The statistical analysis of magnitude and center-of-mass velocity data for binary galaxies will never provide an adequate estimate of the mass within the orbit, and will therefore never furnish useful information about the mass-to-light ratios of galaxies. There are three main reasons. Firstly, it is very difficult to find a sample which is both objectively selected and dynamically isolated, and in those cases with obvious tidal distortions it is not clear which two points of the confused systems can be taken to be the centers of mass. Secondly, there are several different ways to produce a mass estimator from the available data, and it is not clear which estimator corresponds most closely to what is normally thought of as the gravitational mass of the system. Finally, those mass estimates which include it as a variable depend quite strongly on the orbital eccentricity, and any analysis method is necessarily least sensitive to just this parameter. Therefore, variation even within a single estimator is both considerable and only weakly constrained.