Results of two groups of experiments testing phosphate fertilizer for sugar beet were re-examined; there were 53 experiments made between 1957 and 1960 in group 1 and 25 experiments between 1957 and 1959 in group 2. The mean response of total sugar to 126 kg/ha P a O 6 was only 160 kg/ha but on a few fields response exceeded 1000 kg/ha. Response seemed to be greater in 1958 than in the other years, but between-site variance accounted for most of the difference from year to year.Stored soil samples were analysed for phosphorus by four methods; the results were compared and related to the yield response to phosphate fertilizer. After allowing for experimental error, the percentages of the between-sites variance of the responses accounted for by log (soil P) were (group 1 first): sodium bicarbonate, 60 and 62 %; anion resin, 52 and 30%; ammonium acetate/acetic acid, 52 and 0%; and calcium chloride, 42 and 2 %. When soil pH was included in the regression equation, prediction of response by anion resin much improved and equalled that of sodium bicarbonate. Ammonium acetate/acetic acid was ineffective on soils with much free calcium carbonate.As two-thirds of the soils had sodium bicarbonate-soluble phosphorus concentrations between 15 and 45 ppm P with mean response 65 kg/ha sugar, there is only limited scope for increasing the profitability of the crop by improving P manuring. However, more P fertilizer can be recommended for the few soils with =S 10 ppm P and P fertilizer can be withheld from fields with more than 45 ppm. The dressings we recommend are 180, 120, 60, 30 and 0 kg/ha P 2 O 5 (approximately 1-5, 1-0, 0-5, 0-25 and 0 cwt/acre P 2 O 6 ) on soils with < 10, 11-15, 16-25, 26-45 and ^45 ppm sodium bicarbonatesoluble P respectively.