestion ranged from 0.87 to 13.3μg/100g. The bio-accessibility of pVACs ranged from 12.2% to 33.6%. In cooked food, pumpkin fruit contributed high amount of provitamin A followed by palm oil, ‘bira’ and amaranths. ‘Bira’ banana variety contributed high pVACs than local ‘nshakala’ banana variety. Provitamin A Carotenoids (pVACs) from pumpkins leaves were more accessible than those from amaranths and red palm oil fruit. Our results suggest that when carrying out interventions to improve diets, it is very important to take into account the estimation of dietary source of vitamin A and pVCAs and their bio-accessibility to meet nutritional requirements for vitamin A. The contents of iron and protein in separate ingredients and in modified recipes were analysed using flame atomic absorption spectrophotometry and Kjeldahl ́s method, respectively, and bio-accessibility of iron was estimated using in vitro simulating gastrointestinal digestion method. The contribution of iron ranged from 75 to 458% and protein from 106 to 146% of recommended dietary allowance (RDA) ranged from and in a portion of 500g . Iron bio-accessibility in all recipes ranged from 7.4% to 31.1%. Iron in porridge recipes made from orange fleshed sweet potato porridge and ‘bira’ porridge was more bioaccessible. The product might be having high nutritional value but without assessing its acceptability and preference, the product is likely to be not liked by the community. Therefore this study assessed sensory attributes of the modified/improved recipes for children in two rural villages of Tanzania. Consumer preference of eight recipes was assessed using a nine-point hedonic scale test. The results revealed that the improved ‘katogo’ dry beans (recipe 3 OC) and ‘katogo’ fresh beans (recipe 4 OD) had the highest scores for colour (7.9 - 8.1), aroma (7.6 - 7.7), taste (7.6 - 7.8), texture (7.7 - 7.8) and overall acceptability (7.8 - 7.9) compared to local ‘katogo’ steamed sardines (recipe 1 OA) and improved ‘katogo’ with groundnuts and pumpkin (recipe 5 OE). Recipe 3 OD was the most preferred recipe compared to other recipes Porridge prepared using maize flour and orange fleshed sweet potatoes scored highest for all attributes compared to porridge that had eggs in it (recipe 7 OH) and plain local maize porridge (recipe 8 OI). Modified banana-based and porridge-based were widely accepted by panelists. The importance of food diversification for intake of iron, vitamin A and protein is discussed. The communities need to establish home gardens so as to use vegetables and foods which are cheap and good sources of micronutrients and protein. Knowledge on nutrition education to enable community to accept other tastes than their own for better choice of healthy food is highly recommended. Sensory evaluation is important component in developing complementary food. The modified recipes based on locally available and affordable ingredients have a potential to meet RDAs of vitamin A, iron and protein for children aged 6 to 23 months in the banana-based system and other communities in Africa with the same settings.