Aggregate water demand in the Miami area was also esti-
 mated. At a price of $0.80 per thousand gallons it was estimated
 commercial use (from department stores, grocery stores, hotels
 and motels, and eating-drinking establishments) would be in the
 vicinity of 3,382 million gallons per year. This would represent
 nearly half of the total commercial use in the Miami SMSA in the
 early 1970's. The overall demand for water for these four major
 groups of commercial users was also found to be inelastic over
 most price ranges expected in the area.


 Conclusions
 The major conclusion of the study is: commercial business
 establishments appear responsive to the price of water, even
 though generally in an inelastic manner. This has implications
 for pricing policy and water management. Other conclusions
 include the following:
 1. data collection and demand estimation must be accom-
 plished at the 4-digit Standard Industrial Classification
 (SIC) code level.
 2. the derived demand economic model is appropriate to
 the task of guiding the estimation process for com-
 mercial water demand.
 The first of these latter conclusions has implications for further
 research in the area. Demand estimation at this level is an ex-
 pensive process, because demand curves have to estimated for all
 business types in order to be exhaustive. This is a problem simi-
 lar to that faced if derived demands are to be estimated for in-
 dividual agricultural crops or separate industrial processes.
 However, given that economic impacts of alternative water al-
 location strategies are to be established, these demand curves
 must be known.
 The second conclusion has implications for the researcher
interested in estimating commercial water demand. It is argued
herein that the appropriate conceptual basis is the derived de-
mand model, which in turn has its basis in production theory.
Others have used a consumer demand model in an attempt to
explain the same phenomenon [11].
 Further research should also be initiated to establish the
applicability of the estimated models to other areas in Florida
and the nation. We feel confident the general formulations are
appropriate for several areas. Only further empirical study can
serve as the basis for denying the validity of either the general
forms or the particular parameter estimates.