1: \begin{proof}

2:   It has been proved in \cite{LoTo} that $\L(\mathcal{M})=0$. Now,

3:   since $\mathcal{M}$ is a Maxwellian distribution, it is also

4:   well-known that $\Q_1(\mathcal{M},\mathcal{M})=0$ and this proves

5:   that $\mathcal{M}$ is a solution to~\eqref{equi}. To prove that it

6:   is the unique solution with unit mass, one proceeds {in some formal way for the time being assuming that $F$ decays sufficiently fast at infinity; we will see that it is actually the case in the following section. All the proof can then be made rigorous thanks to the subsequent Theorem \ref{gaus}}. For

7:   any distribution $F(\v) \geq 0$ solution to \eqref{equi}, let us

8:   multiply \eqref{equi} with $\log

9:   \left(\frac{F(\v)}{\mathcal{M}(\v)}\right)$ and integrate with

10:   respect to $\v$. One gets

11:   $$0=\IR \Q_1(F,F)(\v)\log

12:   \left(\frac{F(\v)}{\mathcal{M}(\v)}\right)\d\v + \IR \L(F)(\v)\log

13:   \left(\frac{F(\v)}{\mathcal{M}(\v)}\right)\d\v$$ and it is

14:   well-known from \cite{LoTo} and \cite{Ce94} that both the integrals

15:   in the above sum are nonpositive. Therefore,

16:   $$\IR \Q_1(F,F)(\v)\log

17:   \left(\frac{F(\v)}{\mathcal{M}(\v)}\right)\d\v =0.$$ Since

18:   $\mathcal{M}$ is a Maxwellian distribution, it is a well-established

19:   fact that

20:   $$\IR \Q_1(F,F)(\v) \log \mathcal{M}(\v) \d \v=0.$$

21:   Consequently, $F$ is such that

22:   $$\IR \Q_1(F,F)(\v)\log F(\v) \d \v=0$$ and the classical Boltzmann

23:   $H$-Theorem \cite{Ce94} asserts that $F$ is a given Maxwellian and

24:   $\Q_1(F,F)=0.$ Consequently, one has $\L(F)=0$ and, from the

25:   uniqueness result \cite{LoTo}, $F=\mathcal{M}.$

26: \end{proof}