Sulphur formation by the obligately chemolithoautotrophic Thiobacillus o and Thiobacillus neapolitanus was studied in aerobic, substrate-limited continuous cultures. The performance of transient-state and steady-state cultures was compared using different methods for measuring sulphur production. Below a dilution rate (D) of 0.3 h(-1) (at 50% air saturation), sulphate-producing steady states were obtained, and cultures grown with sulphide or thiosulphate (at D = 0.06 h(-1)) showed similar characteristics (e.g. cell yields, oxidation capacities and CO2-fixation capacities). Elemental sulphur was a major product above D = 0.3 h(-1), but steady states were difficult to achieve, because of adherence of sulphur to the fermenter surfaces and the accumulation of sulphide. These problems could be circumvented using transient-state experiments of 1 h. It was then found that elemental sulphur was formed under oxygen limitation or at high substrate load. The rates of sulphur formation obtained by sulphur analysis agreed with the values calculated from stoichiometric balances. Sulphide and thiosulphate proved to be equivalent substrates for both Thiobacillus species during elemental sulphur formation under the conditions tested. It is concluded that transient-state cultures of thiobacilli, pregrown as sulphate-producing steady-state cultures, provide experimental conditions for the quantitative assessment of sulphur formation from (labile) sulphide and from thiosulphate.