4 Discussion

The observations of an amplified stellar image in the H₂O masers of U Her indicate that at least for the brightest spot at this epoch the maser beaming is radial. Reid & Menten (1997) have detected 22 GHz continuum emission from a small sample of Mira stars, finding the typical stellar brightness temperature to be T_* = 1600 K. Compared to their estimate of the maser excitation temperature ($\approx$10 K), this is strong enough to influence the H₂O maser medium and produce a stellar image, as the increased seed radiation from the star will cause the radial maser beam to be brighter. Since the H₂O masers are found to be mostly unsaturated, slight changes in density, pumping and velocity structure have a strong effect on the maser and the relative strength of the maser features and the amplified stellar image may be less dominant than the OH, as is demonstrated by the detection by Reid & Menten (1990) of an H₂O maser feature at the stellar position of W Hya which was several orders of magnitude weaker than the strong feature observed here.

The H₂O masers around U Her have been observed before with MERLIN, the VLA and the VLBA. VLA observations by Colomer et al. (2000), and MERLIN observations by Bains et al. (in preparation) show an incomplete ring structure with a scale of 150-200 mas. The brightest maser spot seen in our MERLIN observations corresponds in velocity with the masers on the edge of the ring structure. So somewhat surprisingly, our astrometric results indicate that the star is not in the center of this ring.

The maser spots detected with high resolution VLBA observations do not show any indication of circular structure (Vlemmings et al. 2002). They have a linear extent of $\approx$60-70 mas and they most likely correspond to the brightest VLA and MERLIN features.