CMB lower multipoles and geometry constraints In recent analysis of the microwave background (e.g. that one which lead to the dodecahedral model) it is argued that, if the power spectrum of the lower multipoles shows some kind of suppression, this may be due to some geometrical condition, like the impossibility for the big fluctuations to take place completely in a curved universe. My question is the following: since the first peak in the cmb corresponds with a fluctuation with horizon size during recombination, why should not this fluctuation be also affected by this power suppression? With other words: how should I imagine the observable universe during recombination in an hypotetical curved or closed universe? If it should be also a curved shape, then I would say the first peak might be -more or less- affected by this suppression, since the first peak spanned the whole observable universe. Only higher multipoles should remain unaffected. But if it should be a 'flat' patch into a big curved shape, then I would say the first peak may be unaffected. Please some help for a qualitative understanding. Thanks.