The QR algorithm is a basic algorithm for computing the eigenvalues of
dense matrices. For efficiency reasons it is prerequisite that the al
gorithm is applied only after the original matrix has been reduced to
a matrix of a particular shape, most notably Hessenberg and tridiagona
l, which is preserved during the iterative process. In certain circums
tances a reduction to another matrix shape may be advantageous. In thi
s paper, we identify which zero patterns of symmetric matrices are pre
served under the QR algorithm.