A polyclonal antiserum produced against spore balls of Spongospora subterranca f.sp. subterranea prepared from potato tubers was able to detect as little as 0·02 spore balls in an enzyme-linked immunosorbent assay (ELISA). In spiked soil samples, the antiserum detected 100 spore balls per g soil. However, the different spore ball contamination levels were discriminated better in ELISA tests al concentrations above 2000 spore balls per g soil than at lower concentrations. In contrast, a bioassay test based on baiting soils with tomato seedlings gave good discrimination of spore ball contamination levels in spiked soils containing <1000 spore balls per g soil and poor discrimination of levels in spiked soils containing >2000 spore balls per g soil. Tests on a limited number of field soils suggested ELISA may be capable of predicting disease levels on tubers group in such soils better than the bioassay. The antiserum did not react with 30 other micro-organisms tested, including many that are saprophytes or pathogens on potatoes and resting spores of the taxonomically related Plasmodiophora brassicae. It detected spore balls of different cultivar origin equally well. It also detected spores from different geographical origins. An attempt to improve the sensitivity of the serological detection through concentrating spore balls from held soils by sieving was unsuccessful. Cross-absorption of the antiserum with uncontaminated field soil increased the sensitivity of detection of spore balls in spiked soil samples four-fold. The ability of the antiserum to discriminate contaminated field soils from an uncontaminated soil was much improved by using the gamma-globulin fraction or cross-absorbed serum. Western blot analysis revealed that the antiserum detected a member of different proteins the most distinct of which had a molecular weight of slightly less than 6.5 kDa. A technique was developed to suppress autofluorescence of spore balls, allowing immunofluorescence studies to be carried out. Using this technique in conjunction with indirect FITC immunofluorescence discrete bright fluorescent spots were visualized using the specific serum. With the non-specific serum, only a very dull background fluorescence was evident.