The objective of the study was to assess the development and evolution of visual and nonvisual hallucinations in patients with Parkinson’s disease over 10 years. Hallucinations increase over time, but minimal attention has been placed on nonvisual domains. We studied 60 patients with Parkinson’s disease who had never hallucinated at baseline and followed them over 10 years. The Rush Hallucination Inventory monitored frequency and type (visual, auditory, tactile, olfactory) of hallucinations at baseline and after 0.5, 1.5, 4, 6, and 10 years. Descriptive statistics were applied, and general estimating equation modeling assessed longitudinal risks. Over 10 years, visual hallucinations were endorsed by patients more frequently than other sensory modalities. Whereas isolated visual hallucinations dominated the early hallucination profile, visual plus nonvisual hallucinations accounted for progressively higher proportions of hallucinators over 10 years: 0.5 years, 0%; 4 years, 26%; 6 years, 47%; 10 years, 60% (odds ratio, 1.17; confidence interval, 1.01–1.37; P = .04). Once visual plus nonvisual hallucinations developed, the risk of continuing to have multidomain hallucinations was high (odds ratio, 3.67; confidence interval, 1.13–11.93; P = .03). Hallucination severity was highly associated with current visual plus nonvisual hallucinations (odds ratio, 4.06; confidence interval, 2.93–5.61; P < .0001) and the continuation of multidomain hallucinations (odds ratio, 1.58; confidence interval, 1.12–2.24; P = .01). Whereas visual hallucinations in isolation are classic in Parkinson’s disease, nonvisual hallucinations emerge over time, and the combination of visual with nonvisual hallucinations predominates in late Parkinson’s disease. To capture the breadth and severity of hallucinations in chronically hallucinating patients with Parkinson’s disease, screening inventories and practice-based interviews must include questions on both visual and nonvisual components.