Breakage of rice kernels is a major problem faced by

Transcription

1 VOUMERIC CHANGES IN RICE KERNES DURING DESORPION AND ADSORPION K. Muthukumarappan, V. K. Jindal, S. Gunasekaran SUDEN MEMBER MEMBER MEMBER ASAE ASAE ASAE ABSRAC Experiments were conducted to obtain information on volumetric changes of rough, brown, and milled rice due to changes in moisture content and temperature. Volumetric change of rough, brown, and milled rice was linearly related to changes in moisture content and temperature. For all three forms of rice kernels, the coefticients of cubical and linear hygroscopic expansion were higher during adsorption than during desorption. he coefficient of linear hygroscopic expansion of thickness was higher than that of length and width during desorption and adsorption; and it was higher for brown rice than for rough and milled rice during desorption. However, the coefhcient of cubical thermal expansion was higher for milled rice than for brown and rough rice during desorption. Decrease in head yield ratio was correlated with rate of volumetric expansion of rough rice during adsorption. KEYWORDS. Desorption, Adsorption, emperature, Rice, Hygroscopic expansion. INRODUCION Breakage of rice kernels is a major problem faced by rice processors. Various external and internal stresses formed in rice kernels during processing cause the kernels to fissure resulting in easy breakage (Kunze, 1977; Kunze and Prasad, 1978). For example, moisture and temperature gradients prevalent within the kernel cause expansion and contraction in the grain leading to development of internal stresses. If these stresses can be calculated, better processes can be designed to reduce fissure development and increase head rice yield. he head rice yield is the mass fraction of cleaned rough rice which remains as whole kernels (three-fourths kernel or more) after complete milling. For ease of comparison and computational convenience the head rice yield can be expressed in terms of head yield ratio (HYR). he HYR is defined for adsorption processes as the ratio of the actual head rice yield to the control head rice yield. Steffe and Singh (1980) estimated the shrinkage of white, brown and rough rice by taking volume measurements at 30% and 15% moisture content (MC) using a Article was submitted for publication in April 1991; reviewed and approved for publication by the Food and Process Engineering Inst, of ASAE in January Presented as ASAE Paper No he authors are K. Muthukumarappan, Graduate Research Assistant, Agricultural Engineering Dept., University of Wisconsin, Madison; Vinod K. Jindal, Professor, Division of Agricultural & Food Engineering, AFP, Bangkok, hailand; and Sundaram Gunasekaran, Associate Professor, Agricultural Engineering Dept., University of Wisconsin, Madison. commercial air-comparison pycnometer. hey determined that on an average the volume of each of the three rice forms decreased 12.3% with the 15% drop in MC. Murthy et al. (1986) investigated the increase in rough rice kernel volume during adsorption from 13.6 to 29.9% for five rough rice varieties. he increase in volume was linearly related to MC for two varieties and nonlinearly for the other three varieties. Banaszek and Siebenmorgen (1990 a,b) recently developed thin layer moisture adsorption equations for rough rice and correlated the head rice yield reduction rates to moisture adsorption. o better understand internal stresses, the information on basic properties like coefficients of linear and cubical hygroscopic expansion and cubical thermal expansion of rice kernels are needed. In the past there have been many studies (Arora et al., 1973; Prasad et al., 1975; Yamaguchi et al., 1985) to determine these properties during drying. However, information on these properties during adsorption is lacking. his article presents the results of an experimental investigation of these properties during moisture desorption and adsorption for rough, brown and milled rice kernels. OBJECIVES he objectives of this study were to: 1. Determine the coefficients of linear and cubical hygroscopic expansion of rice kernels during desorption and adsorption; 2. Determine the coefficient