1. INTRODUCTION

Sometimes a theory is proposed in relatively early stages of the development of a scientific field, and this theory turns out to be not only a useful paradigm for the further development of the field - it also survives confrontation with a vast amount of data, and becomes accepted as the standard theory. This happened with General Relativity [2], and it seems to be happening now with general relativistic cosmology. It appears that the universe on the largest scales can indeed be described by three numbers:

• H₀ 100hkm s^-1 Mpc^-1, the Hubble parameter (expansion rate of the universe) at the present epoch,

• _m / _c, the density of matter in units of critical density _c 3H₀² (8G)^-1 = 2.78 × 10¹¹ h² M Mpc^-3, and

• (3H₀²)^-1, the corresponding quantity for the cosmological constant.

The currently measured values of these and other key parameters are summarized in the Table below. It remains to be seen whether the "dark energy" represented by the cosmological constant is really constant, or is perhaps instead a consequence of the dynamics of some fundamental field as in "quintessence" theories [3].

In particle physics, the first unified theory of the weak and electromagnetic interactions [4] had as its fundamental bosons just the carriers of the charged weak interactions W⁺, W^-, and the photon . The next such theory [5] had a slightly more complicated pattern of gauge bosons - a triplet plus a singlet, out of which came not only W⁺, W^-, and , but also the neutral weak boson Z⁰, and correspondingly an extra free parameter, the "Weinberg angle." It was of course this latter SU(2) × U(1) theory which has now become part of the Standard Model of particle physics. During the early 1970s, however, when the experimental data were just becoming available and some of the data appeared to contradict the SU(2) × U(1) theory, many other more complicated theories were proposed, even by Weinberg [6], but all these more complicated theories ultimately fell by the wayside.

The development of theories of dark matter may follow a similar pattern. By the late 1970s it was becoming clear both that a great deal of dark matter exists [7] and that the cosmic microwave background (CMB) fluctuation amplitude is smaller than that predicted in a baryonic universe. The first nonbaryonic dark matter candidate to be investigated in detail was light neutrinos - what we now call "hot dark matter" (HDM). This dark matter is called "hot" because at one year after the big bang, when the horizon first encompassed the amount of matter in a large galaxy like our own (about 10¹² M) and the temperature was about 1 keV [8], neutrinos with masses in the eV range would have been highly relativistic.

It is hardly surprising that HDM was worked out first. Neutrinos were known to exist, after all, and an experiment in Moscow that had measured a mass for the electron neutrino m(_e) 20 eV (corresponding to _m 1 if h were as small as ~ 0.5, since = m(_e)(92h²eV)^-1) had motivated especially Zel'dovich and his colleagues to work out the implications of HDM with a Zel'dovich spectrum ( P_p(k) = Akⁿ with n = 1) of adiabatic primordial fluctuations. But improved experiments subsequently have only produced upper limits for m(_e), currently about 3 eV [9], and the predictions of the adiabatic HDM model are clearly inconsistent with the observed universe [10, 11].

Cold Dark Matter (CDM) was worked out as the problems with HDM were beginning to become clear. CDM assumes that the dark matter is mostly cold - i.e., with negligible thermal velocities in the early universe, either because the dark matter particles are weakly interacting massive particles (WIMPs) with mass ~ 10² GeV, or alternatively because they are produced without a thermal distribution of velocities, as is the case with axions. The CDM theory also assumes, like HDM, that the fluctuations in the dark matter have a nearly Zel'dovich spectrum of adiabatic fluctuations. Considering that the CDM model of structure formation in the universe was proposed almost twenty years ago [12, 13, 14], its successes are nothing short of amazing. As I will discuss, the CDM variant of CDM with _m = 1 - 0.3 appears to be in good agreement with the available data on large scales. Issues that have arisen on smaller scales, such as the centers of dark matter halos and the numbers of small satellites, have prompted people to propose a wide variety of alternatives to CDM, such as self-interacting dark matter (SIDM) [15]. It remains to be seen whether such alternative theories with extra parameters actually turn out to be in better agreement with data. As I will discuss below, it now appears that SIDM is probably ruled out, while the small-scale predictions of CDM may be in better agreement with the latest data than appeared to be the case as recently as a year ago.

In the next section I will briefly review the current observations and the successes of CDM on large scales, and then I will discuss the possible problems on small scales.