Annu. Rev. Astron. Astrophys. 1994. 32: 531-590 Copyright © 1994 by Annual Reviews. All rights reserved

2. EVIDENCE FOR DARK MATTER

The observational evidence for dark matter arises in many different contexts and baryonic dark matter is not necessarily implicated in all of these. We therefore begin by identifying the observational issues most relevant to the baryonic versus nonbaryonic dilemma.

2.1. Local Dark Matter

Measurements of the stellar velocity and density distribution perpendicular to the Galactic disk provide an estimate of the total disk density. This turns out to be about 0.1 M pc^-3 and it has long been suspected (Oort 1932) that this exceeds the density in visible stars. The possibility of disk dark matter is very important in the present context because - of all the dark matter problems - this is the one most likely to have a baryonic solution. Unfortunately, the evidence is very controversial. Bahcall (1984a, b, c) used counts of F dwarfs and K giants to conclude that the density of unseen material must be at least 50% that of the visible material. He also concluded that the disk dark matter must have an exponential scale height of less than 700 pc, so that it must itself be confined to a disk. However, Bahcall assumed a particular model and Bienayme et al (1987), using a different model, found a best-fit dark matter density of only 0.01 M pc^-3, and even this could be removed if the halo was slightly flattened. Knapp (1988) came to the same conclusion by studying the velocity dispersion and scale height of molecular hydrogen. Further doubt was cast in a series of papers by Kuijken & Gilmore (1989), who used the full distribution function for the velocities and distances of K dwarfs rather than assuming a particular model. Although Gould (1990) used a maximum likelihood analysis to conclude that Kuijken & Gilmore's data were not inconsistent with the Bahcall et al claim, Kuijken & Gilmore (1991) disagreed with this. More recently, Bahcall et al (1992a) have concluded from another analysis of K giants that the no-disk-dark-matter hypothesis is only consistent with the data at the 14% level and their best-fit model has a dark density of 0.15 M pc^-3, which corresponds to 53% more dark matter than visible matter. For present purposes the existence of disk dark matter will be regarded as an open question.