The most important cosmological observation in the last forty years has undoubtedly been the discovery of the microwave background. As well as confirming the existence of a hot early phase of the Universe, by its spectrum, its remarkable isotropy indicates that the Universe must be very nearly spherically symmetric about us. Because of the revolution of thought brought about by Copernicus, we are no longer vain enough to believe that we occupy any special position in the Universe. We must assume, therefore, that the radiation would appear similarly isotropic in any other place. One can show that the microwave radiation can be exactly isotropic at every point only if the Universe is exactly spatially homogeneous and isotropic, that is to say, it is described by one of the Friedmann models. (Ehlers et al., 1968). Of course, the Universe is neither homogeneous nor isotropic locally. This must mean that the background radiation is not exactly isotropic, but only isotropic to within the very good limits set by the observations (about 0.1%). One would like to know, however, what limits the observations place on the large-scale anisotropies and inhomogeneities of the Universe. One would also like to know why it is that the Universe is so nearly, but not exactly, isotropic.