African Journal of Food Science Research Vol. 1 (1), pp. 001-009, September, 2013. Available online at www.internationalscholarsjournals.org © International Scholars Journals Full Length Research Paper Optimization of process conditions for cassava (Manihot esculenta) lafun production Nwabueze, T. U.* and Odunsi, F. O. Department of Food Science and Technology, Michael Okpara University of Agriculture, Umudike, P.M.B. 7267, Umuahia, Abia State, Nigeria. Accepted 28 th September, 2013 Optimization of cassava (Manihot esculenta, NR8082 clone) lafun production and evaluation was carried out using response surface methodology. The selected experimental design had three variables and five levels referred to as central composite design. Process variables were inoculum volume, fermentation time and drying temperature. Inoculum volumes were 1.50, 2.00, 2.50, 3.00 and 3.50 l. Fermentation times were 24, 36, 48, 60 and 72 h while drying temperatures were 35, 41, 47, 53 and 59 o C. Response surface data on lafun yields, colour index and cyanogenic potentials were analyzed in a regression model while three- dimensional response surface plots were made. Effects of process variables on lafun parameters varied depending on which response was analyzed. Maximum and minimum yields were 82.68 and 66.68% for lafun flour and 68.24 and 56.20% for lafun starch, respectively. Photometric colour index ranged from 10.03 to 30.51 while cyanogenic potentials of the flour ranged from 1.88 to 5.31 mg HCN/kg and that of the paste from 0.93 to 3.11 mg HCN/kg. Models developed by response surface analysis for these responses were significant at low (1 to 5%) levels of probabilities and could be explained by regression coefficients in the models. Variabilities explained by the models were low for lafun flour yield (R 2 = 0.3188) but high for photometric colour index (R 2 = 0.8098). Optimum process conditions obtained for starch yield, photometric colour index and cyanide reduction were 2.0 l, 36 h and 41 o C; 2.50 l, 48 h and 35 o C; and 2.50 l, 72 h and 41 o C, respectively. Key words: Cassava, lafun, fufu, hydrogen cyanide, optimization, response surface methodology. INTRODUCTION Cassava (Manihot esculenta) is an important food in the tropical areas of African, Asia and Latin America. It is estimated (IITA, 1990) that the crop provides about 40% of all the calories consumed in Africa and ranks second only to cereal grains as chief source of energy in Nigerian diet (Ngoddy, 1989). By this, cassava plays important role in alleviating African Food Crisis though poor in protein (1.20%) and rich in cyanide (> 10 mg/100g fresh weight) in some varieties such as TMS 50395 (IITA, 1990; Janssens, 2001). The variety/clone NR8082 is low cyanide high-yielding cassava developed by the National Root Crops Research Institute, Nigeria and commonly distributed to Nigerian farmers. *Corresponding author. E-mail : teeubueze@yahoo.co.uk. Raw cassava has been reported (Onabolu, 1989; Cereda and Mattos, 1996) to contain two cyanogenic glycosides known as linamarin and lotaustralin with the former being the most representative glucoside, accounting for about 80% of the total cassava glucoside (Dicar, 1993) . Linamarin and lotaustralin are -glucosides of acetone cyanohydrin and ethyl-methyl- ketone- cyanohydrin, respectively (Cereda and Mattos, 1996). Linamarin produces the toxic compound (hydrogen cyanide, HCN), which can be hazardous to the consumer. Cassava processing by fermentation is one of the most widespread techniques used in Africa and is considered an efficient means of reducing cyanogenic potential in the resulting food (Brainbridge, 1994). Lafun is fermented cassava flour popular among the people in the Southwestern States of Nigeria (Cereda and Mattos, 1996) . The traditional method of processing cassava into lafun though unique for its ability to reduce