Using photographs from the DMSP satellite Lui et al. (1982) have found several examples of large scale undulations of the edge of the diffuse aurora. Several satellite experimenters have previously reported intense sheared flow near the edge of the diffuse aurora as evidenced either by detection of large localized meridional electric fields or by narrow regions of high azimuthal flow velocity. It is suggested here that the diffuse auroral undulations are caused by an instability of this sheared flow. In order to test this idea, a search for in situ satellite flow data associated with the events presented by Lui et al. was made. In addition the shear flow events reported by Rich et al. (1980) were used as a basis to search for near simultaneous DMSP photos. The cross-section of these data sets was not large. The closest example was a satellite pass indicating shear 20 minutes after a DMSP photo showing strong undulations which had been developing for at least an hour. A theory recently published by Viñas and Madden (1986) is in good agreement with the observations in that, (1) several events studied by Rich et al. satisfied the magnetic Richardson number criterion for instability and (2) the observed wavelengths are in good agreement with the theory. A recent numerical simulation by Pritchett (1985) also seems to support the concept that the edge of the diffuse aurora may be unstable to a shear driven process. An interesting feature of the analysis presented here is that when scaled from ionospheric heights to the equatorial plane, the shear frequency is only reduced by 30–40% from ionospheric values.