Fluctuation in farm incomes resulting from variation in crop yield is one of the most significant features in agriculture. Crop insurance is a feasible method by which the farmer can protect his income and his investment from the disastrous effects of crop losses due to natural hazards. This study has attempted to cover two parts. First, it has examined the most important‐factors influencing crop yields in connection with the premium rate scheme (i.e. the long‐run average yield and the level of coverage). These factors include resource inputs, technology, weather, and stochastic variable. Second, it has developed a refined method of approximating the premium rate. The data used in testing normality were based on the Manitoba Crop Insurance Corporation's annual yield survey covering the years 1916 to 1964. The test shows that none of the annual yield distributions within the area surveyed was normally distributed, revealing that a cartful and exact delineation of a crop risk area is necessary. The findings also show that the cyclical pattern of weather and the upward trend in crop production due to technology were evidently important for the adjustment of the level of coverage and premium rate over time. Additional research relating the effects of weather and technology on crop yields would help to establish a more realistic insurance program. Other aspects should not be overlooked. These include (a) other possible levels besides the existing level of coverage and (b) a livestock or a combined crop‐livestock insurance program along with the crop insurance program. The purpose of these additional aspects is to provide farmers with a fuller measure of prote