Mixtures of collectors have been widely used for many years in sulphide flotation, and a range of performance benefits have been reported for many different systems. The combinations of collector types have varied, as have the ratios that have been used. Synergistic effects have been obtained (greater than the sum of the parts) and in some cases the mechanisms of this improved behaviour have been identified. These benefits have been attributed to increased carrying capacity of the froth phase, faster kinetics, and more successful recovery of middling or coarse particles. It is the interaction between the various components of the mixed collector system, rather than the individual main effects, that dominate the performance benefits. The process benefits include increased paymetal recoveries and grades – as well as increased rates of recovery whilst using lower dosages of reagents. Various mechanisms have been reported and are discussed. These have been shown to affect different composition/liberation classes and sizes of mineral particles. In recent years, automated quantitative mineralogy and surface analysis technology such as ToF-SIMS have enabled the development of better information, to establish what aspect of the process has been affected. This has been successful mostly for use in a diagnostic capacity. Candidate selection for the mixed collector suite is presently based on experience and contextual knowledge. Predictive properties from these systems are a desirable future goal. Currently optimum combinations are preferably identified experimentally at laboratory scale prior to any plant trial. It is recommended that such laboratory work be performed using a factorial design with replicates and quality controls, such as may be delivered from High-Confidence Flotation Testing. The purpose of this paper is to summarise and review current theory and practice in the usage of mixtures of collectors in sulphide flotation – both in the application and in research in order to develop insights and guidelines to develop a methodology for use in a predictive capacity. A case study demonstrating this approach will be published at a later date.