Several kinds of astronomical observations, interpreted in the framework of the standard Friedmann-Robertson-Walker cosmology, have indicated that our universe is dominated by a Cosmological Constant. The dimming of distant Type Ia supernovae suggests that the expansion rate is accelerating, as if driven by vacuum energy, and this has been indirectly substantiated through studies of angular anisotropies in the cosmic microwave background (CMB) and of spatial correlations in the large-scale structure (LSS) of galaxies. However there is no compelling direct evidence yet for (the dynamical effects of) dark energy. The precision CMB data can be equally well fitted without dark energy if the spectrum of primordial density fluctuations is not quite scale-free and if the Hubble constant is lower globally than its locally measured value. The LSS data can also be satisfactorily fitted if there is a small component of hot dark matter, as would be provided by neutrinos of mass 0.5 eV. Although such an Einstein-de Sitter model cannot explain the SNe Ia Hubble diagram or the position of the `baryon acoustic oscillation' peak in the autocorrelation function of galaxies, it may be possible to do so e.g. in an inhomogeneous Lemaitre-Tolman-Bondi cosmology where we are located in a void which is expanding faster than the average. Such alternatives may seem contrived but this must be weighed against our lack of any fundamental understanding of the inferred tiny energy scale of the dark energy. It may well be an artifact of an oversimplified cosmological model, rather than having physical reality.

12 pages, 5 figures; to appear in a special issue of General Relativity and Gravitation, eds. G.F.R. Ellis et al; Changes: references reformatted in journal style - text unchanged