A stochastic process is said to be stationary (or strictly stationary) if the joint distribution of the sequence of measurements x_1+l, x_2+l,…, x_k+l is, for all k, independent of l. A less restrictive requirement is that the expected value of the x-values should be constant and that, for all k, the covariance between the values of x_k and x_k+l should depend only on the lag l—such a process is described as weakly stationary (or second-order stationary). See also Markov process.