Nwabueze and Odunsi 002 the toxic cyanogenic compound to a least possible level (Numfor, 1983) imparts a strong smell to the product (Cereda and Mattos, 1996). Lafun, like gari, another fermented product from cassava, is a dry product, which can be preserved for a long time under the prevailing local environment. It is a cheap and popular meal in the rural areas of Nigeria and can be prepared into ready food much more quickly than gari (Latunde-Dada, 1997). Lafun is generally prepared into a thick paste in boiling water and eaten with vegetable soup or stew. Oyewole and Ogundele (2001) reported that lafun quality varies with processing methods and with rocess- sors. In spite of various reports in literature (Oyewole and Odunfa, 1990; Oyewole and Afolami, 2000; Oyewole and Ogundele, 2001) on lafun production from different cas- sava varieties, there is scarcity of information on efforts to optimize known process conditions for maximum product quality in mind. The need to optimize process conditions to enhance product quality in terms of yield, cyanogenic potentials and sensory acceptability through response surface methodology, becomes relevant given the pres- ent need for small scale food processing and spread of consumers. In many applications, the system to be optimized can be formulated as a mathematical model. With the advent of high-speed computers, very large and complex sys- tems can be modeled and optimization can yield substan- tially improved benefits. Although optimization has beco- me popular in many sectors of the Food Industry, it is a procedure for developing the best possible product in its class (Ruguo, 1999). The intention of optimization is to provide a more precise map of the path that has the highest probability of leading to a successful food product (Ruguo, 1999; Crapiste, 2000). The objective of this stu- dy is to use the response surface application to optimize process variables to maximize cassava lafun yields, col- our and sensory attributes or reduce cyanogenic poten- tials without large changes in the operational parameters. These process variable conditions investigated included inoculum volume (l), fermentation time (h) and flour dry- ing temperature (C) . The process variable combination that gave maximum cassava lafun response is recom- mended for scale up production in that response. MATERIALS AND METHODS Source of cassava roots Cassava roots of the NR8082 clone were obtained from the Nation- al Root Crops Research Institute, Umudike, Abia State, Nigeria. The Cassava roots were freshly harvested at 10 – 12 months of age. Cassava roots preparation Freshly harvested cassava roots of the NR8082 clone were peeled, washed and fermented according to the method described in litera- ture (Oyewole, 1994; Oyewole and Odunfa, 1990) . The cassava roots were then grated with a local grating machine and fermented with appropriate inoculums from a 4-day-old starter culture containing about 10 cfu/g of microorganisms comprising mainly Corynebacterum manihot and Geotricum candida. Processing variables Process variable conditions investigated for optimization in lafun production included inoculum volume (l), fermentation time (h) and flour drying temperature (C). Inoculums The methods described by Mathew (2002) and Numfor (1983) were used with minor modifications. Inoculums from the starter culture were introduced into the cassava pulp at various volumes of 1.5, 2.0, 2.5, 3.0 and 3.5 l (x1 ). The medium was stirred to bring about uniform distribution of the microorganisms. Drying temperature of lafun flour The fermenting liquor was decanted and the pulp processed using a local screw press. The dewatered pulp was dried in the oven (Carbolite MD 1430, England) at variable temperatures of 35, 41, 47, 53 and 59C (x2). Fermentation time of the cassava pulp The cassava pulp was allowed to ferment at room