2.3. What could the dark matter be?

The dark matter is not primarily baryonic. The amount of deuterium calculated to emerge from the big bang would be far lower than observed if the average baryon density were ~ 2 (rather than ~ 0.3) per cubic metre. Extra exotic particles that do not participate in nuclear reactions, however, would not scupper the concordance.

Beyond the negative statement that it is non-baryonic, the nature of the dark matter still eludes us. This key question may yield to a three-pronged attack:

1. Direct detection. As described by other contributors to this meeting, several groups are developing cryogenic detectors for supersymmetric particles and axions This is an exciting quest. Of course, not even optimists can be confident that the actual dark matter particles have parameters within the range that these experiments are yet sensitive to. But the stakes are high: detection of most of the gravitating stuff in the universe, as well as a new class of elementary particle. So it seems well worth committing to these experiments funding that is equivalent to a small fraction of the cost of a major accelerator

2. Progress in particle physics. Important recent measurements suggest that neutrinos have non-zero masses; this result has crucially important implications for physics beyond the standard model. The inferred neutrino masses seem, however, too low to be cosmologically important. If the masses and cross-sections of supersymmetric particles were known, it should be possible to predict how many survive, and their contribution to , with the same confidence with which we can compute the nuclear reactions that control primordial nucleosynthesis. Associated with such progress, we might expect a better understanding of how the baryon-antibaryon asymmetry arose, and the consequence for _b. Optimists may hope for progress on still more exotic options.

3. Simulations of galaxy formation and large-scale structure. When and how galaxies form, the way they are clustered, and the density profiles within individual systems, depend on what their gravitationally-dominant constituent is. A combination of better data and better simulations is starting to set generic constraints on the options. The CDM model works well. But there are claimed discrepancies, though many of us suspect these may ease when the galaxy formation process is better understood. For instance the centre of a halo would, according to the simulations, have a `cusp' rather than the measured uniform-density core: this discrepancy has led some authors to explore modifications where the particles are assumed to have significant collision probabilities, or to be moving with non-negligible velocities (i.e. `warm' not cold.). These calculations are in any case offering interesting constraints on the properties of heavy supersymmetric particles. (Also, straight astronomical observations can rule out a contribution to of more than 0.01 from neutrinos - this is compatible with current experimental estimates.)