People find their way through cluttered environments with ease and without injury. How do they do it? Two approaches to wayfmding are considered: Differential motion parallax (DMP) is a retinal motion invariant of near and far objects moving against fixation; the information in optical flow (IOF) is a radial pattern of vectors, relying on decomposition of retinal flow. Evidence is presented that DMP guides wayfinding during natural gait, accounting for errors as well as correct responses. Evidence against IOF is also presented, and a space-time aliasing artifact that can contaminate IOF displays is explored. Finally, DMP and IOF are separated, showing they can yield different results in different environments. Thus, it is concluded that (a) DMP and IOF are different, (b) DMP and not IOF is used for wayfinding, (c) moving observers do not usually decompose retinal flow, and (d) optical flow may be a mathematical fiction with no psychological reality.One of the most compelling of all visual phenomena occurs when one hurtles through the environment. The resulting radial streams of motion by surrounding objects, sometimes called optical flow, have captured the imagination of writers, artists, and cinematographers, as well as psychologists, neuroscientists, and computer scientists. This global motion was probably first noticed by the general populace in the mid19th century, but then only in industrializing nations and with the widespread use of railroads.The reasons for the relatively recent focus on optical flow are probably threefold: On a train one could, for the first time, travel (a) at velocities greater than about 4 eye heights/ s for a sustained period of time; 1 (b) on a relatively smooth roadbed that eliminated the bouncing caused by one's own footfall or that of a horse, or by the jostling of a coach; and (c) with free time to look about, unfettered by the demands of guiding one's course through the environment. Naturally, railway travel offered much more than noticeable optical flow. Not all of it was good. Indeed, around 1860 The Lancet published a series of articles on putative ill effects of rail travel This research was supported by National Science Foundation Grant BNS-8818971 to James E. Cutting.Experiment 1 was reported at the 30th Annual Meeting of the Psychonomic Society, Atlanta, November 1989; parts of Experiments 2 and 3 at the Society's 31st Annual Meeting, New Orleans, November 1990; and parts of Experiments 7 and 8 at its 32nd Annual Meeting, San Francisco, November 1991. Experiments 1, 2, 4, 5, and 6 were also reported at a perception-action workshop at Storrs, Connecticut, in July 1990.We thank Laurence G. Kaplan for initial and sustaining help in programming the Personal Iris Workstation; David Sabean for discussions of 19th-century perception; Gretchen Van de Walle, Jack Loomis, and several anonymous reviewers for their comments; and Julian Hochberg and Nan Karwan for various discussions of this project over the past 6 yearsCorrespondence concerning this article should be addressed to J...