Yields of sugar and responses to fertilizers were determined in about 400 fertilizer experiments on farms throughout the sugar-beet growing areas of Britain during 14 years from 1957 to 1970. The soil at each experimental site was described and classified and the records of the experiments have been examined to determine which properties of the soil influence sugar yield. The effect of year, rainfall, elevation, region and other factors such as sowing and harvesting date were also investigated.Year-to-year variation accounted for 20 % of all variation in yield; increasing amounts of rainfall during the growing season appeared to decrease yield. There was a significant long-term trend of increasing sugar yield from the experiments of 0-042 t/ha/year. The experimental yields closely followed national yields each year but were always greater. Delay in sowing and early harvesting depressed yield by 0-02 and 0-01 t sugar/ ha/day respectively. Yields in Scotland (average 5-23 t/ha) were approximately 1-4 t/ha less than in England and Wales, but there were no evident regional differences within England and Wales.Soil regarded as moderately drained yielded better than either well-drained or imperfectly drained soil, which in turn yielded better than droughty and poorly drained soil. The difference in the adjusted yield between drainage classes was 1-0 t/ha. Surprisingly, topsoil texture had no consistent effect on mean yield (as distinct from response to fertilizer). Subsoil texture, however, had an appreciable effect, the crop on sandy subsoil and chalk or limestone yielding poorly whilst that on silt or peat yielded best. The range of differences in sugar yield due to subsoil texture was almost 2 t/ha. Yields were also examined in relation to soil profile type. Broad division into major soil groups gave meaningful differences but fine division by soil series was only useful for the 11 series on which at least ten experiments had been made. The crop yielded most sugar on gleyed calcareous soils, peats and humic gleys, and least on rendzinas and brown calcareous soils. Responses to nitrogen and potassium but not to phosphorus were affected by both topsoil and subsoil texture. Nitrogen and potassium both increased yield most on sandy soils and least on fine silts and peats.The morphology, chemical and physical condition of soil clearly affect sugar yield greatly and further research is needed in experiments planned specifically to measure their influence and provide more precise guidance for selecting the best land for the crop.