Oral candidosis has become an increasingly important problem in HIV-infected individuals. At present, the small body of published literature on the characterization of the Candida strains and species found in HIV+ patients is full of confusion and contradictions. Some of these difficulties are the result of the methodological shortcomings of a number of the techniques that have been used. Examples of the problems that may be encountered on primary isolation and subculture are described and the drawbacks associated with the systems used to date for phenotyping Candida are quoted. While molecular characterization techniques would appear to offer a reliable and objective alternative, they too have their strengths and weaknesses. An attempt is made to summarize the progress that has been made recently in the detection and identification of Candida albicans and also the non-albicans species from HIV-infected individuals. What emerges is that the commensal Candida species that inhabit the oral cavities of HIV+ patients are subjected to a number of significant pressures that probably promote the selection of organisms with unusual phenotypes and genotypes. These Candida are more difficult to characterize and behave differently compared to their counterparts in HIV- individuals. It is clear that uncovering the factors that are important for the selection of treatment regimens and will be predictive of outcome will not be easy. Candida organisms are neither as benign nor as simple as once thought.