This article provides an efficient algorithm for generating a random matrix
according to a Wishart distribution, but with eigenvalues constrained to b
e less than a given vector of positive values. The procedure of Odell and F
eiveson provides a guide, but the modifications here ensure that the diagon
al elements of a candidate matrix are less than the corresponding elements
of the constraint vector, thus greatly improving the chances that the matri
x will be acceptable. The Normal hierarchical model with vector outcomes an
d the multivariate random effects model provide motivating applications.