The Millstone Hill radar, Westford, Massachusetts (&Lgr;=56 ¿), has been upgraded for observations in support of the international magnetosphere study (IMS) by means of the addition of a 150-ft diameter fully steerable antenna. This antenna allows measurements of F region ion drifts, from which F region electric fields can be derived, using the existing UHF (68-cm wavelength) radar. The paper describes measurements of the electric field in the auroral zone over the interval 60 ¿?&Lgr;?75 ¿. The electric field is obtained by means of a least mean squares fit between the drifts obtained in four beam directions and a smoothly varying model with 204 degrees of freedom. Results are presented for 11 moderately quiet days in 1978 and these are compared with the models published by Heppner (1977). All but one of the days exhibit a simple two-cell convection pattern. Two days appear reasonably consistent with the pattern expected when the IMF is toward the sun (symmetric cells) and six when the IMF is away (dusk cell larger). The Harang discontinuity is identifiable in all these cases in extending from 22 to 01 LT. The agreement between the observations and the Heppner (1977) models can be improved by moving the center of the dawn cell into the nightside by perhaps as much as 3 hours in the case of the asymmetric pattern. Two of the 11 days fit neither pattern in that a feature corresponding to the Harang discontinuity cannot be identified. The size and speed of the flow in the convection cells appear to increase with Kp (and/or decreasing Dst) without any obvious change in the direction of the flow (except in one instance near noon). We find no evidence for a permanent narrow 'throat' region in the dayside cleft through which all the flux tubes convect, though it seems possible that this behavior may manifest itself during disturbed periods (Kp>3). At other times, the Heppner (1977) models seem to describe our observations on the dayside quite adequately.