temperature for variable time intervals of 24, 36, 38, 60 and 72 h (x3). Lafun flour yield The percent yield of lafun flour was calculated from the weight of lafun obtained from known weight of the peeled cassava roots used (Oyewole and Ogundele, 2001). Lafun starch yield A 50 g sample of lafun flour was thoroughly mixed with sufficient water and filtered through a 50 sieve. The mixture was allowed to stand overnight (Oyewole and Afolami, 2001). It was decanted and recovered starch was dried in oven to a constant weight. Weight of dried starch Starch (%) = x 100 (1) Weight of sample Photometric colour index Photometric colour of various lafun flours was determined on 1 g sample according to the method described by Pike (2003). The sample was weighed and dissolved in 20 ml water/ethanol mixture. The mixture was filtered after standing for 30min. The absorbance of the filterate was measured at 400, 550, 620 and 670 nm using 003 Afr. J. Food Sci. Res. Table 1. Central composite experimental design of 3 variable five level process conditions for production of lafun flour from cassava root. Exptal a Run Inoculum volume (I) Fermentation time (h) Drying temperature(C) 1 2.0 (-1) 36 (-1) 41 (-1) 2 2.0 (-1) 36 (-1) 53 (+1) 3 2.0 (-1) 60 (+1) 41 (-1) 4 2.0 (1) 60 (+1) 53 (+1) 5 3.0 (+1) 36 (-1) 41 (-1) 6 3.0 (+1) 36 (-1) 53 (+1) 7 3.0 (+1) 60 (+1) 41 (-1) 8 3.0 (+1) 60 (+1) 53 (+1) 9 3.5 (+a) 36 (-1) 41 (-1) 10 1.5 (-a) 36 (-1) 53 (+1) 11 2.5 (0) 72 (+a) 41 (-1) 12 2.5 (0) 24 (-a) 53 (+1) 13 2.5 (0) 48 (0) 59 (+a) 14 2.5 (0) 48 (0) 35 (-a) 15 2.5 (0) 48 (0) 47 (0) 16 2.5 (0) 48 (0) 47 (0) 17 2.5 (0) 48 (0) 47 (0) 18 2.5 (0) 48 (0) 47 (0) 19 2.5 (0) 48 (0) 47 (0) 20 2.5 (0) 48 (0) 47 (0) 21 2.5 (0) 48 (0) 47 (0) 22 2.5 (0) 48 (0) 47 (0) 23 2.5 (0) 48 (0) 47 (0) Numbers in brackets are the coded values of the independent variables in the experimental design. Exptal = experimental. spectrophotometer (Unican He 105Y, England). The solvent was A small consumer acceptance panel (Stone and Sidel, 1985) used as blank. Photometric colour index was calculated as: familiar with lafun consumption and with the sensory scale and method of assessment were drawn from Southwestern Nigeria pci = 1.29 (A400) + 69.70 (A500) + 41.20 (A620) – 56.41 (A670) (2) students of Michael Okpara University of Agriculture, Umudike and used to evaluate sensory attributes of the cooked paste. where A = absorbance. The samples were arranged randomly in similar plates each with coded 3-digit non-misleading or biasing numbers. A 9-point Hedonic scale was used to make panelists express their feelings of Cyanogenic potentials of lafun flour and paste like and dislike for the flavour, colour, texture, taste and overall acceptability of the pastes. On the scale score, 9 represented like Cyanogenic potentials of lafun flour and paste were determined extremely, 1 represented dislike extremely while 5 represented neither like nor dislike. Data generated from scores were analyzed using picrate paper kits method described by Bradburg (1999). for variance as described by Larmond (1977). One-gram sample of lafun flour was homogenized in a 250 ml conical flask containing 25 ml water. A strip of spot paper soaked in an alkaline sodium picrate solution was fixed in the solution with the Experimental design cork. The flask was kept for 18 h at room temperature. The strip was removed and eluted in 60 ml water and the absorbance was The experimental design of 3-factor factorial in Completely read at 540 nm using a spectrophotometer (Unican He 105Y, Randomized Design (CRD) as reported by Meilgaard et al. (1999) England). as shown in Eq. (3) was adopted. Thus: Sensory evaluation of lafun paste Lafun flour was first sifted through a 50 µ sieve. It was then cooked (3) where Y is the dependent variable, Xi and Xj the independent into a paste by turning the flour in boiled water at