... say.1
``Wovon man nicht reden kann, darüber muss man schweigen'' (L. Wittgenstein). 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... average2
It follows that all moments of the distribution are weighted averages of the moments of the conditional distribution. Then, expected value and variance of $\lambda _s$ can be easily obtained from the conditional expected values and variances:
$\displaystyle \mbox{E}(\lambda_s)$ $\textstyle \propto$ $\displaystyle \sum_{n_s}\, \mbox{E}(\lambda_s\,\vert\,n_s)\,f(n_s)\,$  
$\displaystyle \mbox{Var}(\lambda_s)$ $\textstyle \propto$ $\displaystyle \sum_{n_s}\, [\mbox{Var}(\lambda_s\,\vert\,n_s) + \mbox{E}^2(\lambda_s\,\vert\,n_s)]\,f(n_s)\,.$  

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.