of cubical thermal expansion of rice kemels and its dependence on moisture content; 3. Develop generalized models for estimating volumetric changes in rice kernels due to changes in moisture content and temperature; and 4. Correlate head yield ratio to hygroscopic expansion of rough rice. MAERIAS AND MEHODS SAMPE PREPARAION Rough rice used in the experiments was a long-grain variety (RD-29) that was hand harvested in the last week of January 1988 from Klong uang district in hailand. he grains were manually separated from the stalks and sealed in 10 kg lots in double polyethylene bags and stored in a refrigerator at 5 C until used. he average MC of the cleaned rough rice was 33.1% (unless stated otherwise all moisture contents are on a dry basis). Each MC was determined by oven-drying at 130*^ C for 13.5 h as recommended by Jindal and Siebenmorgen (1987). VO. 35(1): january-ftbruary American Society of Agricultural Engineers / 92 /

2 Experiments were performed using samples subjected to both desorption (drying) and adsorption (rewetting). For the desorption experiments, high-moisture rough rice was exposed to room conditions maintained at 28^ C and 65% relative humidity (RH). During the desorption process, samples of about 3 kg each were periodically collected. Moisture content of these samples were 33.1, 27.9, 23.3, 21.0, and 17.4%, respectively. Additional samples of 15.0, 12.3 and 8.5% MC were obtained through artificial drying in a controlled air-oven (Hsu Hui Chemicals Co.) set at 35^Cand50%RH. For the adsorption experiments, samples with initial MC, of 8.5, 15.0, and 17.4%, each of about 15 kg, were rewetted in a controlled environment chamber maintained at 30** C and 95% RH. Final MC obtained by this rewetting process was 25.0%. Rewetted samples of MC higher than 25.0% were obtained by applying additional water as fine spray to the samples in the chamber. Although samples with an initial MC of 8.5% might have fissured during adsorption at 30"^ C and 95% RH, the volume change due to fissure formation is negligible. Moreover, the fissure does not add to absolute volume. he process of rewetting samples of 8.5, 15.0, and 17.4% initial MCs were labelled as Adsorption 1, Adsorption 2, and Adsorption 3, respectively. Samples were removed from the chamber periodically. he intermediate MCs attained were 11.0, 12.1, 15.6, 17.3, 21.9, 24.9, 28.0, and 33.3% for Adsorption 1; 17.7, 20.4, 23.3, 25.0, and 30.9% for Adsorption 2; and 20.1, 23.7, 27.4, and 28.8% for Adsorption 3. All the samples obtained through desorption and adsorption processes were sealed in double polyethylene bags. he samples with MC less than 16% were stored at room conditions and the other samples were stored in a refilgerator at 5 C until testing. Each sample, of different MC obtained fi-om desorption and adsorption processes, was divided into three lots. One lot was returned and kept as rough rice in storage and the other two lots were used to get brown and milled rice. Brown rice was obtained by using a Satake rubber roll laboratory huuer passing 125 g of rough rice twice through the rotating rolls for complete removal of the husk. For samples above 20% MC, clearance between the rolls was increased to minimize damage to the bran layer. Milled rice was obtained by whitening the brown rice using a horizontal abrasive whitener. Most kernels in the samples labelled Adsorption 1 were broken because of fissuring during adsorption. A larger sample size was used to get milled rice fi-om samples above 20% MC. A Satake-type rice grader was used to separate whole grains from the brokens and only whole kernels were used for this study. he brown and milled rice samples were stored until testing as explained before. HYGROSCOPIC AND HERMA EXPANSION EXPERIMENS Coefficients of cubical hygroscopic expansion (a), linear hygroscopic expansion (P), and cubical thermal expansion (y) are defined as: a = (l/v)(av/am) p = (l/x)(ax/am) Y=(1/V)(AV/A) V X AM A AV Ax volume of a kernel (m^), characteristic dimension (length, width or thickness) (m), increase in MC (%), increase in temperature ( C), increase in volume (m^), and increase in characteristic dimensions (m). Cubical