Good quality commercial wheat flour (Brand: Teer Maida) was collected from local market, and potato and peanut flours were prepared in the laboratory of the Department of Food Technology and Rural Industries under the Faculty of Agricultural Engineering and Technology, Bangladesh Agricultural University, Mymensingh. All other necessary ingredients and chemicals were used from laboratory stocks. The potatoes were cleaned and washed using potable water and peeled. They were cut into 3 mm slices, blanched, dried and used for further processing. The peanuts were dried in cabinet dryer, peeled, grinded and used for further processing. Chemical analysis of potato, peanut and wheat flour The potato, peanut and wheat flour were analyzed for moisture content, protein, ash, crude fat, crude fiber and total carbohydrate. All the determinations were done in triplicate and the results were expressed as the average value. Moisture content, protein content, ash content, fat content and crude fiber were determined adopting AOAC (2005) method. The carbohydrate content was determined by subtracting the measured moisture, ash, protein and fat content from 100 (Pearson, 1970). Formulation and preparation of composite cake The basic formulation for plain cake (multi-stage mixing) and composite flour cakes are outlined in Table 1 (Atkins, 1971; Anon, 2000). The replacements of wheat flour in the formulations were made with different levels of composite flour. Cakes were prepared by replacing wheat flour with different levels of composite flour in the basic formulation of cake (Table 1) as per the methods of Rajchel et al., (1975). The wheat, potato and peanut flours, and other ingredients for each cake sample were weighed. The flours were mixed together according to the percentages. Then sugar and shortening were mixed in a mixing machine for 20 minutes to produce a cream. At later stages, one egg and other ingredients, and, finally, the mixed flour were added and mixed well using a mixer at low speed (145 rpm) for 10 minutes to ensure even distribution of the components. The bowl was scrapped and stirred for an additional two minutes at medium speed (250 rpm). Then another egg was added, and the mixture was stirred at low speed for two minutes. After the bowl was scrapped, the mixture was stirred an additional two minutes at medium speed. Then a portion of batter was scaled into pre-greased cake pan. All cakes were baked in air oven for 45 minutes at 170°C. The schematic diagram for preparation of composite cake is shown in Fig. 3. Objective analysis of cakes The cake samples S0 (control cake), S1 (70% wheat flour, 5% potato flour& 25% peanut flour), S2 (70% wheat flour, 15% potato flour& 15% peanut flour) and S3 (70% wheat flour, 20% potato flour& 10% peanut flour) were analyzed for moisture content, protein, ash, crude fat and crude fiber as per the methods of AOAC (2005) and total carbohydrate content as per Pearson (1970). Cake volume was initially used as an important parameter of cake quality. It was determined by seed displacement method (Ott, 1987). Moisture content was determined according to the methods outlined in AOAC (2005). The weights and specific volumes of baked cakes were also measured. Subjective (sensory) evaluation of cakes The symmetry and characteristics of crust and crumb of the cakes supplemented with wheat, potato and peanut flour samples were evaluated and recorded. Cakes were evaluated organoleptically for color, flavor, texture and overall acceptability. A 1-9-point hedonic rating test was also performed to assess the degree of acceptability of cakes containing different levels of potato and peanut flour. One slice from each lot of cake was presented to 10 panelists as randomly coded samples. The taste panelists were asked to rate the sample for color, flavor, texture and overall acceptability on a 1-9-point scale where, 1 = dislike extremely, 2 = dislike very much, 3 = dislike moderately, 4 = dislike slightly, 5 = neither like nor dislike, 6 = like slightly, 7 = like moderately, 8 = like very much, 9 = like extremely. The results were evaluated by analysis of variance and Duncan’s New Multiple Range Test procedures of the Statistical Analysis system (SAS, 1985).