> if you combine the spectrum of two objects of different temperatures you do not get anything like a black body curve. Also true for two normally distributed

> if you combine the spectrum of two objects of different temperatures you do not get anything like a black body curve. Also true for two normally distributed variables and yet the normal distribution has the central limit property. Presuming the hypothesis, do you suspect that you'd be looking at exactly two grains of dust, one of which has a crazy high temperature and the other has a crazy low temperature in any given pixel of the CMB?