Hygroscopic Expansion. Before testing, MC of the samples after storage was determined using the oven method. here was no loss of moisture in storage. All the samples were removed from storage and held overnight in a controlled environment room (28 C and 55% RH). During the holding period the samples were kept in moisture proof containers to avoid moisture loss. Sample sizes of 40, 50, and 50 g were used for rough, brown, and milled rice, respectively. All experiments were conducted in the controlled environment room. Volumes of all the rice samples were measured in triplicate using an aircomparison pycnometer as described by Mohsenin (1980). he air-comparison pycnometer was constructed (fig. 1) similar to the one described by Reyes, Jr. (1983). he pycnometer consisted of two tanks of different volumes with the reference tank slightly bigger than the sample tank. A U-tube manometer with a resolution of 0.5 mm H2O was used to measure the pressure readings. With the material in the sample tank, control valve 2 was closed and air was supplied to the reference tank. When the desired reference tank pressure was reached, as indicated by the manometer, control valve 1 was closed and the pressure in the reference tank was allowed to stabilize and Pj was read. hen control valve 3 was closed and control valve 2 was opened allowing the air pressure to equilibrate. his pressure was recorded as P2. he sample volume was calculated using the ideal gas equation as shown below: rearrangmg P.V,=P2V,+P2(V2-V,) Vs=(V,+V2)-{P,/P2)V, V = volume of reference tank plus that of control valves 1 and 2, and the connecting tube (m^). cv, Diaphragm Dump " Reference tank rf\ \ ^ CV, \~ Manomotor (1) (2) CV,, CV,, CV, - control valves Figure 1-Schematic of the air comparison pycnometer. ^ CV, Sample tank =1 236 RANSACIGNS OF HE ASAE

3 V2= volume of sample tank plus that of control valves 2 and 3, and the connecting tube (m^), and Vs= volume of sample (m^). he pycnometer was calibrated using water as the medium. A linear fit between the pressure ratio (P1/P2) and the sample volume was obtained with a correlation coefficient, r = Maximum difference between calculated and measured volumes was 1.37% with r = During the calibration, it was noticed that the initial pressure in the reference tank had an influence on the volume measurement. So the initial pressure was kept constant at m H2O. he accuracy of volume measurements using the pycnometer was determined by measuring the volume of commercial bearing steel balls 7.9 mm in diameter. he measured volumes of the steel balls were in close agreement with the actual (manufacturer listed) volume. On an average the pycnometer yielded 1.41% higher (r = 0.998; standard error of estimate = 0.494) volume in comparison with the actual volume. inear Hygroscopic Expansion. Characteristic dimensions (length, width and thickness) of rice kernels were measured from their projected image, 30 times larger than actual size, using an overhead projector (Reyes, Jr., 1983). ength, width, and thickness refer, respectively, to the longest, the intermediate, and the shortest dimensions of the rice kernel. he magnification factor was obtained by comparing the actual and projected distance between markings on a transparent straight-edge. Fifty kernels were randomly selected from the rough and brown rice samples of different MCs labelled Desorption and Adsorption 1 and their characteristic dimensions were measured. Since most kernels in the milled rice samples labelled Adsorption 1 were broken, the characteristic dimensions of milled rice were measured during desorption only. Average values of the length, width and thickness were considered for data analyses. Cubical hermal Expansion. he coefficient of cubical thermal expansion of rice kernels was determined according to the Standard ASM est D (ASM, 1984) for plastics. For this test, six pyrex glass dilatometers were constructed and used. he dilatometers were about 65 mm long and 10 mm in internal diameter with a wall thickness of 1.0 mm and with a removable stopper having 1.0 mm diameter capillary pore (fig. 2). Before the experiments, all the dilatometers were calibrated using