Diet survey in this country has indicated that the adequacy of the supply of vitamin B is mainly determined by the quantity of this vitamin presence in the staple food consumed. Analysis of milled kaoliang grain obtained in the market shows a much lower than adequate content of the vitamin. In a previous report, it has been shown that the distribution of various nutrients in this grain is similar to those in other cereals. It is apparent that the study of the effect of the milling process on the nutritive value of kaoliang is of considerable importance both to the improvement of human nutrition and to the conservation of this staple cereal.Experiment was made by running a quantity of red kaoliang sample eight times through a standard commercial mill. Samples were withdrawn after each run for determining the amount of various parts of the grain retained and for the analysis of thiamin, riboflavin, nicotinic acid, protein, fat, crude fibre and xylose. The weight of the whole grain removed was also estimated.More than 70％ of the pericarp was found to have been removed in the first run. When 20％ of the total weight was removed, the amount of pericarp left was negligible. This was in accordance with the loss of crude fibre and xylose which showed a greater rate of removal during the first few millings.With the increase of the degree of milling, the ateurone layer, the scutellum and the embryo were removed nearly at a similar rate. At the earlier stage of milling, the loss of thiamin was at a lower rate than that of nicotinic acid and riboflavin. After 15％ of the bran had been removed, its loss became more rapid and ran parallel with that of nicotinic acid, while the removal of riboflavin was at a comparatively lower rate. This can be satisfactorily explained by the uneven distribution of these vitamins in different parts of the grain.The percentage of fat removed at different stages of milling was about the same as that of riboflavin. High degree of milling also caused a lower content of protein in the grain.For better controlling the degree of milling and for reducing the friction in the milling process, a laboratory mill was used in a second experiment. The results obtained were nearly similar to those stated above with the exception of a better retention of scutellum during the early stage of milling.It is suggested that 80％ extraction is likely to be the most rational degree of milling. Owing to the differences between varieties, this may be varied accordingly.It seems that the content of xylose may be used as an index for the degree oi milling.