1999A&A...349..475S

We search for radio counterparts to two recently-detected strong X-ray sources associated with highly embedded young stellar objects (SVS4/EC 95 and SVS16). We detect a radio source (S68-2) consistent with the position of EC95. We fail to detect a counterpart for SVS16, and place upper limits on its quiescent radio brightness. For S68-2, we show that the radio source has a falling spectrum, suggestive of a gyrosynchrotron emission mechanism, and that it is variable on a timescale of years. We search for, but do not detect, evidence for flaring activity on timescales of minutes to hours. We also search for, but do not detect, circular polarisation. We derive the radio luminosity and compare the object to an empirical X-ray - radio luminosity relationship established for dMe stars. We find that the object is consistent with the dwarfs relation, but is unusually X-ray rich compared to other high-luminosity coronal sources. By comparing the objects to a sample of active galactic nuclei in the L_X-L_R diagram, we rule out the possibility that either object is a background AGN. We discuss the ways in which a normal stellar coronal model might be modified to explain the strong, X-ray rich characteristics of the source, which appears to be the most extreme stellar corona yet found.