This paper is a review of many of the dozens of procedures currently available for testing a data set for goodness-of-fit to the multivariate normal distribution. A majority of the procedures can be placed into one of four basic categories. Most procedures are multivariate extensions or adaptations of procedures used for testing univariate normality. Results of several power studies are summarized, and an extensive bibliography of literature pertaining to testing for multivariate normality is provided.
"An Appraisal and Bibliography of Tests for Multivariate Normality." Internat. Statist. Rev. 72 (1) 123 - 138, April 2004.