mercury as the medium. he ground glass joint between bulb and stopper was sufficiently tight to prevent any leakage of air or mercury. he dilatometers were filled with mercury, sealed and placed in a controlled temperature water bath. he water batii was heated from 28 to 68 C at a rate of 0.3 C/min. Mercury spilled from the dilatometers was collected at 4 C intervals and weighed. From the mass and density of mercury (Kaye and aby, 1966), volume of the spilled mercury was determined. he calibration procedure was repeated three times for each dilatometer. he coefficient of cubical thermal expansion of mercury was calculated from the relative changes in the volumes of the glass bulb and mercury that resulted during the heating of the dilatometers as follows: ^IOmm ^ SOPPER FIING EVE RICE KERNES PYREX GASS BUB Figure 2-Pyrex glass dilatometer. AV =AV,+AV, (3) AVn^ = change in volume of mercury in glass bulb (ni3), AVj, = change in volume ofglass bulb (m^), and AVg = changein volume of mercury spilled out (m^). Equation 3 can be written as: C V d = C,V,d+AV, (4) rearranging C = Cb(V,/V )+(l/v )(AV,/A) (5) ^m C5 ~ coefficientof cubical thermal expansion of mercury (per C), = coefficient of cubical thermal expansion of glass bulb (per ^^C), l.le-osr C (Michael et al., 1957), AVg/A = slope of the volume spilled out and temperature difference (m^/ C). When the volumes of mercury and glass bulb were assumed the same, equation 5 becomes: C,=Cb+(l/V )(AV,/A) (6) hree undamaged rice kernels were randomly selected from the previously conditioned samples and placed in the dilatometers. After filling the dilatometers with mercury, they were heated as described above. hree replications were made at each MC for all three forms of rice. Since the volume change due to temperature was very small compared to the volume change due to moisture, the cubical thermal expansion experiments were conducted during desorption only. he coefficient of cubical thermal expansion of the rice kernels was determined from the relative changes in the volume of rice kernels, glass bulb, and mercury during heating of the dilatometer. Now equation 3 takes the following form: VO. 35(1): january-ffebruary

4 AVk+AV = AV,+ AV, (7) AVjj is the change in volume of rice kernels (m^). Equation 7 can be expressed as: qv,a+ CjV,- VjA= C,V,A+ AV, (or) ABE 1. Model parameters* for volumetric changet with moisture content^, and coefficient of cubical hygroscopic expansion of rough, brown, and milled rice during desorption and adsorption Moisture change step Rough rice Desorption Adsorption 1 Adsorption 2 Adsorption 3 Adsorption M^% A a (B)xlO^ (a)xlo^ Ck=Vj(C,-Cj/V,+ (AV,/A)(l/Vj+C (8) Cj^ is the coefficient of cubical thermal expansion of rice kernel (per ^^C). In equation 8 the volume of kernels Vj^ was determined from the volume of the glass bulb, V^,, by knowing the volume of mercury in the presence of rice kernels. RESUS AND DISCUSSION CUBICA HYGROSCOPIC EXPANSION inear and quadratic regression models were used to relate the changes in the volumes of rough, brown, and milled rice with MC. It was observed that the change in volume of rough, brown and milled rice with MC conformed to a linear model with r better than for all moisture change steps (Desorption and Adsorption 1, 2, and 3). he quadratic models offered only a very small increase in the r values and, therefore, were not considered. he parameters of the linear model (A and B of V = A + B M^ib) and the coefficient of cubical hygroscopic expansion values calculated based on the volume change predicted by the linear model are presented in able 1. he coefficient of cubical hygroscopic expansion values of rough, brown, and milled rice during adsorption were different for Adsorption 1, 2, and 3. However, the values did not follow any definite trend. So, for ease of comparison with the desorption values, the data from Adsorption 1, 2, and 3 were pooled and linear regressions were redone. For all rice forms, the coefficient of cubical hygroscopic expansion was higher during adsorption than during desorption. he difference between the desorption and adsorption values were much higher for rough rice than for brown and