Under Section 2.1, "Multivariate Normal", the equations could be more generalized. The first equation AX \sim N(A \mu,A \Sigma A^T) could be generalized to AX+b \sim N(A \mu + b,A \Sigma A^T).
Also, when talking about transformations between X and a N(0,I), the first equation holds, even if \Sigma is not full rank. The first transformation could be given, then after introducing that \Sigma must be full rank, the inverse transformation could be given.
Also under that section, there's a stray "Conditional distribution".
Under Section 2.1, "Multivariate Normal", the equations could be more generalized. The first equation AX \sim N(A \mu,A \Sigma A^T) could be generalized to AX+b \sim N(A \mu + b,A \Sigma A^T).
Also, when talking about transformations between X and a N(0,I), the first equation holds, even if \Sigma is not full rank. The first transformation could be given, then after introducing that \Sigma must be full rank, the inverse transformation could be given.
Also under that section, there's a stray "Conditional distribution".