# Plotting the CMB power spectrum - Why $C_\ell \ell (\ell+1)$ rather than only $C_\ell$?

I can't find any convincing answer for the following question :

Why do we always (or often) plot the CMB power spectrum in this way? I mean the vertical axis is $$C_\ell \ell (\ell+1)$$ and not only $$C_\ell$$. Why?

I know it's because of the scale invariance, but why do we absolutely want to show the flat line at low $$\ell$$? And I do not understand why the power spectrum is flat in this scale.

As you mention, for the Sachs-Wolfe effect the $C_{\ell}$ values drop off as approximately $\ell(\ell + 1)$ so plotting $C_{\ell}\ell(\ell + 1)$ on the $y$ axis gives an approximately horizontal line and this makes it easy to see deviations from Sachs-Wolfe behaviour. However I suspect the main reason the graphs are drawn this way is that it nicely highlights the doppler peaks. If you just plotted $C_{\ell}$ you'd need to use a log axis and that would make all the peaks look smaller.
• Well, because we're really plotting the anisotropy i.e. variations. So they're the Fourier modes not of the temperature $T$ itself but its Laplacian over the sphere, $\Delta T$, and the Laplacian has a simple well-defined effect on the component $C_l$ which is multiplied by a spherical harmonic function $Y_{lm}$: it just multiplies the spherical harmonic function by $-l(l+1)$. That's why $l(l+1)$ may be identified with the (minus) Laplacian. It's more natural to insert the Laplacian rather than not to really measure "variations" and to get rid of the huge constant term prop. to $Y_{00}$, too. – Luboš Motl May 19 '12 at 8:48
$C_{\ell}\ell(\ell + 1) \propto (\Delta T)^2$