Staphylococcus aureus Protein A (SpA) labelled with [125I] by the Bolton & Hunter (1973) method performed about as well as labelled sheep anti-rabbit globulin (SAR) in an indirect immunoradiometric assay (IRMA) for Bacillus anthracis spores immobilized on multispot microscope slides. SpA labelled with [3H] by propionylation also performed well but would be expensive to use. SpA labelled with [3H] fluorodinitrobenzene, or labelled with [125I] by the chloramine T reaction gave erratic assay results, high noise values and low signal-to-noise ratios, indicating substantial direct binding of labelled SpA to the slide surface and to the bacterial preparation. The uptake of radioactively labelled SpA in the IRMA was compared with the fluorescence intensity of individual spores in a microfluorometric immunofluorescence (IF) test involving dual labelled fluorescein-[125I]-SpA. The maximum number of SAR molecules bound to the mixture of spores and cell-free antigens in the B. anthracis IRMA was about twice the maximum number of radioactively labelled SpA molecules bound. The SAR:SpA saturation binding ratio on the surface of the spores, however, was approximately the inverse of this. It is concluded that radioactively-labelled SpA is not recommended in preference to anti-species antibody reagents in bacterial IRMA tests but fluorescein-conjugated SpA deserves further consideration for use in microscope-based IF tests for bacterial antigens.