Field experinents on soil erosion under natural rainfall were conducted in 1986 and 1987 on Gretna clay, Leary sandy loam, Ryerson sandy clay loam, and Carroll clay loam soils. The aim was to develop a data base for evaluating the soil erodibility (K) and the crop-management (C) factors of the Universal Soil Loss Equation (USLE) under Manitoba agricultural conditions. Experimental design conformed closely to that used when the USLE was developed. Crop-management treatments were continuous and included alfalfa, conventional tillage wheat, minimum tillage wheat, conventional tillage corn and summerfallow. Rainfall data were measured with a tipping bucket rain gauge and were used to determine rainfall erosivity (R) values. Surface runoff and soil loss data measurements were done using a Coshocton sampling system. Crop cover measurements were done regularly with a modified point-line method to determine different crop growth stage periods. Antecedent soil moisture was measured using the gravimetric analysis method. Soil losses were observed to be extremely variable among different crop-management treatments and soil types. This was mainly due to the extreme variability of rainfall erosivity, antecedent soil moisture, cultivation, crop cover changes, residual effects of mulch and previous crop cover and possible undersampling of soil loss. As a result, the measured K and C factor values were observed to be extremely variable. The measured K values were compared to those estimated using the USLE nomograph equation (NE) and the Modified Young and Mutchler Equation (MYME). The estimated values for both equations were not very different for most soils except Gretna clay soil. The average measured K values for Gretna clay and Leary sandy loam soils were quite comparable to the NE estimated values. These values showed that the two equations could possibly be underestimating the K values for these soils. The measured K values for Ryerson sandy clay loam and Carroll clay loam soils were extremely low, possibly due to the effects of the previous crop residues. The measured ratios of soil loss fron the cropped treatments to that from summerfallow treatment were observed to be extremely variable among similar and different treatments for similar growth stages. These ratios were not easily comparable to the USLE estimated values, due to the differences in the crop-management systems for the experiments and those used when the USLE was developed. The effects of antecedent soil moisture and cultivation in modifying observed soil losses and the measured K and C factor values were found to be important. This observation suggested the need for the modification of the USLE to account for the effects of these field factors so as to ensure accurate estimation of soil loss. The results obtained reflect the need for long-term measurements of soil loss and influencing factors before effective evaluation of factor values can be obtained. The short-term duration of this study and the absence of comparable crop-managenent systems and proper summerfallow conditions in the experiments limited the ability of this study to come up with conclusive results.