The effect of different parameters on the decarboxylation of acetolactate (ALA) to diacetyl and acetoin were studied. The distillation volume and the milk solids concentration had no significant effect on decarboxylation of ALA, whereas breakdown of ALA increased with decreasing pH and increasing temperature. Oxygenation increased diacetyl production from ALA, but diacetyl was lost from the model system. Oxygenation did not have an effect on acetoin production from ALA. Metal ions (Cu2+, Fe2+) and haemin caused high breakdown of ALA to diacetyl during steam distillation, with Cu2+ being the most effective. The decarboxylation of ALA was a first order reaction. A new method was developed for the determination o f ALA based on steam distillation at pH 3.5 in the presence of Cu2+, which caused complete decarboxylation of ALA to diacetyl. ALA concentrations were calculated from the difference between diacetyl levels in a sample in which ALA was completely converted to diacetyl, and diacetyl levels in a sample with minimal decarboxylation of ALA to diacetyl, which was achieved by distillation at pH 0.8. The method compared well to the Jordan and Cogan [1995] method. Trials were carried out at laboratory and industrial scale to improve the manufacturing process for lactic butter. An increase in temperature during manufacture increased diacetyl concentrations and converted most of the ALA to diacetyl; it had no significant effect on acetoin concentrations. Cit+ strains of Lactococcus lactis subsp. diacetylactis, 999 and 1166, were grown in the presence and absence of oxygen, leucine, valine, CuS04, FeS04 and haemin. Except for oxygen, which increased diacetyl production and decreased growth, there was no significant effect o f these compounds on metabolite production.