Abstract.A field experiment was set up in 1988 to study the development of rhizomania disease of sugar beet at different inoculum levels of beet necrotic yellow vein virus (BNYVV) in soil. Five, tenfold different, inoculum levels were created by addition of the approximate amounts of 0, 0.5, 5, 50 and 500 kg infested soil per ha (the latter corresponding to 0.01% v/v calculated to the tillage layer). A drip irrigation treatment was applied to study the influence of soil moisture on disease. Susceptible sugar beet, cv. Regina, was grown for three consecutive years.In the first year, root symptoms were not observed, but BNYVV-infected plants were detected by ELISA in low numbers at all inoculum levels at harvest. After late drilling in 1989, high numbers of infected plants, up to 90-100% in plots with the highest inoculum level, were detected already in June. Root symptoms were also observed from June onwards. In both these years disease incidence increased in time and was significantly influenced by the initial inoculum level. In the third year, the whole field was heavily diseased, and only for the non-irrigated plots incidence differed for different initial inoculum levels. The expression of symptoms by BNYVV-infected plants was influenced by initial inoculum level, thus by the amount and timing of primary infection.Root weight at harvest was not affected, but sugar content decreased with increasing inoculum level already in 1988, leading to a reduction in sugar yield of 10% at the highest inoculum level. In 1989, both root weight and sugar content decreased progressively with increasing inoculum level, resulting in sugar yield reductions of 11-66% (down to approximately 3000 kg ha -1) for low to high inoculum levels compared to the control. As the control plots became contaminated, all yields were low in 1990, still showing a decrease with increasing inoculum level in the non-irrigated plots, but an overall mean sugar yield of 3323 kg ha -1 for the irrigated ones.Sodium and c~-amino nitrogen content of the root, additional quality parameters determining extractability of sucrose, showed an increase and decrease, respectively, with increasing initial inoculum level already in the first year. The relative differences in contents compared to those from the control were largest for Na content. A significant negative correlation was found between Na (mmol kg -I root) and sugar content (% of fresh weight); linear for 1988, exponential for 1989 and 1990. In spring 1989, the infestation of individual plots was assessed using a quantitative bioassay estimating most probable numbers (MPNs) of infective units of BNYVV per 100 g 20 dry soil. The relationship between the MPNs determined and root weight, sugar content and sugar yield at harvest could be described by Gompertz curves. The increase in disease incidence with increasing MPN in 1989 was adequately fitted with a logistic equation.