Top: the new astrometric solution computed with an assumed radio Cosmic Noise of 2.4 mas. The 1997 HIAD Hipparcos data are shown for illustration (see Harper et al. 2008). On this scale, the 1997 Hipparcos data are not distinguishable from the 2007 data. The apparent angular diameter of the star at different wavelengths is shown in the upper-right corner for reference—the inner circle is H-band (﹩1.6\,\mu {\rm{m}}﹩), the yellow is ALMA (338 GHz), and blue is Q-band (43 GHz). Bottom: the 2007 Hipparcos solution, which shows how the radio data constrains the proper motions. The radio data show that Hipparcos systematically overestimates ﹩{\mu }_{\delta }﹩ and, to a lesser extent, ﹩{\mu }_{\alpha \cos \delta }﹩.