How Many Numerical Eigenvalues Can We Trust?

When using finite element and finite difference methods to approximate eigenvalues of 2mth-order elliptic problems, the number of reliable numerical eigenvalues can be estimated in terms of the total degrees of freedom $$N$$N in resulting discrete systems. The truth is worse than what we used to believe in that the percentage of reliable eigenvalues decreases with an increased $$N$$N, even though the number of reliable eigenvalues increases with $$N$$N.