milled rice. he coefficient of cubical hygroscopic expansion values obtained from this experimental investigation compared well with other reported values for rough and brown rice during desorption. For example, Wratten et al. (1969) and Murthy et al. (1986) reported values of 1.489x10-2 and 0.952x10-2 mvm^. he percent MC for rough rice in the range of 14 to 29% MC; and Yamaguchi et al. (1985) and Prasad et al. (1975) reported values of 1.452x10-2 and 1.737x10-2 m^/m^. he percent MC for brown rice in the range of 0 to 43% MC. INEAR HYGROSCOPIC EXPANSION he characteristic dimensions (X) were found to increase linearly with MC. he linear model parameters (A and B of X = A + B M^^) for desorption and adsorption are presented in able 2. Due to variations among the Brown rice Desorption Adsorption 1 Adsorption 2 Adsorption 3 Adsorption Milled rice Desoiption Adsorption 2 Adsorption 3 Adsorption AandBofV = A + BMdb t V,x 10"^ m^/leg dry matter t Mdb»% a,mvm^%mc kernels of the sample, the r values were occasionally low (0.58 to 0.62) and most of them were in the range of about 0.77 to he model parameters A and B represent the slope and intercept of the linear relationship between the characteristic dimensions and MC. he slope gives the change in dimension with change in MC (i.e. B = AX/AM). But the coefficient should be relative to the original dimension, which is the intercept of the fit. So B/A (=(AX/AM)/X) represents the coefficient of linear hygroscopic expansion. Values of the coefficient of linear ABE 2. Model parameters for characteristic dimensions vs. moisture content, and coefhcient of linear hygroscopic expansion of rice samples during desorption and adsorption* Moisture change step(md5,%) Rough rice Desorption ( ) Adsorption ( ) Brown rice I>esorption ( ) Adsorption ( ) Milled rice Desorption ( ) CDt w w W W W (A)xlO^ (B)xlO^ (p)xlo^ * See able 1 for mathematic equivalents. t CD-Characteristic dimension, -ength, W-Width, -hiclcness. p,m/m%mc 238 RANSACnONS OF HE ASAE

5 hygroscopic expansion for desorption and adsorption are also summarized in able 2. In general, the coefficient of linear hygroscopic expansion values of rough, brown, and milled rice were higher during adsorption than during desorption. It can be observed that the coefficient of linear expansion of thickness was the highest followed by those of length and width during desorption and adsorption. his means that the relative expansion per unit of linear expansion is not the same in all directions, which may be one of the possible factors causing cracks in rice kernels. Among the grains tested, brown rice gave higher coefgcients of linear expansion in all three directions than rough and milled rice during desorption indicating that the expansion of the bran was higher than other parts of the rice kernel. he values obtained in this study are very close to other reported values for rough and brown rice during desorption. For example, the coefficient of linear hygroscopic expansion values of length, width, and thickness reported by Morita and Singh (1979) and Wratten et al. (1969) are 1.66x10-3, 2,65xl(^3 and 4.07x10-3 and 1.64x10-3, 2.52x10-3 and 4.83x10-3 m/m, %MC, respectively, for rough rice in the MC range of 12 to 29%. he values reported by Prasad et al. (1975) are 4.39x10-3, 3.7x10-3, and 4.26x10-3 m/m, % MC, respectively, for brown rice in the MC range of 2 to 29%. CUBICA HERMA EXPANSION he rate of expansion of rice kernels with temperature was uniform up to 58 C and different thereafter. his observation was similar to Arora et al. (1973), who found the thermal expansion of milled rice to take place at two uniform rates with a transition at 53 C. Since the thermal expansion of kernels beyond 60 C may not be needed for most practical applications, a single rate for the entire temperature range was considered. here was no change in the kernel mass after the expansion experiment, indicating that there was no moisture transferft-omthe kernel. Among the many models fit to the data, a linear model best related the coefficient of cubical thermal expansion and MC for all rice forms (fig. 3). he cubical thermal expansion of the milled rice was the highest