|Annu. Rev. Astron. Astrophys. 1979. 17:
Copyright © 1979 by . All rights reserved
Is there more to a galaxy than meets the eye (or can be seen on a photograph)? Many decades ago, Zwicky (1933) and Smith (1936) showed that if the Virgo cluster of galaxies is bound, the total mass must considerably exceed the sum of the masses of the individual member galaxies; i.e. there appeared to be ``missing mass'' in the cluster. As more data became available, the discrepancy persisted between masses of individual galaxies determined from optical rotation curves and the larger average galaxy mass needed to bind groups and clusters (e.g. Neyman et al. 1961).
Recently, however, new information has pointed toward larger total masses for individual galaxies, thus decreasing the traditional discrepancy between various methods of mass measurement. Arguing that thin self-gravitating stellar disks are unstable against bar-like modes, Ostriker & Peebles (1973) suggested that the disks of normal spiral galaxies must be imbedded in optically undetected, stabilizing massive halos. Ostriker et al. (1974) and Einasto et al. (1974) collected observational evidence in support of the existence of such halos (although Burbidge 1975 used similar data to reach the opposite conclusion). At nearly the same time, high-resolution 21-cm observations of nearby galaxies were showing that H I often extends well beyond the optical boundaries of galaxies and that rotation velocities are constant at large galactocentric distances. Simply interpreted, these measurements implied the presence of substantial mass outside the optically visible dimensions of galaxies.
In this review, then, we are especially concerned with the current status of the ``missing mass'' problem: has it been resolved by new data, or does it linger on essentially unchanged in magnitude? To answer this question we rely on mass-to-light ratios as our primary tool, since they provide a direct intercomparison of galaxy masses measured for many different samples using varied techniques. We further assume that all Doppler shifts are caused by actual velocities of recession, though this view is not universally held (e.g. Arp 1974). Finally, in discussing the possible presence of invisible mass in galaxies, we do not wish to assume any model for its structure or spatial distribution. For this reason, we choose the neutral term ``massive envelope'' to describe the unseen mass. Although ``massive halo'' is often used in a similar context, it connotes a more or less smooth and spherical mass distribution, a possibly misleading notion since the present data contain little actual information on the spatial structure of any extended components.
In comparing mass-to-light ratios from different sources, we use a standard system of M / LB. We define total magnitudes of galaxies on the BT system of the Second Reference Catalogue of Bright Galaxies (de Vaucouleurs et al. 1976, hereafter RC2). We have corrected the BT magnitudes for internal extinction using the precepts of the RC2, which are nearly independent of type. The resultant luminosities are ``face-on'' values only; an additional 10-20% increase would be necessary to produce totally absorption-free magnitudes, which, strictly speaking, are those with which the local M / LB for the solar neighborhood should be compared (Section 2.1). The galactic extinction assumed is AB = 0.133(csc |b| - 1), close to Sandage's (1973) formulation and in reasonable agreement with the more recent results of Burstein & Heiles (1978). The value of the solar absolute magnitude is here taken to be +5.48 in B (Allen 1973) or +5.37 in the photographic system used by Holmberg (Stebbins & Kron 1957). Finally, H0 = 50 km s-1 Mpc-1 is used consistently throughout.
All values of M / LB used in this paper are corrected to this standard system. Failure to adopt a standard system of M / LB can easily lead to errors of a factor of two or three in comparisons among mass-to-light ratios from various authors. The standard system adopted here is sensitive to the adopted magnitude and extinction corrections. The resulting uncertainties in the overall scale are ± 30-40%.