**4.4.1. Growing mode perturbations**

In general perturbations evolving on scales larger than the horizon have both growing and decaying modes, which on horizon entry become the baryonic oscillations. However because inflationary perturbations were laid down in the distant past, they evolve to become completely dominated by the growing mode by the time they enter the horizon. This leads to a fixed phase of oscillations for the baryons on a given scale; for example, at any given time (such as decoupling) there are particular scales on which the oscillations are at zero amplitude. On those scales the corresponding microwave anisotropies will be at their smallest. This contrasts with topological defect scenarios, where the modes are sourced after entering the horizon and in general have a mixture of the two perturbation modes, meaning that the phase of oscillation is not determined.

The growing mode prediction is perhaps the most important generic prediction of inflation, and one that cannot be avoided. It leads to characteristic predictions; the phase coherence of the oscillations that results is what leads to the oscillatory structure in the microwave anisotropy spectrum [21]. Such oscillations will arise whether the inflationary perturbations are gaussian or non-gaussian, and whether they are adiabatic or non-adiabatic. The prediction of an oscillatory structure in the microwave anisotropies is therefore a key testable prediction of inflation. That is not to say that the discovery of such would prove inflation, as other mechanisms may prove capable of also creating oscillations. But if the oscillations are not seen, the paradigm of inflation as the sole origin of structure will be excluded. The combined BOOMERanG and Maxima data are not quite sufficient for a definitive verdict, but tests of this prediction are imminent.