The unsatisfactory state of index number theory is demonstrated in this paper by showing that the theory lacks a consistent set of rules of choice. It is further proved, by an example, that at least two standard rules conflict with the requirements of economic theory. An error in the literature of index numbers is also corrected.