[1] A comparative study of the neutral wind in the polar upper mesosphere/lower thermosphere is conducted using two radars, the European Incoherent Scatter (EISCAT) UHF radar (n = 931 MHz) and the Tromsø MF radar (n = 2.8 MHz) colocated in northern Scandinavia. Comparisons of winds are made in case studies as well as on a statistical basis. On the basis of simultaneous observations obtained by the two radars, comparisons over a height range from $90 to $100 km are made for about 20 days between February and October 1999. In these comparisons we directly compare temporal wind variations. On the other hand, mean winds as well as diurnal and semidiurnal components are derived using 1-month-averaged wind data obtained by the medium frequency (MF) radar for 1999. They are compared with those derived from 56 days of EISCAT data studied by Nozawa and Brekke [1999a, 1999b]. In the case studies, we have also utilized altitude profiles of electron density obtained by the EISCAT radar at and above 62 km height to determine the total reflection height as well as to estimate the effect of group retardation for 2.8 MHz radio wave. It can be seen that the effect of particle precipitation sometimes penetrates into the D region (down to $84 km). Generally, wind velocities derived by EISCAT with the monostatic method exhibit significant scatter with time below 100 km, while the number of the MF radar wind data sharply decreases above 90 km except for summer. These facts make the comparison difficult, but no significant departures between different radar winds are identified on an instantaneous basis. In the statistical studies, generally both altitude profiles are well connected, or continuous, but significant disagreements are observed for the mean wind and the tidal components for summer. On the basis of these comparisons, we raise strong concerns regarding the use of summer wind data above 91 km obtained by MF radar located at high latitude under high solar activity conditions. The wind values during the winter and equinox observed by the two radars are complementary. Thus here is a good opportunity to utilize wind data from the two different radars through a year, which then enables us to cover a wide height range (approximately 70-120 km). This wide range gives us good opportunities to examine the mesosphere/lower thermosphere coupling processes based on observational data. INDEX