Parallax data from the Hipparcos mission allow the direct distance to open clusters to be compared with the distance inferred from main-sequence (MS) Ðtting. There are surprising differences between the two distance measurements, indicating either the need for changes in the cluster compositions or reddening, underlying problems with the technique of MS Ðtting, or systematic errors in the Hipparcos parallaxes at the 1 mas level. We examine the di†erent possibilities, focusing on MS fitting in both metallicity-sensitive B[V and metallicity-insensitive V [I for five well-studied systems (the Hyades, Pleiades, a Per, Praesepe, and Coma Ber). The Hipparcos distances to the Hyades and a Per are within 1 p of the MS-Ðtting distance in B[V and V [I, while the Hipparcos distances to Coma Ber and the Pleiades are in disagreement with the MS-Ðtting distance at more than the 3 p level. There are two Hipparcos measurements of the distance to Praesepe; one is in good agreement with the MS-fitting dis-tance and the other disagrees at the 2 p level. The distance estimates from the different colors are in connect with one another for Coma but in agreement for the Pleiades. Changes in the relative cluster metal abundances, age related effects, helium, and reddening are shown to be unlikely to explain the puzzling behavior of the Pleiades. We present evidence for spatially dependent systematic errors at the 1 mas level in the parallaxes of Pleiades stars. The implications of this result are discussed.