followed by those of the brown and rough rice. COMBINED EFFEC OF MOISURE AND EMPERAURE ON VOUME EXPANSION he coefficient of cubical hygroscopic expansion of rice was determined at 28 C only. herefore, the hygroscopic expansion values at other temperatures were synthesized mathematically assuming (1) a change in volume from one moisture-temperature combination to another is independent of the order in which the process takes place, and (2) the volume of grain (V) is unity at 0% moisture and 0 C. he assumption of unit volume is necessary for developing generalized models to determine relative volume rather than actual volume. A non-linear, least square multivariate secant method (SAS, 1987) was used to establish the following equations. Rough rice: V = E-02 Mdb E.04 (9a) Moisture content, (%, d.b.) Figure 3-Variation of coefficient of cubical thermal expansion of rice Icemels witli moisture content Brown rice: V= E-02 M^j, E-04 (9b) Milled rice: V= E-02 M^ E-04 (9c) Equations 9a, 9b, and 9c may be directly used to determine the relative volume (V, m^/kg dry matter) at different MCs (M^b, %db) and temperatures (, C). Further information on the models developed can be obtained from Muthukumarappan (1988). EFFEC OF VOUME EXPANSION RAE ON HYR he effect of volumetric expansion rate of rough rice on HYR after milling was studied. he results of the present investigation were used along with the results of Dash (1986) who studied the effects of moisture adsorption on HYR using the same rice variety. Further information about HYR and models used can be obtained from Dash (1986). From the experimental investigation the volumetric change of rough rice (V, xlo-^ m-^/kg dry matter) was correlated with dry basis MC (M^b, % db) during moisture adsorption in the following form: V = E-03 M. db (10) Equation 10 correlates the volumetric change of rough rice with dry basis MC. But Dash (1986) expressed his results in terms of wet basis MC. So data used in equation 10 were converted in terms of wet basis MC (M^b, % wb) and the volume expansion of rough rice was correlated with MC as: V = E-03 M^ (11) Dash (1986) proposed the following models to correlate MC with exposure time during moisture adsorption: VO. 35(1): JANOJARY-FfeBRUARY

6 M^= M, + (Mi- M, )exp(-k t«-^) k = exp ( (M./) () (RH) ln(rh)) (12) Mwb = MC of rough rice at exposure time t(%wb), Mj = initial MC of rough rice (%wb), Mg = equilibrium MC of rough rice (%wb), = exposure temperature ( C), t = exposure time (h), and RH = exposure relative humidity (%). o correlate the HYR with the rate of volume expansion (RVE, xlo-3 m^/kg dry matterh), the relationship between RVE and exposure time was needed. Using equations 11 and 12 the RVE of rough rice, when exposed at 30** C and 90% RH for different Mj, was best related to the exposure time by logarithmic models as follow: Mi = 8.0: RVE = 6.40E E-03 log(t) Mi = 10.0: RVE = 4.62E E-03 log(t) Mi = 12.0: RVE = 3.09E E-03 log(t) (13a) (13b) (13c) Dash (1986) also developed a relationship for HYR as a function of exposure time and initial MC. In(HYR) = -A exp (-B( Mj)) (14) A = exp[ ln(t) (l/t)] B = ln(a) (l/t) (l/t) Equations 13a, 13b, 13c, and 14 were combined to correlate the HYR with the rate of volume expansion (RVE, xl0^3 m^/kg dry matterh). he regression lines for different Mj are presented in figure 4. It is known that rough rice with low initial MC adsorbs moisture at a high rate which results in high head rice yield reduction (Banaszek and Siebenmorgen, 1990b). his is due to the high volumetric expansion associated with adsorption in kernels with low initial MC. he curves presented in figure 4 represent HYR as a function of volumetric expansion rate for 0.25 h intervals when the kernels were exposed to 30 C and 90% RH conditions. he RVE was highest during the initial 0.5 h of exposure and subsequently decrease with increasing exposure time until RVE approached zero. he figure 4 should not be interpreted as that high RVE gives high HYR and viceversa. he HYR reduction was due to the cumulative effect of RVE with time. About 80% of HYR reduction occurred during the first 6 h of exposure for different Mj. For a given RVE, the HYR is a