Fluctuations in the microwave background will have been imprinted at z ≃ 1000, when the photons and the plasma decoupled. On angular scales greater than a few degrees these fluctuations provide a clear view of any primordial density perturbations, and therefore a clean test of theories which invoke such fluctuations from which to form the structure we see in the universe. On smaller angular scales the predictions are less certain: reionization of the gas may modify the spectrum of the primordial fluctuations, and secondary fluctuations may be generated.
Here I shall review some recent theoretical developments. A brief survey is made of the currently popular theories for the primordial perturbations, with emphasis on the predictions for large scale anisotropy. One major uncetainty in the predictions arises from the normalisation of the fluctuations to e.g. galaxy clustering, and much attention is given to the question of ‘biased’ galaxy formation. The effect of reionization on the primordial fluctuations is discussed, as is the anisotropy generated from scattering off hot gas in clusters, groups and galaxies.