flour/water ratio of 1:4 (w/v). Sensory evaluation was carried out within ten minutes of variables in the model, k the number of independent variables, ßo preparation (Oyewole and Afolami, 2001) following the method the intercept (constants and regression coefficients of the model), described by Stone and Sidel (1985) with minor modifications. and the random error term. Nwabueze and Odunsi 004 Table 2. Analysis for quality parameters of lafun flour and starch yields (%), photometric colour index and cyanogenic potentials (mg HCN/kg) for flour and paste a . Exptal b Run Yields (%) Photometric colour index Cyanogenic potentials c Flour Starch Flour Paste 1 82.68 68.24 14.47 3.75 2.17 2 80.00 64.62 14.52 4.06 2.17 3 78.68 56.72 19.74 3.13 1.86 4 66.68 64.32 29.26 3.44 2.17 5 80.26 58.74 14.54 3.75 2.48 6 78.68 64.32 13.80 4.06 2.48 7 80.08 56.20 15.48 3.75 2.48 8 73.32 56.70 17.59 4.06 2.17 9 71.04 58.68 12.04 4.06 1.86 10 67.04 65.70 13.44 4.38 2.48 11 77.32 55.80 10.03 1.88 0.93 12 66.68 57.20 10.08 5.63 3.11 13 77.32 57.80 12.74 5.31 2.80 14 66.68 64.66 30.51 4.38 2.48 15 73.32 60.86 20.83 4.69 2.48 16 77.32 60.72 22.34 4.69 2.48 17 66.68 62.72 24.79 4.69 2.17 18 73.32 60.24 29.59 5.00 2.48 19 66.68 62.30 26.62 5.00 2.48 20 78.68 60.30 21.06 4.69 2.17 21 78.68 61.28 18.41 4.69 2.48 22 77.32 60.74 21.10 5.00 2.48 23 77.32 60.24 20.46 4.69 2.17 a lafun flour cooked into paste by stirring in boiled water ready for consumption, b Experimental, c Cyanogenic potentials (mg HCN/kg). The three variables; inoculum volume (l), fermentation time (h) and RESULTS AND DISCUSSION flour drying temperature (C) and five levels coded -a, -1, 0, +1 and +a gave 15 variable combinations which when replicated 8 times at Response surface analysis for lafun flour the center point (0) generated a total of 23 experimental runs. The experimental design had upper (+a), intermediate (0) and lower (-a) Data on quality parameters of cassava lafun are presen- values of process variable conditions (Table 1). ted in Table 2. Optimum process condition for maximum yield (82.68%) was 2.0 l, 36 h and 41C. (inoculum Statistical analysis volume, fermentation time and drying temperature, res- pectively). Lower or upper process conditions reduced Data collected were analyzed using an appropriate Statistics flour yields. The independent variable with most signi- package (SPSS/PC +) in a regression model. Three-dimensional ficant (p 0.1) effect on flour yield was the length of response surface plots were made using a Statgraphic computer fermentation. Short fermentation times impair root tissue package (Statistica, Statsoft, Inc., Tulsa, OK). softening while long fermentation times as in the tradi- tional practice favour it and result in increased flour yield. Tissue softening is a product of associative interaction of (4) several microbes and enzymes such as polygalactu- ronase, pectinase and cellulase with tissue degrading Where activities (Meilgard et al., 1999; Okolie and Ugochukwu, 1988). Y = dependent variable, 0 …. 3 =estimated regression The polynomial equation (Eq.5) shows that only cross coefficient, X1….X3 = independent variables, and = random error. product effects of process variables had significant (p 0.05) effect on lafun flour yield. Essentially, length of 005 Afr. J. Food Sci. Res. Figure 1. Effect of process conditions on lafun starch yield (IN = inoculum volume (l) and DT = drying temperature ( o C)). Table 3. Estimated regression coefficients for starch content of lafun four. Source Coeficient Std error df p–value Regression on Constant 33.576623 84.594281 X1 -258.193308 54.030059 1 0.004 X2 5.933780 3.073175 1 0.0756 X1.X1 10.087864 4.386095 1 0.0387 X2.X2 -0.102818 0.021418 1 0.0003 X3.X3 -0.042018 0.009088 1 0.0005 X1.X2 4.105445 1.210963 1 0.0048 X1.X3 4.553212 0.865245 1 0.0002 X2.X3 0.060312 0.054932 1 0.2921 R 2 0.85464 fermentation