On the significance of velocity dispersion measurements

Small number statistics may heavily affect the structure of the broadening function in integrated spectra of galactic globular cluster centers. As a consequence, it is a priori unknown how closely line broadening measurements gauge the intrinsic velocity dispersions at the cores of these stellar systems. We have tackled this general problem by means of Monte Carlo simulations. An examination of the mode and the frequency distribution of the measured values of the simulations indicates that the low value measured for the velocity dispersion of M30 (Zaggia et al. 1992b) is likely a reliable estimate of the velocity dispersion at the center of this cluster. The same methodology applied to the case of M15 suggests that the steep inward rise of the velocity dispersion found by Peterson, Seitzer and Cudworth (1989) is real, although less pronounced. Large-aperture observations are less sensitive to statistical fluctuations, but are unable to detect strong variations in the dispersion wich occur within the aperture itself.