We study active region NOAA 8541, observed with instruments on board SOHO, as well as with TRACE. The data set mainly covers the transition region and the low corona. In selected loops studied with SUMER on SOHO, the VIII 770 Å line is systematically redshifted. In order to estimate the plasma velocity, we combine the Doppler shifts with proper motions (TRACE) along these loops. In the case of an ejection, apparently caused by the emergence of a parasitic polarity, proper motions and Doppler shifts give consistent results for the velocity. A cooler loop, observed in the same active region with CDS, shows a unidirectional motion reminiscent of a siphon flow. The derived electron temperature and density along a large steady loop confirm that it cannot be described by hydrostatic models.