The questions here are: can indirect methods relying on estimating Engel curves provide mathematically simpler ways of estimating price elasticities in demand systems than direct methods and, will the results be an improvement over available direct methods? In answering these questions a number of issues surface. The first has to do with specifying the size of the problem. As a practical matter, separability must be assumed. It can, however, appear in various ways. Second, if an indirect method is to be used to recover demand function parameters then we must consider the functional form of the estimating equations and the number of samples needed to obtain estimates. The samples may be from the same population at different points in time, or may be from different parts of the same population at the same time. In both cases price regimes must differ among samples.