Roasting of refractory gold ores has been a commercial process for many years. Improvements in design of fluidised bed roasting technologies and low capital and operating costs compared with other refractory treatment processes can potentially make roasting a viable option.However, gold extraction from roasted ores carried out on laboratory scale equipment often invariably yields lower recoveries than other processes such as bio-oxidation or pressure oxidation. Non-optimal laboratory roasting conditions can incorrectly write roasting off as a process option.Low gold recoveries from the cyanidation of calcine products from roasting can often be attributed to roasting conditions not favouring gold liberation and agglomeration processes, particularly if lockups of elements such as arsenic are being pursued for environmental reasons.This paper reviews the conditions favoured for gold agglomeration and accessibility of gold to cyanide, test conditions to achieve gold liberation and agglomeration on laboratory scale and designs that incorporate staged decomposition and roasting to make roasting a more commercial option.CITATION:Aylmore, M G and de Klerk, L W, 2013. Conditions and design considerations for maximising recoverable gold in roasting of refractory gold ores, in Proceedings World Gold 2013 , pp 375-388 (The Australasian Institute of Mining and Metallurgy: Melbourne).