Before the COBE detection, there had been about 25 years of quoted upper limits to CMB anisotropy. There was much talk in the literature about how new paradigms would be required if COBE returned yet stronger upper limits. But in fact the detection was just at the right level for gravity alone to have grown the structure from amplitudes of ~ 10-5 at z ~ 1000 to the non-linear structures we see today. This is easy to arrange for models in which the Universe is dominated by non-baryonic dark matter, and with adiabatic perturbations. Thus, probably the most important thing to come out of the COBE anisotropy measurements (apart from the general good news that we are on the right track!) is the realization that
This `fact', arising from the CMB, has added to the Big Bang paradigm, so that the picture is of a hot expanding Universe, which at early times contained small amplitude density perturbations. The obvious next question is where those perturbations came from - an issue we shall return to in a minute.
Figure 3. The result of binning the data in the previous figure. More precisely, what was done was to split the multipoles into 16 bins between l = 2 and l = 1000, and to weight each experiment by the fraction of the window in each bin. The precise height of the peak depends to some extent on the choice of bins, on details of the window functions used, and on the weights given to individual experiments. The points here are not un-correlated, but provide a reasonable visual impression of the current data - more sophisticated treatments (e.g.  give similar results). The solid line is a -dominated CDM model, with parameters which are consistent with most current cosmological constraints.
Meanwhile, there are a few other things that the current suite of CMB measurements tells us. First of all, it is pleasing that the approximate scale of the peak (apparent in the binned plot, Figure 3) is just where it is theoretically predicted in simple models. This acoustic peak corresponds to the length scale which a sound wave can travel at the time of recombination, projected onto the sky, and was contained in papers at least as early as Doroshkevich, Sunyaev & Zel'dovich (1978) . The position in angle is also a good test for the geometry of the Universe, since it comes from the projection of a fixed physical scale onto the sky.
We know (from the lack of complete absorption shortwards of the Lyman edge in distant quasars) that most of the material in the nearby Universe is ionized, out to redshifts z > 5. Whether the Universe reionized at z ~ 10 or z ~ 1000 is, however, not obvious. But, very early reionization would lead to the erasing of the small-scale CMB anisotropies, which patently has not occurred. Hence we can infer that
(see e.g. ).