The notion of "neutral" want association is introduced. The relationship between this and strictly additive utility, or "want independence," is discussed. It is argued that the expression of neutral want association always in terms of want independence obscures the generality of the former concept and inhibits further progress. A simple empirical test for neutral want association is proposed. It is shown that the number of parameters to be estimated in a complete set of demand equations can be drastically reduced by grouping in accordance with want relationships. By following the initial grouping with further groupings of successively higher orders, all parameters may be estimated from as few as two sets of observations of price and quantity changes. It is shown that in general no index number problem is involved in this process.