was central in this effect on lafun flour yield. Fermentation time (x 3) had a cross product order effect when interacted with either the inoculum volume (x1) or the drying temperature (x2). Inoculum volume increased microbial load in the medium and hence favoured fer- mentation while drying temperature limited their activities. Since flour yield was relative to the intact cassava root, loss of moisture and volatile components of fermentation (Bokanga, 1992) could affect the actual values obtained for lafun flour in Table 2: Yflour = 30.14444–0.52332x1.x3–0.03695x2.x3 (5) The model for this parameter explained only 31.89% of the total variations in flour yield showing a significant (p 0.05) lack of fit. Response surface analysis for lafun starch Lafun flour was essentially of starch material with maximum yield of 68.24% at optimum process condition (Table 2) . Response surface plot of lafun starch yield had a twisted shape (Figure 1) and linear, quadratic and cross product order effects were significant at 1 – 5% levels of probability. The plot showed that starch yield increased with inoculum volume and drying temperature effect up to a maximum of 2.0 l and 41C, respectively, before a twist effect of these variables occurred at an increased fermentation time. At longer fermentation times, disintegrated tissue pulps have the chances of leaching into the fermenting water, there by reducing starch yield. Corynebacterum manihot population in the fermenting medium which initiates cassava starch break down could have bond breaking effect thereby lowering water holding capacity of the pulp. This activity invariably culminates in loss in moisture and consequently starch as a soluble cassava constituent. The response surface equation developed from the reg- ression Table (Table 3) after removal of non significant (p>0.05) terms becomes: Ystarch = 33.57662 – 258.19331x1 + 10.08764x1 2 – 0.10282x2 2 – 0.04202x3 2 + 4.10545x1.x2 + 4.55321x1.x3 (6) The model accounted for 85.46% of the total variation in lafun starch yield and significantly (p 0.05) fitted. Response surface analysis of lafun colour index All lafun flour samples were visually white in colour. However the photometric determinations presented in Table 2 showed that flour whiteness reading varied from 10.03 to 30.51. The highest index was obtained at pro- cess variable condition of 2.5 l (inoculum volume), 48 h (fermentation time) and 35C (drying temperature). Cas- sava pulps fermented with higher or lower ( 2.5 l ) inoculum volumes and for longer or shorter ( 48 h ) periods, produced lafun flours with lower photometric col- our indices particularly when dried at higher temperature range of 47 to 59C (Figure 2). The Figure confirms that white flour is produceable with >1.5l (inoculum volume) and <47C (drying temperature). Lafun flour from 72 h fermented pulp had the least colour index (10.03) indica- ting that long fermentation times could encourage micro- bial breakdown of colour pigments as well as increase availability of reactive groups which can go on to take part in reactions in low molecular weight carbohydrates during drying. This effect was stronger at higher drying temperatures (47-59C) where mild reactions such as caramelization of carbohydrates and Millard browning involving amino acid compounds of natural proteins and Nwabueze and Odunsi 006 Table 4. Estimated regression coefficients for photometric colour index of lafun flour. Source Coeficient Std error df p–value Regression on Constant 63.100223 80.738146 X1 -82.826304 51.567161 1 0.1322 X2 -5.308687 2.933088 1 0.0935 X1.X1 -6.880421 4.186160 1 0.1242 X2.X2 -0.008258 0.020442 1 0.6928 X3.X3 -0.050130 0.008674 1 0.0001 X1.X2 4.207685 1.155763 1 0.0030 X1.X3 3.866535 0.825804 1 0.0004 X2.X3 0.214220 0.052428 1 0.0013 R2 0.80980 Figure 2. Effect of process conditions on photometric colour index (PCI) of lafun flour. IN = inoculum volume (l) and DT = drying temperature ( o C) reducing sugars affect flour whiteness. Badrie and Mello- wes (1992) who considered effect of cassava starch or amylose on characteristics of cassava extrudate, repor- ted the potentials for interactions between lipids, proteins and carbohydrates and their breakdown products, which under the present processing conditions could affect flour colour. The response surface analysis of data on the photo- metric colour index (Table 4) showed that the response depended more significantly (p 0.05) on cross product order effects of independent variable interactions than on linear order effects of individual variables. Fermentation time had significant (p 0.05) quadratic and cross product effect on the flour colour. The interac- tive cross product effects of inoculum volume with drying temperature was most significant (p 0.05) (Eq. 7) follo- wed by that with fermentation time, while drying tempera- ture and fermentation time was the least. On removal of non significant (p > 0.05) terms and recomputing, the polynomial becomes: Ycolour = – 63.10022 – 0.05013x3 2 + 4.20769x1.x2 + 3.86654x1.x3 + 0.21422x2.x3 (7) The model accounted for 80.98% of total variation in lafun colour and significantly (p 0.05) fitted. Response surface analysis of cyanogenic potentials of lafun flour and paste The independent process variable most affecting cyanide reduction was the length of fermentation. Fermentation times of 60 to 72 h reduced cyanogenic potentials from 79.80 mg HCN/kg originally present in the raw cassava root (NR8082 clone) to a range of 3.75 to 1.88 mg HCN/kg in the flour and 2.17 to 0.93 mg HCN/kg in the paste. Shorter fermentation times (24 –48 h) reduced the cyanide potentials from 5.63 to 3.75 mg HCN/kg in flour and from 3.11 to 2.17 mg HCN/kg in the paste (Table 2). Optimum process variable condition of 2.5 l, 72 h and 41C produced lafun flours with least cyanogenic poten- tials in flour (1.88 mg HCN/kg) and paste (0.93 mg HCN/kg). Long fermentation periods gave fermenting microorganisms ample opportunity to produce enzyme linamarase which when combined with retting and drying drastically reduced cyanide content in the flour. Optimum independent process condition of 2.0 l, 36 h and 41C (inoculum volume, fermentation time and drying tempera- ture), reduced cyanogenic potentials to 95.30% in the flour and 97.28% in the paste. Cooking the flour into paste resulted in a loss of 50.53% and 98.83% of cyanide 007 Afr. J. Food Sci. Res. Table 5. Estimated regression coefficients for cyanogenic potentials of lafun paste a . Source Coeficient Std error df p–value Regression on Constant - 3.469850 6.468992 X1 5.184381 4.131722 1 0.2317 X2 -0.043815 0.235008 1 0.8550 X1.X1 -0.868747 0.335408 1 0.0224 X2.X2 0.001506 0.001638 1 0.3747 X3.X3 -0.000705 0.000695 1 0.3291 X1.X2 -0.026021 0.092603 1 0.7831 X1.X3 0.019166 0.066166 1 0.0466 X2.X3 0.000196 0.004201 1 0.0636 R2 0.68678 a lafun flour cooked into paste by stirring in boiled water ready for consumption. Figure 3. Effect of process conditions on cyanogenic potential (COCNP) of lafun paste. IN = inoculum volume (l) and ST = fermentation time (h). present in the flour and raw cassava root (NR8082), respectively. The high percentage loss is attributable to its solubility and heat labile nature particularly in the paste. This is one major nutritional benefit of traditional practice of combining long period of fermentation (about 96 h) and twice cooking as unit operations in local cassa- va processing and utilization in Africa. This observation agrees with the report (Brainbridge, 1994) that fermentation is the most widespread techniques used in Africa as effective means of reducing cyanogens in cassava products. Response surface analysis of the cyanogenic potentials of lafun paste (Table 5) indicated that