Measurements of total electron content (TEC) variations are easy to perform. As several authors have pointed out, however, TEC variations must be interpreted carefully since TEC is an integrated quantity. Prior studies of TEC variations have considered these variations to stem from classical gravity-wave interaction with the ionosphere at midlatitudes; that is, the component of gravity-wave neutral wind perturbations along the local magnetic field moves the charged particles up and down the field lines. Recent evidence suggests that at night, in the absence of photoproduction and when E layer conductivity becomes small, electrodynamic effects come into play and gravity-wave perturbations can yield vertical movement of the F layer, rather than movement along the field lines. Here we consider the TEC variations resulting from horizontally propagating disturbances in ionospheric height. First, we develop an analytic model, assuming sinusoidal height variations and a geostationary satellite, to estimate the magnitude of TEC variations produced. In this case, the dependence of TEC variations on the orientation of the line of sight, relative to the direction of disturbance propagation, takes on a different form from the classical case. Total electron content variations may also be observed using nonstationary satellites, such as Global Positioning System (GPS) satellites. We present GPS measurements from Puerto Rico for the night of August 22--23, 1995, which show TEC variations of ~¿5¿1015 m-2 in the presence of a southwesterly propagating ionospheric height disturbance (as independently corroborated by 630-nm airglow images at Arecibo and Ramey digisonde measurements). Although GPS satellite motion does not permit direct application of the geostationary theory, geostationary theory still allows us to make a crude estimate of the magnitude of TEC variations that result.¿ 1997 American Geophysical Union