A high yielding long grain rice variety (BR4), developed by the Bangladesh Rice Research Institute, was used throughout the experiment. The paddy was grown in Bangladesh during August to December 1997. After harvesting, the grains were separated from panicles by the drumbeating method and dried by spreading on a solid floor under the sun to achieve a moisture content of typically c. 14%. About 97% matured filled grain was obtained. The mean length/breadth ratio of ten randomly selected grains was 2.7.The amylose content of this rice variety is 25% (Biswaset al., 1992) and the granule size was within the range of 5–8 mm (Ali & Pandya, 1974).About 20 kg of paddy sample was brought to the UK in January 1998 (about a month after harvesting). Each of 20 polythene bags (250lm thick) was filled with 1 kg of sample, sealed and stored in an incubator at 25°C. Cold soaking Paddy was soaked with tap water for 15, 30, 45, 60and 120 min. The temperature of soaking water was measured and ranged from 21 to 22°C. A 20-gsample of paddy was weighed in a container with 26 mL of cold water and placed in an incubator at a constant temperature of 25°C. Hot soaking A stainless steel pan containing 762 mL of water was heated to 100°C and a 600-g sample of paddy was poured into the pan, which resulted in the temperature dropping by 20°C within 30 s. The pan with a lid was transferred to an oven maintained at a constant temperature of 80°C. Samples were soaked for 15, 30, 45, 60 or 120 min and their moisture contents determined after each soaking treatment, on removal of surface moisture by blotting paper. Steaming Samples were steamed in a 4-L capacity pressure cooker at an excess pressure of 1 atmosphere for10 min, during which the temperature inside the pressure cooker increased to 121°C. After steaming, tap water was sprayed onto the cooker, through a rubber tubing, at a rate of 140 mL s–1 for fast cooling before opening the lid. Moisture content was then determined on a representative sample and found to be between 32 and 37%. Moisture meter calibration One kilogram of paddy was parboiled for calibration of a moisture meter (model Grainer IIPM 300, Farm-Tec, Scunthorpe, UK), a capacitance type meter with a specified accuracy of±0.5%. The initial moisture content of steamed paddy was determined by the oven method using c. 20 g of sample. The remaining paddy was placed in a thin-layer drier. After 1 h, two samples of 100 and 20 g were removed and the moisture content was determined three times by the moisture meter and oven method, respectively. The 100-g sample was then sealed in a plastic jar for 16 h at room temperature (c. 20–22°C) for tempering. The same procedure of moisture determination was followed after every 30 min of drying for 4h. The moisture content of the tempered grain was also determined by both methods.Drying and storage of steamed paddy A thin-layer laboratory scale drier with six sieves was used, in which hot air passed through the grain. The diameter, cross-sectional area and depth of each sieve was 195 mm, 0.03 m2and 45 mm, respectively. The initial moisture content was determined by the oven method. The dry bulb temperature, wet bulb temperature and relative humidity were monitored by a sling psychrometer.The wet paddy was divided into six equal parts and placed on the sieves (2 mm aperture) for drying. The average grain depth and drying air velocity were 2 mm and 0. 4m s–1, respectively. All samples were initially dried at 35°C until the moisture content was c. 18–20%. The drying temperature chosen was based on a study by Uddinet al. (1987), where a drying temperature ranging from 31 to 37°C was used. Samples were then subjected to 16-h tempering in a sealed plastic container at c. 20–22°C.Finally, after about 45 min of further drying at c.22–24°C, a moisture content reading of 16% was achieved, which, as will be seen later, is equivalent to a moisture content of 14% as determined by the oven method.Tempering of paddy during drying is very important for obtaining better milling quality. When the parboiled paddy was dried rapidly without tempering, a steep moisture gradient developed between the surface and the centre of the kernel, setting up internal stresses. At a certain stage, the kernel relieves these stresses by cracking. These cracks are irreversible and set up lines of weakness along which fractures may easily occur under mechanical stress during milling (Bhattacharya & Swamy, 1967). To avoid this happening, the samples, at a moisture content of c.18–20%, were kept in airtight plastic containers at c. 22–24°C for 16 h. During this equilibration process, moisture distributed evenly throughout the grain, allowing a precise moisture content to be measured. After drying, the samples were cooled and kept in sealed polythene bags. All samples were stored at a temperature of 25°C in an incubator. Milling Rice milling involves two operations, husking and polishing. A 150-g sample of parboiled paddy was dehusked using a Satake rubber roll huller (model THU-35 A, Satake Ltd, Tokyo, Japan). The weight of the cleaned brown rice was recorded. The brown rice was then polished by an abrasive type polisher (McGill Miller no. 2, Rapsco, Brookshire, TX, USA) for 1 min to obtain a typical degree of polish of c. 8%. The weight of the cleaned polished rice was recorded. All samples were replicated three times for the husking and polishing operations. Individual replicate sizes were 150 g samples from each 600 g of the original hot soaked paddy. Milled rice yield was expressed as the percentage ratio of milled rice after husking and polishing, to the weight of unmilled paddy. Head rice yield was expressed as the percentage ratio of the weight of whole (unbroken) kernel to the weight of the total milled rice yield. Broken rice was quantified as100 – head rice yield.