An extraordinary variety of experimental (e.g., flicker, magnetic fields) and clinical (epilepsy, migraine) conditions give rise to a surprisingly common set of elementary hallucinations, including spots, geometric patterns, and jagged lines, some of which also have color, depth, motion, and texture. Many of these simple hallucinations fall into a small number of perceptual geometries--the Kluver forms--that (via a nonlinear mapping from retina to cortex) correspond to even simpler sets of oriented stripes of cortical activity (and their superpositions). Other simple hallucinations (phosphenes and fortification auras) are linked to the Kluver forms and to pattern-forming cortical mechanisms by their spatial and temporal scales. The Kluver cortical activity patterns are examples of self-organized pattern formation that arise from nonlinear dynamic interactions between excitatory and inhibitory cortical neurons; with reasonable modifications, this model accounts for a wide range of hallucinated patterns. The Kluver cortical activity patterns are a subset of autonomous spatiotemporal cortical patterns, some of which have been studied with functional imaging techniques. Understanding the interaction of these intrinsic patterns with stimulus-driven cortical activity is an important problem in neuroscience. In line with this, hallucinatory pattern formation interacts with physical stimuli, and many conditions that induce hallucinations show interesting interactions with one another. Both types of interactions are predictable from neural and psychophysical principles such as localized processing, excitatory-inhibitory neural circuits, lateral inhibition, simultaneous and sequential contrast, saccadic suppression, and perceptual opponency. Elementary hallucinations arise from familiar mechanisms stimulated in unusual ways. (Contains 13 figures, 1 table and 9 footnotes.)