quadratic and cross product coefficients of the model were significant (p 0.05) while the plot showed a dome shape (Figure 3). Reduction in cyanide increased quadratically with inocu- lum volume corresponding to 48 h fermentation time be- fore declining. The cross product order effect of fermenta- tion time and inoculum volume was significant up to 55 h before linear effect of fermentation time took over, being maximum at 72 h. Effect of inoculum volume on cycanide reduction was due to heavy microbial concentration, which favoured fermentation. On removing the non significant (p>0.05) terms and recomputing, the polynomial equation beco- mes: Ypaste =– 3.46985 – 0.8687x1 2 – 0.01917x1.x3 (8) The model contributed 68.68% of the total variation in cyanide content of lafun paste. Sensory evaluation Table 6 shows sensory evaluation of the lafun paste. One significant observation in this study is that all the lafun paste samples were acceptable to the panelists for the sensory attributes evaluated. The samples visually looked alike but for the experience of the traditional lafun consumers. Differences in sensory data were not statistically significant (p >0.05). Panelists had been used to consuming traditionally processed cassava lafun pastes, which are odourous, stickier, and darker in colour than products of this experimentation. They describe lafun pastes with little or no odour, having a characteristic white colour and good texture as good quality lafun (Oyewole and Afolami, 2001). Perhaps loss of volatile products of fermentation during decanting and subsequent cooking operations reduced the offensive smell of the product. Conclusion Optimization of process variables to maximize cassava lafun yields, colour and sensory attributes or reduce cyanogenic potentials without large changes in the operational parameters was achieved in this study. The process variable condition that gave maximum cassava lafun flour (82.68%) and starch (68.24%) yields were 2.0 l, 36 h and 41C (inoculum volume (l), fermentation time (h) and drying temperature (C)). Under these variable condition, cyanogenic potentials reduced by 95.30% in the flour and 97.28% in the paste from its 79.80 mg HCN/kg in the raw cassava pulp (NR8082 clone). Since all lafun flours were visually white and the corresponding pastes generally liked by sensory panelists, the optimum Nwabueze and Odunsi 008 Table 6. Sensory evaluation of lafun paste* Run Colour Flavour Texture Taste Overall acceptability 1 7.3±0.25 ab 7.3±0.37 a 7.5±0.44 a 7.6±0.22 a 7.6±0.23 a 2 6.9±0.38 ab 7.00±0.44 a 7.7±0.27 a 7.3±0.32 a 7.1±0.34 a 3 6.7±0.27 a 6.8±0.53 a 6.8±0.38 a 7.2±0.34 a 7.0±0.27 a 4 7.5±0.25 ab 7.3±0.42 a 7.9±0.24 a 7.8±0.22 a 7.7±0.24 a 5 7.0±0.24 ab 7.2±0.27 a 7.5±0.36 a 7.3±0.25 a 7.3±0.20 a 6 7.6±0.23 ab 7.8±0.13 a 7.2±0.23 ab 7.6±0.23 a 7.7±0.23 a 7 7.8±0.21 ab 7.5±0.13 a 7.3±0.46 a 7.7±0.23 a 7.9±0.23 a 8 7.3±0.31 ab 7.7±0.23 a 7.5±0.26 a 7.2±0.23 a 7.8±0.20 a 9 7.4±0.33 ab 7.3±0.31 a 7.7±0.28 a 7.6±0.30 a 7.4±0.28 a 10 7.0±0.30 ab 7.8±0.40 a 7.6±0.23 a 7.9±0.19 a 7.4±0.20 a 11 7.6±0.23 ab 7.7±0.22 a 7.7 ±0.31 a 7.7±0.31 a 7.7±0.24 a 12 7.8±0.21 ab 7.4±0.31 a 7.5±0.43 a 7.6±0.35 a 7.5±0.25 a 13 7.4±0.22 ab 7.6±0.20 a 7.1±0.51 a 7.7±0.28 a 7.3±0.33 a 14 7.3±0.32 ab 7.3±0.39 a 7.0±0.45 a 7.3±0.37 a 7.6±0.31 a 15 7.1±0.29 ab 7.4±0.32 a 7.5±0.41 a 7.4±0.28 a 7.4±0.23 a 16 7.6±0.16 ab 7.3±0.28 a 7.5±0.38 a 7.2±0.20 a 7.5±0.17 a 17 6.6±0.36 ab 7.2±0.33 a 7.6±0.23 a 7.4±0.43 a 7.6±0.32 a 18 7.3±0.26 ab 7.3±0.28 a 7.9±0.30 a 7.1±0.35 a 7.1±0.29 a 19 7.4±0.16 ab 7.3±0.36 a 7.4±0.32 a 7.2±0.26 a 7.3±0.34 a 20 7.3±0.10 ab 7.3±0.23 a 7.6±0.24 a 7.3±0.34 a 7.3±0.24 a 21 6.7±0.38 ab 6.7±0.32 a 7.1±0.32 a 7.0±0.36 a 6.9±0.42 a 22 7.6±0.19 ab 7.6±0.27 a 7.0±0.44 a 7.9±0.32a 7.8±0.20 a 23 7.9±0.17 a 7.0±0.16 a 7.9±0.45 a 7.7 ±0.32a 7.2±0.20 a * lafun flour cooked into paste by stirring in