The age of β Pictoris

We have reanalyzed data for the proposed moving group associated with beta Pictoris in order to determine if the group (or part of it) is real and, if so, to derive an improved age estimate for beta Pie. By using new, more accurate proper motions from PPM and Hipparcos and a few new radial velocities, we conclude that on kinematic grounds, most of the proposed members of the moving group are not, in fact, associated with beta Pie. However, two M dwarfs-or three, actually, since one of them is a nearly equal mass binary-have space motions that coincide with that of beta Pic to within 1 km s(-1) with small error bars. Based on a color-magnitude diagram derived from accurate photometry and Hipparcos parallaxes, these two possible proper-motion companions to beta Pic are very young; we derive an age of similar to 20 Myr by comparison with theoretical tracks from F. D'Antona & I. Mazzitelli. In fact, the proposed beta Pic companions comprise two of the three youngest M dwarfs in the sample of 160 dM stars for which we have data. The chromospheric and coronal activities of these two stars also confirm that they are quite young. We argue that the probability that two of the three youngest nearby M dwarfs would accurately share the space motion of beta Pic by chance is quite small, and therefore we believe that beta Pic and the two M dwarfs (GL 799 and GL 803) were formed together. The estimated age for beta Pic is then 20 +/- 10 Myr, where the uncertainty in the age arises primarily from possible errors in the pre-main-sequence isochrones and in the conversion from color to effective temperature. This young age for beta Pic supports the contention that the IR excess for the Vega-like stars is age dependent.