Comments on "On estimating the orders of an autoregressive process"

This correspondence makes some critical remarks on the method suggested in the above paper for estimating the order of an autoregressive process. An alternative method is then suggested.