At present, nearly 99% of the world's niobium concentrate is produced by the Catalao and Araxa mines in Brazil and the Niobec mine (St. Honore) in Canada. In total, Araxa produces approximately 150,000 tonnes of niobium per annum (tpa), while Niobec and Catalao produce 5300 tpa and 4500 tpa, respectively. Niobium recovery at all three operations is near or above 60%, with final concentrates grading 55-65% Nb2O5. The geographic diversification of niobium supply depends on the development of operable beneficiation flowsheets for the upgrading of known niobium deposits worldwide. The beneflciation of niobium oxide ores is predominantly carried out by froth flotation, which generally involves niobium mineral flotation at acid pH using a cationic amine collector. In the case of the Niobec and Catalao mines, there is a reverse gangue mineral flotation step prior to niobium flotation. The main issues associated with currently employed flowsheets are loss of niobium to the gangue mineral concentrate and the slime fraction. Further, there are a number of gaps in understanding with regard to the mechanisms by which amine collectors and auxiliary reagents interact with one another and with niobium minerals. To date, several alternate flotation reagents such as hydroxamates, sulfosuccinates, phosphorous compounds, and hydroxyquinolines have been examined with the aim of improving niobium recovery and concentrate grade in the flotation process. Notably, alkyl hydroxamates were identified in the 1960s as a potential collector for direct niobium mineral flotation eliminating the need to first deslime the ore. Subsequent testwork performed on Niobec ore using this collector has corroborated its effectiveness without desliming. However, due to increased reagent addition requirements when using hydroxamate collectors, the cost benefit of this reagent scheme compared to amine collectors is unclear. Overall, greater understanding of mechanisms at play in currently employed reagent systems is required to effectively improve plant operating conditions and maximize recovery at the industrial scale. Future testwork should also be conducted on real ores using reagents that have been identified as selective collectors at the single mineral flotation level. (C) 2015 Elsevier B.V. All rights reserved.