Astronomers have already eliminated a number of plausible candidates for the dark matter. X-ray observations of galaxies imply that only a small fraction of the mass of a typical galaxy is in the form of hot gas [6,44]. Even in rich clusters, hot gas makes up less than 20% of the total mass of the system [14]. Neutral hydrogen gas is detectable through its 21 centimeter emission: in most galaxies, neutral gas comprises only 1% of the mass of the system [56] and in only a handful of dwarf galaxies does the neutral gas mass exceed the stellar mass. Even in these systems (e.g., DDO 240 [17]), neutral gas does not account for more than 20% of the system mass. Molecular gas is detectable through dipole emission of CO and other non-homopolar molecules: in most galaxies, the molecular gas mass appears to be less than the neutral gas mass. Low luminosity (low mass) stars, M dwarfs, have often been proposed as a dark matter candidate but HST observations show that faint red stars contribute less than 6% of the unseen matter in the galactic halo [7].
If the dark matter is composed of baryons, then these baryons must be clumped into dense bound objects to evade detection. Gerhard and Silk [29] have proposed that the dark matter consists mostly of very dense tiny clouds of molecular gas. Their model, while provocative, is only marginally consistent with current observational limits. A more widely accepted proposal is that the dark matter consists of very low mass stars, called brown dwarfs. These brown dwarfs are not massive enough to burn hydrogen, so that their only energy source is gravitational energy.
While these brown dwarfs are difficult to detect through their own emission, they are potentially detectable through the gravitational effects. Paczynski [45] proposed gravitational lensing searches for these objects. Several groups have begun searching for these events in an effort to probe the nature of the dark matter.
So far, MACHO searches are not finding as many events as predicted by spherical halo models [3]; however, they can not yet rule out MACHOs as the dominant component of the halo. The current experiment is limited by both small number statistics and by uncertainties in galactic parameters. Many important galactic parameters such as the circular speed, disk scale length and the local surface density are still quite uncertain. Because of these uncertainties, the local halo density is not certain to a factor of two.
It is particularly important to accurately determine the local circular speed as our estimates of the local dark matter density is very sensitive to its value:
(Deriving this formula is a good exercise for a student
new to dynamics. For an excellent introduction
to the subject, see Binney & Tremaine [12]).
Thus, a 10% uncertainty in local circular speed translates
into a 40% uncertainty in the local dark matter density.
Without more accurate determinations of 
, it is
difficult to definitively argue that MACHOs can
not comprise much, if not all, of the mass of the dark halo.
There is also a need for better models of the LMC and more accurate measurements of its properties. Some of the lensing events reported by the MACHO and EROS collaborations may be due to "self-lensing" by the LMC [55] rather than dark matter in the halo.