Linear combination of jointly normal random variables

Every linear combination $a^{T}X$ of the coordinates of a jointly normal random vector is a normal random variable. Its mean is $a^T\mu$ and its variance is $a^T\Sigma a$.