boiled water ready for consumption. process condition producing highest lafun yields in terms of flour and starch with low flour and paste cyanide concentration (2.0 l, 36 h and 41C), is recommended for scale up production when using cassava NR8082 clone, on the basis of yield, safety and acceptability. REFERENCES Badrie N, Mellowes WA (1992). Cassava starch or amylase effects on characteristics of cassava (Manihot esculenta Crantz) extrudate. J. Food Sci. 57: 103 –107. Bokanga M (1992). Mechanisms of the elimination of cyanogens from cassava during traditional processing. In: Traditional African Foods – Quality and Nutrition. A Westby, PJA Reilly (eds.) International Foundation for Science, Stockolm, Sweden. Bradburg GM, Egan SV, Bradburg JH (1999). Determination of all forms of cyanogens in cassava roots and cassava products using picrate paper kits. J. Sci. Food Agric. 79: 593 –601. Brainbridge Z (1994). Case studies in cassava processing and cyanogens elimination. Proc. fifth symposium. ISTRC–AB, 343 –347. Cereda MP, Mattos MCY (1996). Linamarin: The toxic compound of cassava. J. Venom. Anim. Toxins 2: 6 –12. Crapiste GH (2000). Simulation of drying rates and quality changes during the dehydration of foodstuffs. In: Trends in Food Engineering. JE Lozano, C Anon, E Parada-Arias, GV Barbosa-Canovas eds Technomic Publishing Co, INC Pennsyvania, USA. Dicar HA (1993). The indigenous Foods of the Sudan: A study in African Foods and Nutrition. Center for Agriculture and Biosciences International (CBS) publication, Walling Ford, UK. IITA (1990). Cassava in Tropical Agric: A Reference Manual. IITA, Ibadan, Nig. pp. 87 –120. Janssens M (2001). Cassava (Manihot esulenta Crantz). In: Crop production in Tropical Africa. RH Raemaekers ed. DGIC, Brussels, Belgium. Larmond E (1977). Laboratory method of sensory evaluation of Foods. Department of Agriculture, Canada Publications, 1637:33. Latunde–Dada GO (1997). Fermented foods and cottage industries in Nigeria. Food Nutr. Bull. 18: 4. Mathew G (2002).Cell separation and softening of cassava after fermentation with a mixed culture inoculum. Proc 12 th symp. ISTRC – Japan, pp. 104 –105. Meilgard M, Civille GV, Carr BT (1999). Sensory evaluation techniques. 3 rd edition. CRC. Press, Boca Raton, Florida. Ngoddy PO (1989). Processing and storage of cassava. Proceedings of a course held at University of Nigeria Nsukka, 10 –23 April. 1988. Dotam Publications Ltd. pp. 51-57. Numfor FA (1983). An improved technique of processing cassava fufu. Proceedings of 2 nd Symposium. ISTRC –AB, 11 –113. Okolie PN, Ugochukwu EN (1988). Changes in activities of cell wall degrading enzymes during fermentation of cassava (Manihot escu- lenta Crantz) with Citrobacter freundii. J. Sci. Food Agric. 44: 51 –61. Onabolu AO (1989). Cassava Utilization and Nutrition. Proceedings of cassava held at the University of Nigeria Nsukka, 10 –13April, 1988. Dotam publication limited. pp. 90 –95. Oyewole OB (1994). Upgrading traditional cassava fermentation through process improvement. Proceedings of 9 th Symposium. ISTRC – AB. pp.261 –266. Oyewole OB, Afolami OA (2001). Quality and preference of different cassava varieties for lafun production. J. Food Technol. Afr. 6: 27–29. Oyewole OB, Odunfa SA (1990). Characterization and distribution of 009 Afr. J. Food Sci. Res. lactic acid bacteria in cassava fermentation for fufu production. J. Appl. Bact. 68: 45 –52. Oyewole OB, Ogundele SL (2001). Effects of length of fermentation on the functional characteristics of fermented cassava fufu. J. Food Technol. Afr. 6: 38-40. Pike OA (2003). Colour in Food Analysis. 3 rd edn. Kluwer Academic Hoenin Publisher, NY. Ruguo HU (1999). Food product design: A computer aided statistical approach. CRC Press, Boca Raton, Florida. Stone H, Sidel JL (1985). Sensory evaluation practices. San Diego. Califonia: Academic Press, Inc. p. 311.