In modern terms, by far the most important property of inflationary cosmology is that it produces spectra of both density perturbations and gravitational waves. The density perturbations may be responsible for the formation and clustering of galaxies, as well as creating anisotropies in the microwave background radiation. The gravitational waves do not affect the formation of galaxies, but as we shall see may contribute extra microwave anisotropies on the large angular scales sampled by the COBE satellite. [19, 20] An alternative terminology for the density perturbations is scalar perturbations and for the gravitational waves is tensor perturbations, the terminology referring to their transformation properties.
Studies of large-scale structure typically make some assumption about the initial form of these spectra. Usually gravitational waves are assumed not to be present, and the density perturbations to take on a simple form such as the scale-invariant Harrison-Zel'dovich spectrum, or a scale-free power-law spectrum. It is clearly highly desirable to have a theory which predicts the forms of the spectra. There are presently two rival models which do this, cosmological inflation and topological defects. At present inflation is favoured both on observational grounds and because it provides a simpler framework for understanding the evolution of structure