function of initial kernel MC Rate of volume expansion (RVE), xlo'vn^g dry matterh Figure 4-Chaiige in liead yield ratio of rougli rice samples with rate of volume expansion at ao"" C and 90% RH for different initial moisture contents (Mf). More specifically, as the Mj level decreased, the corresponding RVE increased and HYR decreased. CONCUSIONS 1. Rough, brown, and milled rice kernels exhibited linear relationships for: a) change in characteristic dimension with change in moisture content; b) change in volume with change in moisture content; c) change in volume with change in temperature; and d) change in coefficient of cubical thermal expansion with change in moisture content. 2. For all three rice forms, the coefficients of cubical and linear hygroscopic expansion were higher during adsorption than during desorption. 3. Coefficients of linear hygroscopic expansion were higher for brown rice than for rough and milled rice and the coefficient of cubical thermal expansion was higher for milled rice than for brown and rough rice during desorption. For all three rice forms, the coefficient of linear hygroscopic expansion was the highest for thickness followed by those for length and width for both desorption and adsorption. 4. Models were developed for estimating the relative volume of rough, brown and milled rice due to combined changes in moisture content and temperature of the kernels. 5. he rate of volume expansion of rough rice, which is a function of initial grain moisture content and exposure time during adsorption, results in a reduction of head yield ratio. REFERENCES Arora, V. K., S. M. Henderson and. H. Burkhardt Rice drying cracking versus thermal and mechanical properties. ransactions oftheasae 16(2): ,327. ASM Standard test method for coefficient of cubical thermal expansion of plastics. ASM designation: D ASM Standards. Banaszek, M. M. and. J. Siebenmorgen. 1990a. Moisture adsorption rates of rough rice. ransactions of the ASAE 33(4): RANSACIONS OF HE ASAE

7 _. 1990b. Head rice yield reduction rates caused by moisture adsorption. ransactions oftheasae 33(4): Dash, P. K Moisture adsorption in paddy and its effects on head rice yields. M.S. thesis No. AE 86-15, Asian Institute of echnology, Bangkok, hailand. Jindal, V. K. and. J. Siebenmorgen Effects of oven drying temperature and drying time on rough rice moisture content determination. ransactions of the ASAE 30(4): Kaye, G. W. C. and. H. aby ables of Physical and Chemical Constants, ondon: ongman Group united. Kunze, O. R Moisture adsorption influences on rice. Journal of Food Process Engineering 1: Kunze, O. R. and S. Prasad Grain Assuring potentials in harvesting and drying of rice. ransactions of the ASAE 21(l): Michael, F. Jr., H. B. emon and R. J. Stephenson Analytical Experimental Physics. Chicago, I: he University of Chicago Press. Mohsenin, N Physical Properties of Plant and Animal Materials. New Yoiic: Gordon and Breach Science Publishers. Morita,. and R. P. Singh Physical and thermal properties of short-grain rough rice. ransactions of the ASAE 22(3): Murthy,. S. N., B. N. Rao and K. K. Rao Physical properties of paddy grains. J, Agric. Engng., ISAE 23(4): Muthukumarappan, K Volumetric changes in rice kernels due to moisture desorption and adsorption. Unpub. M.S. thesis No. AE 88-19, Asian Institute of echnology, Bangkok, hailand. Prasad, S., J. D. Mannapperuma and F.. Wratten hermal and hygroscopic expansion of brown rice. Presented at the 1975 South West Region Meeting of the ASAE, Oklahoma. ASAE Paper. St. Joseph, MI: ASAE. Reyes, V. G., Jr A study of resistance to air flow through grains. M.S. thesis No. AE 83-17, Asian Institute of echnology, Bangkok, hailand. SAS Institute Inc SAS/SA Guide for Personal Computers, Ver. 6 Ed. Gary, NC. Steffe, J. F. and R. P. Singh Note on volumetric reduction of short grain rice during drying. Cereal Chemistry 57(2): Wratten, F.., W. D. Poole, J.. Chesness, S. Bal and V. Ramarao Physical and thermal properties of rough rice. ransactions of the ASAE 12(6): Yamaguchi, S., W. Kaichiro and S. Yamazawa Properties of brown rice kernel for calculation of drying stresses. Drying 85(l): VO. 35(1): JANUARY-FIEBRUARY