A scalar, \lambda, such that there exists a non-zero vector x (a corresponding eigenvector) for which the image of x under a given linear operator \mathrm{A} is equal to the image of x under multiplication by \lambda; i.e. \mathrm{A} x = \lambda x.
The eigenvalues \lambda of a square transformation matrix \mathrm{M} may be found by solving \det(\mathrm{M} - \lambda\mathrm{I}) = 0.