Three new methods for inoculating brassicas with Plasmodioohora brassicae were developed for glasshouse clubroot resistance tests. The ,1:modified slurry method" produced markedly higher and more consistent disease levels than the widely used "root dip procedure". It was furthermore simpler and required less inoculum. Improved techniques for single spore manipulation and inoculation were also devised. The resting spore maturation rate in diseased tap roots depended upon host species and P.brassicae population. Under glasshouse conditions infective spores were first detected 6 or 7 weeks after the sowing date, maximum infectivity being reached in 8- to 10-week-old roots, usually in galls which were firm and white and, in some cases, still swelling. Numbers (log10) of detectable spores per unit fresh tissue weight appeared to be almost constant and independent of plant age, h o s t 'species, pathogen population and gall size. The pathogenicity and virulence of inocula stored as concentrated resting spore suspensions for up to 6-12 months at -18 °C or h °C showed no decrease, but occasional marked increase. Loss of infectivity was eventually rapid and unpredictable and depended upon the P.brassicae population used. Inocula stored as frozen galls retained pathogenicity and virulence for at least 2 years. Differences in pathogenicity and virulence were recorded between spore populations from roots and soil at the same site. There were also marked differences between inocula from different host species, different plants of the same cultivar and different galls on the same plant. Eighty P.brassicae populations occurring naturally in South East Scotland were tested for physiologic specialisation on the European Clubroot Differential set. Most were virulent on the Brassica napus and B .oleracea differentials but avirulent on the stubble turnips (B.camnestris). One population carried virulence factors corresponding to all the differential hosts. The preponderance of intermediate disease indices was probably caused by pathogen, and to a lesser extent host, heterogeneity. The influence of an arbitrarily selected resistance/susceptibility cut-off value on population classification was often considerable. Many combined virulence pairs occurred more frequently than would have been expected from the individual virulence frequencies. This was mainly attributable to the high degree of similarity amongst resistance factors in the ECD set, but may also have been due partly to host selection on the soil population, particularly in the case of interspecific virulence pairs. The frequencies of the reaction patterns recorded on the B .nanus differentials suggested that this group is more likely to carry five resistance genes than three. Virulence to B.nanus occurred most frequently in populations from the Borders region. As unit doses of mature resting spores from different populations apparently differed in their capacity to cause disease, an assay could not be used for the accurate determination of absolute concentrations of infective spores in field soils. However, the test devised provided a reliable and simple means for evaluating comparative soil infectivities.