Kinetic flotation models were applied to data from tests with reactive gases (hydrogen and sulphur dioxide) on a complex sulphide ore. The models were evaluated by statistical techniques, after non-linear regression on the model parameters. It is found that a first order model with rectangular distribution of floatability fits the test data very well. Statistical tests on copper, silver and lead kinetic data show that reactive gases have a greater influence on silver mineral flotation compared to copper-lead minerals. Sulphur dioxide generally gives higher ultimate silver recoveries than sodium bisulphite, and the magnitudes depend on the interactions between the modifiers and the flotation gases (5% H-2, N-2 and air). Correlation analysis suggests that silver flotation is probably determined by the floatability of a separate mineral such as tetrahedrite. Multivariate analysis on pulp chemistry data reveals that the pulp pH is the most significant variable among the pulp chemistry data, provided the grinding is iron-free. The pH effect is attributed to the addition of sulphur dioxide. After autogenous or non-reducing grinding, the naturally-occurring variations in redox levels are too small to influence the flotation results, despite the tests being made with a complex sulphide ore.