Olfactory hallucinations (OHs), so it has been argued, are prognostic of a poorer outcome, are unpleasant, and cannot be well explained within current theoretical accounts of hallucinations. We examined these and related issues by conducting structured interviews with 51 participants who experienced OHs and who were previously diagnosed with schizophrenia or schizoaffective disorder. We found no relationship between disease severity measures and type or frequency of OHs. As with prior research, we too noted the predominance of negative OHs, but with many reports of positive OHs, and also found significant relationships between frequency of OHs and severity of tactile hallucinations. We then examined whether odor imagery or involuntary memory might account for the presence of OHs, but these possibilities were not well supported. We then explored, using cluster analysis, whether or not our sample was homogenous. Two clusters were of especial interest; one which may reflect a 'sensory dysfunction group' and one characterized by more severe tactile hallucinations. The presence of tactile hallucinations may suggest a further novel cause of OHs, which we discuss. Our data suggest diverse causes for OHs in schizophrenia, none of which are consistent with current models of hallucinations in other modalities.