Studies prerequisite to introduction of protein segregation of wheat were undertaken in collaboration with the Canadian Grain Commission. The principal data comprised a 5% stratified sampling of all carlots of wheat unloaded in the Western Division since 1 August, 1939, with records of origin, destination, unloading date, grade, protein content, etc.. Nine crop years were selected with protein levels of 14.4, 14.1, 13.7, 13.7, 13.6, 13.5, 13.4, 13.1, and 12.2%, and data for the 94,894 carlots were read onto magnetic tape for computer studies. Nos. 1 and 2 Northern were combined to simulate the new grade of No. 1 Canada Western. The Commission’s original proposal, that this grade be divided into three subgrades by protein content, namely, over 14.2%, 13.0 to 14.2%, and under 13.0%, was then examined by superimposing it on data for the 9 selected yr. The middle subgrade was reasonably satisfactory, containing 40–60% of the main grade in all but extreme years, and with a protein level varying between 13.75 and 13.35%; but the other two subgrades varied in amount from less than 10% to over 50%, and the protein level of the upper subgrades varied from 14.5 to 15.1%. In accordance with revised thinking, systems were then examined for guaranteeing subgrades at 15.0, 14.0 and 13.0%, or at 14.5, 13.5 and 12.5% (and at higher or lower levels when available). If boundaries are set low enough to meet guarantees in low protein years, levels 0.25–0.6% above guarantees occur in high protein years. Three subgrades, though not necessarily the same three, differing in protein level by 1.0% were feasible in all years, but strict uniformity in protein level could not be achieved within subgrades with systems involving constant boundaries between subgrades. Promotion of carlots from one subgrade to the next, on unload at terminal elevators, was then examined. In effect, this involves moveable boundaries, e.g. to guarantee 14.0% protein, the middle subgrade would have boundaries of 13.8–14.8% when the main grade has a mean of 13.0% protein, and of 13.9–14.3% with a mean of 14.5%. By comparison with constant boundary systems, from 8 to 25% of all carlots would be promoted by one subgrade, and strict uniformity of protein levels is achieved for all but the bottom subgrade. New data were collected to examine the practicability of systems involving moveable boundaries, and these studies will be reported in a second paper.