Cross-talk between output channels is a common problem that can occur in the wet-end or dry-end electronic cabling of a sensor array. This paper examines artifacts in spatial spectral estimates caused by cross-talk between sensor output channels. It is shown that cross-talk can result in spurious peaks in spatial spectral estimates, particularly for high resolution methods that involve the covariance matrix inverse. An analysis of the eigenvalues and eigenvectors of the array output covariance matrix in the presence of cross-talk contamination is given in the noise-only case. Cross-talk contamination is linked to a reduction in the magnitude of the smallest eigenvalue of the covariance matrix and thus the most pronounced artifacts are found in spatial spectral estimates that involve eigenvalue reciprocals. Various examples using both simulated and real towed-array data are discussed.