Nymphaea petersiana Klotzsch (Nyika) is an important wild tuber eaten in some districts of Malawi. The tubers were processed by boiling/freeze-drying(BFD) and sun-drying(USD). The tuber's nutrient and antinutrient composition was determined to produce a preliminary nutrient data base for use in sub-Saharan Africa. There was no significant difference (P > 0.05) in protein content of BFDand USD samples. Sun-dried samples were significantly (P < .05) higher in ash than boiled samples while boiled samples were significantly higher (P < .05) in crude fat and total carbohydrate. The protein content of the tubers (8.0 and 8.1 %) was higher than that of the staple maize (7.9%), African millets (unspecified) (7.5%), and polished rice (7.0%), but lower than sorghum (10.7%). Protein content was higher than tubers like cassava (1.3%), potato (2%), sweet potato (1.6%), yams (1.5%) and N. lotus (5.2). Nyika tubers have a well balanced amino acid content, limiting only in lysine. There were no significant (P > 0.05) differences in the mineral content of BFD and USD samples, except for iron, which was lower in the boiled samples. Nyika tubers have a higher calcium (1376 and 946 ug/g) and phosphorus (2250 and 2883 ug/g ) content than wild and domesticated cassava, potatoes, sweet potatoes and wild and domesticated yams. Sun-dried tubers have a higher iron content (88ug/g) than maize (20ug/g). The zinc content of tuber was higher (21and 25ug/g) than that of boiled maize flour, boiled sorghum flour, rice, cassava, and sweet potato. The predominant fatty acids in the tubers were oleic (47%), linoleic (32%), palmitic (21%) and linolenic (7%) acids. Ascorbic acid content was very low, only 0.1 and 0.003 mg/100g. Tannin content was lower (1 and 1.7 %) in the tubers than in Vulgare Pers. sorghum, DeKalb sorghum from U.S. and Kabale sorghum from Uganda. There was a significantly (P < 0.05) lower content of phytate in boiled (3.9ug/g) than in sun-dried tubers (6.0 ug/g). Phytate content of the tubers was lower than that of cooked maize flour, unrefined maize flour, cassava and sweet potato. Trypsin inhibitor activity in the tubers was reduced from 463 to 55 TIU/g tuber and chymotrypsin inhibitor activity was reduced to 50 from 267 CIU/g tuber by cooking. Nyika is a good source of iron and quality protein limiting only in lysine. Protein is comparable to staple maize and higher than root crops consumed in Malawi. It is not a good source of fat and ascorbic acid. Tannin, phytate, trypsin, and chymotrypsin inhibitor content lower than most food crops consumed in Malawi. / Ph. D.

The effect of inclusion rate of pea meal (PM), pea protein concentrate (PPC), soybean meal (SBM), soy protein concentrate (SPC), canola meal (CM) and canola protein concentrate (CPC) in salmonid diets was determined through six corresponding meta-analyses of all data available in the literature for these six feed ingredients, which was followed by weighted regression analysis. Increasing dietary inclusion levels of SBM, SPC, CM and CPC reduced specific growth rate (SGR). Regression analysis determined all of these relationships to be linear declines in SGR (P < 0.05). Inclusion levels of PM or PPC did not influence salmonid SGR (P > 0.05). These results showed that the influence plant proteins have on salmonid SGR is dependent on ingredient type and inclusion level. PM, PPC, SBM, SPC, CM an aqueous-extracted CPC and a high phytate CPC (PCPC) were analyzed for chemical nutrient (proximate, amino acid and phosphorus analysis) and antinutrient composition and total tract digestibility (two separate digestibility trials) in rainbow trout (Oncorhynchus mykiss). The digestibility of proximate chemical components and amino acids were significantly higher for the soy products than the pea products. These digestibilities were also significantly higher in protein concentrates than in plant meals. Dry matter and gross energy digestibility was higher in CPC than in CM (P < 0.05). Phosphorus digestibility was higher in CPC-fed fish than in CM-fed fish (P < 0.05), which is likely due to the fact that CPC did not contain phytic acid. Six consecutive growth studies (one trial per test ingredient) were conducted over a 361-day period to determine the effects of feeding increasing inclusion rates of PM, PPC, SBM, SPC, CM and CPC on the growth performance of rainbow trout. Diets were formulated based on the digestible nutrient content of all ingredients as determined in the previously conducted digestibility trials, to contain 0, 75, 150, 225 or 300 g/kg of each test ingredient. All diets were nutritionally equal and contained 17.6 MJ/kg digestible energy, 386.2 g/kg digestible crude protein and were balanced for digestible essential amino acids to meet or exceed the requirements of rainbow trout. Linear and quadratic analysis was conducted on the experimental data. There were no significant regressions resulting from feeding PM, SPC or CPC at 0-300 g/kg for average daily feed intake (ADFI), specific growth rate (SGR), feed conversion ratio (FCR) or protein efficiency ratio (PER). A positive relationship was associated between PPC inclusion and ADFI (P < 0.05). There was a significantly negative quadratic equation associated with the inclusion level of SBM on SGR and FCR and significantly negative linear and quadratic equations for PER. There were significantly negative linear relationships between the inclusion rate of CM and the SGR, FCR and PER of rainbow trout (P < 0.05). Growth trial results suggest at inclusion levels up to 300 g/kg, PM, PPC, SPC and CPC are feasible plant-based fish meal replacements with predictable growth effects, provided the nutritional constraints set in this experiment are followed. The results of these growth experiments were further analyzed using structural equation modeling to determine the relationship between ANF in the six ingredients and ADFI and SGR, which were transformed (tSGR and tADFI, respectively) to enable comparisons between experiments. All possible models between ingredient ANF (starch, phytic acid, glucosinolates, tannins, isoflavones, total NSP, soluble NSP, insoluble NSP and saponins) and ADFI and SGR were calculated. The model with the highest likelihood, as determined by the Akaike Information Criteria0, contained 29 parameters and six degrees of freedom. tADFI positively influenced tSGR. Glucosinolates, saponins, and phytates had a significantly negative impact on tADFI, whereas tannins had a significantly positive impact. The presence of saponins in the diet resulted in a decrease in tSGR. This structural equation model had significant correlations between all ANF, with the exception of phytates and saponins. Future applications of this work will be to develop a nutritional model for optimal inclusion of plant-based feed ingredients in rainbow trout diets, based on their ANF content, which may improve the accuracy of diet formulation and growth prediction.