In this letter, two direction-of-arrival (DOA) estimators are proposed: one is a new fourth-order cumulant-based technique, and the other is a joint second- and fourth-order cumulant technique. Using the full set of cumulants for DOA estimation is computationally expensive; conversely, using a downsized quadricovariance matrix leads to poor DOA estimates. This letter proposes the choice of efficient quadricovariance matrices to be used in the fourth-order estimator. Then, together with these matrices, a joint second- and fourth-order DOA estimator is derived. Both algorithms perform reasonably better than existing fourth-order and second-order algorithms in a spatially correlated Gaussian noise environment