Several attentive readers have asked whether and how the present findings account for the "rotating snakes" illusion (http://www.ritsumei.ac.jp...) - a very powerful variant of the peripheral drift illusion (see also Kitaoka A & Ashida H, 2003, Vision, 15, 261-2). Both our illusion and the "rotating snakes" illusion use similar drift patterns. However, in the "rotating snakes" illusion the drift patterns are arranged in a circle. Accoring to our notion, this rotational arrangement has the benefit that eye movements in all directions contribute to illusory motion. Moreover, the size of the patterns in the "rotating snakes illusion" steadily increases from small in the center to large in the periphery of the circles. Pattern size likely has an impact on the illusion as our and previous data (Conway BR et al., 2005, J Neurosci, 25, 5651-6) suggests that the illusion involves motion-sensitive cells in early visual cortex including the primary visual cortex (V1). Varying the size of the patterns assures that there are always some patterns that match the small receptive fields of neurons associated with the illusion. Consequently, the "rotating snakes illusion" is relatively insensitive to the direction of eye movements and the viewing distance. In our study, drift patterns were arranged horizontally. Accordingly, we expected that only (or primarily) horizontal eye movements had a relevant effect on illusory motion. Our goal was to simplify the illusion. Horizontal patterns allowed us to better control and test for the role of central patterns on peripheral patterns than more complex patterns. As we kept viewing distance constant a large variation of pattern sizes was not necessary. The horizontal motion observed with our pattern arrangement likely involves area MT, which is known to pool motion signals across large proportions of the visual field (e.g., Zeki SM, 1974, J Physiol, 236, 549–73; Rust NC et al., 2006, Nat Neurosci, 9, 1421-31). The rotational motion observed with the rotating snakes illusion likely involves neurons in area MST, which is known to encode optic flow patterns such as rotational or radial motion (e.g., Duffy CJ & Wurtz RH, 1997, Exp Brain Res, 114, 472-82), in addition to (or instead of) area MT. Therefore, the mechanisms involved in our variant of the peripheral drift illusion may differ from the mechanisms involved in the "rotating snakes illusion" at a higher level of processing. However, at a lower level the mechanisms should be similar for both variants of the illusion.